The field of computer graphics was revolutionized by models such as NeRF and 3D Gaussian Splatting, displacing triangles as the dominant representation for photogrammetry. In this paper, we argue for a triangle comeback. We develop a differentiable renderer that directly optimizes triangles via end-to-end gradients. We achieve this by rendering each triangle as differentiable splats, combining the efficiency of triangles with the adaptive density of representations based on independent primitives. Compared to popular 2D and 3D Gaussian Splatting methods, our approach achieves competitive rendering and convergence speed, and demonstrates high visual quality. On the Mip-NeRF360 dataset, our method outperforms concurrent nonvolumetric primitives in visual fidelity and achieves higher perceptual quality than the state-of-the-art Zip-NeRF on indoor scenes. Triangles are simple, compatible with standard graphics stacks and GPU hardware, and highly efficient. Our results highlight the efficiency and effectiveness of triangle-based representations for high-quality novel view synthesis. Triangles bring us closer to mesh-based optimization by combining classical computer graphics with modern differentiable rendering frameworks. The project page is https://trianglesplatting.github.io/

Alhim Adonai Vera Gonzalez, Carlos Hinojosa, Karen Sanchez, Haidar Bin Hamid, Donghoon Kim, Bernard Ghanem. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

This paper presents a comprehensive study on the role of Classifier-Free Guidance (CFG) in text-conditioned diffusion models from the perspective of inference efficiency. In particular, we relax the default choice of applying CFG in all diffusion steps and instead propose to search for more efficient guidance policies. We formulate the discovery of such policies in the framework of differentiable neural architecture search. Our findings suggest that, as denoising progresses, the updates produced by CFG become increasingly aligned with simple conditional steps, which renders CFG's additional neural network evaluation redundant, especially in the second half of the denoising process. Building upon this insight, we propose "Adaptive Guidance" (AG), an efficient variant of CFG that adaptively omits network evaluations when the denoising process displays convergence. Our experiments demonstrate that AG preserves CFG's image quality while reducing computation by 25%. Thus, AG constitutes a plug-and-play alternative to Guidance Distillation, achieving 50% of the speed-ups of the latter, while being training-free and retaining the capacity to handle negative prompts. We conclude by uncovering further redundancies of CFG in the first half of the diffusion process, showing that entire neural network evaluations can be replaced by simple affine transformations of past score estimates.

Large video-language models (VLMs) have demonstrated promising progress in various video understanding tasks. However, their effectiveness in long-form video analysis is constrained by limited context windows. Traditional approaches, such as uniform frame sampling, often inevitably allocate resources to irrelevant content, diminishing their effectiveness in real-world scenarios. In this paper, we introduce BOLT, a method to BOost Large VLMs without additional Training through a comprehensive study of frame selection strategies. First, to enable a more realistic evaluation of VLMs in long-form video understanding, we propose a multi-source retrieval evaluation setting. Our findings reveal that uniform sampling performs poorly in noisy contexts, underscoring the importance of selecting the right frames. Second, we explore several frame selection strategies based on query-frame similarity and analyze their effectiveness at inference time. Our results show that inverse transform sampling yields the most significant performance improvement, increasing accuracy on the Video-Mme benchmark from 53.8% to 56.1% and MLVU benchmark from 58.9% to 63.4%. Our code is available at https://github.com/sming256/BOLT.

Recent advances in radiance field reconstruction, such as 3D Gaussian Splatting (3DGS), have achieved high-quality novel view synthesis and fast rendering by representing scenes with compositions of Gaussian primitives. However, 3D Gaussians present several limitations for scene reconstruction. Accurately capturing hard edges is challenging without significantly increasing the number of Gaussians, creating a large memory footprint. Moreover, they struggle to represent flat surfaces, as they are diffused in space. Without hand-crafted regularizers, they tend to disperse irregularly around the actual surface. To circumvent these issues, we introduce a novel method, named 3D Convex Splatting (3DCS), which leverages 3D smooth convexes as primitives for modeling geometrically-meaningful radiance fields from multi-view images. Smooth convex shapes offer greater flexibility than Gaussians, allowing for a better representation of 3D scenes with hard edges and dense volumes using fewer primitives. Powered by our efficient CUDA-based rasterizer, 3DCS achieves superior performance over 3DGS on benchmarks such as MipNeRF360, Tanks and Temples, and Deep Blending. Specifically, our method attains an improvement of up to 0.81 in PSNR and 0.026 in LPIPS compared to 3DGS while maintaining high rendering speeds and reducing the number of required primitives. Our results highlight the potential of 3D Convex Splatting to become the new standard for high-quality scene reconstruction and novel view synthesis. The project page is https://convexsplatting.github.io

Online continual learning (OCL) seeks to learn new tasks from data streams that appear only once, while retaining knowledge of previously learned tasks. Most existing methods rely on replay, focusing on enhancing memory retention through regularization or distillation. However, they often overlook the adaptability of the model, limiting the ability to learn generalizable and discriminative features incrementally from online training data. To address this, we introduce a plug-and-play module, S6MOD, which can be integrated into most existing methods and directly improve adaptability. Specifically, S6MOD introduces an extra branch after the backbone, where a mixture of discretization selectively adjusts parameters in a selective state space model, enriching selective scan patterns such that the model can adaptively select the most sensitive discretization method for current dynamics. We further design a class-conditional routing algorithm for dynamic, uncertainty-based adjustment and implement a contrastive discretization loss to optimize it. Extensive experiments combining our module with various models demonstrate that S6MOD significantly enhances model adaptability, leading to substantial performance gains and achieving the state-of-the-art results. The code is available at https://github.com/MyToumaKazusa/S6MOD.

Masked video modeling, such as VideoMAE, is an effective paradigm for video self-supervised learning (SSL). However, they are primarily based on reconstructing pixel-level details on natural videos which have substantial temporal redundancy, limiting their capability for semantic representation and sufficient encoding of motion dynamics. To address these issues, this paper introduces a novel SSL approach for video representation learning, dubbed as SMILE, by infusing both spatial and motion semantics. In SMILE, we leverage image-language pretrained models, such as CLIP, to guide the learning process with their high-level spatial semantics. We enhance the representation of motion by introducing synthetic motion patterns in the training data, allowing the model to capture more complex and dynamic content. Furthermore, using SMILE, we establish a new self-supervised video learning paradigm capable of learning strong video representations without requiring any natural video data. We have carried out extensive experiments on 7 datasets with various downstream scenarios. SMILE surpasses current state-of-the-art SSL methods, showcasing its effectiveness in learning more discriminative and generalizable video representations. Code is available: https://github.com/fmthoker/SMILE

Facial expression datasets remain limited in scale due to the subjectivity of annotations and the labor-intensive nature of data collection. This limitation poses a significant challenge for developing modern deep learning-based facial expression analysis models, particularly foundation models, that rely on large-scale data for optimal performance. To tackle the overarching and complex challenge, instead of introducing a new large-scale dataset, we introduce SynFER (Synthesis of Facial Expressions with Refined Control), a novel synthetic framework for synthesizing facial expression image data based on high-level textual descriptions as well as more fine-grained and precise control through facial action units. To ensure the quality and reliability of the synthetic data, we propose a semantic guidance technique to steer the generation process and a pseudo-label generator to help rectify the facial expression labels for the synthetic images. To demonstrate the generation fidelity and the effectiveness of the synthetic data from SynFER, we conduct extensive experiments on representation learning using both synthetic data and real-world data. Results validate the efficacy of our approach and the synthetic data. Notably, our approach achieves a 67.23% classification accuracy on AffectNet when training solely with synthetic data equivalent to the AffectNet training set size, which increases to 69.84% when scaling up to five times the original size. Code is available here.

Match-cuts are powerful cinematic tools that create seamless transitions between scenes, delivering strong visual and metaphorical connections. However, crafting match-cuts is a challenging, resource-intensive process requiring deliberate artistic planning. In MatchDiffusion, we present the first training-free method for match-cut generation using text-to-video diffusion models. MatchDiffusion leverages a key property of diffusion models: early denoising steps define the scene's broad structure, while later steps add details. Guided by this insight, MatchDiffusion employs "Joint Diffusion" to initialize generation for two prompts from shared noise, aligning structure and motion. It then applies "Disjoint Diffusion", allowing the videos to diverge and introduce unique details. This approach produces visually coherent videos suited for match-cuts. User studies and metrics demonstrate MatchDiffusion's effectiveness and potential to democratize match-cut creation.

Neural Radiance Field (NeRF)-based segmentation methods focus on object semantics and rely solely on RGB data, lacking intrinsic material properties. This limitation restricts accurate material perception, which is crucial for robotics, augmented reality, simulation, and other applications. We introduce UnMix-NeRF, a framework that integrates spectral unmixing into NeRF, enabling joint hyperspectral novel view synthesis and unsupervised material segmentation. Our method models spectral reflectance via diffuse and specular components, where a learned dictionary of global endmembers represents pure material signatures, and per-point abundances capture their distribution. For material segmentation, we use spectral signature predictions along learned endmembers, allowing unsupervised material clustering. Additionally, UnMix-NeRF enables scene editing by modifying learned endmember dictionaries for flexible material-based appearance manipulation. Extensive experiments validate our approach, demonstrating superior spectral reconstruction and material segmentation to existing methods. Project page: https://www.factral.co/UnMix-NeRF.

Recovering the 3D geometry of a scene from a sparse set of uncalibrated images is a long-standing problem in computer vision. While recent learning-based approaches such as DUSt3R and MASt3R have demonstrated impressive results by directly predicting dense scene geometry, they are primarily trained on outdoor scenes with static environments and struggle to handle human-centric scenarios. In this work, we introduce HAMSt3R, an extension of MASt3R for joint human and scene 3D reconstruction from sparse, uncalibrated multi-view images. First, we exploit DUNE, a strong image encoder obtained by distilling, among others, the encoders from MASt3R and from a state-of-the-art Human Mesh Recovery (HMR) model, multi-HMR, for a better understanding of scene geometry and human bodies. Our method then incorporates additional network heads to segment people, estimate dense correspondences via DensePose, and predict depth in human-centric environments, enabling a more comprehensive 3D reconstruction. By leveraging the outputs of our different heads, HAMSt3R produces a dense point map enriched with human semantic information in 3D. Unlike existing methods that rely on complex optimization pipelines, our approach is fully feed-forward and efficient, making it suitable for real-world applications. We evaluate our model on EgoHumans and EgoExo4D, two challenging benchmarks con taining diverse human-centric scenarios. Additionally, we validate its generalization to traditional multi-view stereo and multi-view pose regression tasks. Our results demonstrate that our method can reconstruct humans effectively while preserving strong performance in general 3D reconstruction tasks, bridging the gap between human and scene understanding in 3D vision.

Several video understanding tasks, such as natural language temporal video grounding, temporal activity localization, and audio description generation, require "temporally dense" reasoning over frames sampled at high temporal resolution. However, computing frame-level features for these tasks is computationally expensive given the temporal resolution requirements. In this paper, we make three contributions to reduce the cost of computing features for temporally dense tasks. First, we introduce a vision transformer (ViT) architecture, dubbed ResidualViT, that leverages the large temporal redundancy in videos to efficiently compute temporally dense frame-level features. Our architecture incorporates (i) learnable residual connections that ensure temporal consistency across consecutive frames and (ii) a token reduction module that enhances processing speed by selectively discarding temporally redundant information while reusing weights of a pretrained foundation model. Second, we propose a lightweight distillation strategy to approximate the frame-level features of the original foundation model. Finally, we evaluate our approach across four tasks and five datasets, in both zero-shot and fully supervised settings, demonstrating significant reductions in computational cost (up to 60%) and improvements in inference speed (up to 2.5x faster), all while closely approximating the accuracy of the original foundation model.

Bimanual manipulation is crucial in robotics, enabling complex tasks in industrial automation and household services. However, it poses significant challenges due to the high-dimensional action space and intricate coordination requirements. While video prediction has been recently studied for representation learning and control, leveraging its ability to capture rich dynamic and behavioral information, its potential for enhancing bimanual coordination remains underexplored. To bridge this gap, we propose a unified diffusion-based framework for the joint optimization of video and action prediction. Specifically, we propose a multi-frame latent prediction strategy that encodes future states in a compressed latent space, preserving task-relevant features. Furthermore, we introduce a unidirectional attention mechanism where video prediction is conditioned on the action, while action prediction remains independent of video prediction. This design allows us to omit video prediction during inference, significantly enhancing efficiency. Experiments on two simulated benchmarks and a real-world setting demonstrate a significant improvement in the success rate over the strong baseline ACT using our method, achieving a \textbf{24.9\%} increase on ALOHA, an \textbf{11.1\%} increase on RoboTwin, and a \textbf{32.5\%} increase in real-world experiments. Our models and code are publicly available at https://github.com/return-sleep/Diffusion_based_imaginative_Coordination.

Understanding videos that contain multiple modalities is crucial, especially in egocentric videos, where combining various sensory inputs significantly improves tasks like action recognition and moment localization. However, real-world applications often face challenges with incomplete modalities due to privacy concerns, efficiency needs, or hardware issues. Current methods, while effective, often necessitate retraining the model entirely to handle missing modalities, making them computationally intensive, particularly with large training datasets. In this study, we propose a novel approach to address this issue at test time without requiring retraining. We frame the problem as a test-time adaptation task, where the model adjusts to the available unlabeled data at test time. Our method, MiDl~(Mutual information with self-Distillation), encourages the model to be insensitive to the specific modality source present during testing by minimizing the mutual information between the prediction and the available modality. Additionally, we incorporate self-distillation to maintain the model's original performance when both modalities are available. MiDl represents the first self-supervised, online solution for handling missing modalities exclusively at test time. Through experiments with various pretrained models and datasets, MiDl demonstrates substantial performance improvement without the need for retraining.

Recent segmentation methods, which adopt large-scale data training and transformer architecture, aim to create one foundation model that can perform multiple tasks. However, most of these methods rely on heavy encoder and decoder frameworks, hindering their performance in real-time scenarios. To explore real-time segmentation, recent advancements primarily focus on semantic segmentation within specific environments, such as autonomous driving. However, they often overlook the generalization ability of these models across diverse scenarios. Therefore, to fill this gap, this work explores a novel real-time segmentation setting called real-time multi-purpose segmentation. It contains three fundamental sub-tasks: interactive segmentation, panoptic segmentation, and video instance segmentation. Unlike previous methods, which use a specific design for each task, we aim to use only a single end-to-end model to accomplish all these tasks in real-time. To meet real-time requirements and balance multi-task learning, we present a novel dynamic convolution-based method, Real-Time Multi-Purpose SAM (RMP-SAM). It contains an efficient encoder and an efficient decoupled adapter to perform prompt-driven decoding. Moreover, we further explore different training strategies and one new adapter design to boost co-training performance further. We benchmark several strong baselines by extending existing works to support our multi-purpose segmentation. Extensive experiments demonstrate that RMP-SAM is effective and generalizes well on proposed benchmarks and other specific semantic tasks. Our implementation of RMP-SAM achieves the optimal balance between accuracy and speed for these tasks. The code is released at \url{https://github.com/xushilin1/RAP-SAM}

We propose a novel approach for disentangling visual and semantic features from the backbones of pre-trained diffusion models, enabling visual correspondence in a manner analogous to the well-established semantic correspondence. While diffusion model backbones are known to encode semantically rich features, they must also contain visual features to support their image synthesis capabilities. However, isolating these visual features is challenging due to the absence of annotated datasets. To address this, we introduce an automated pipeline that constructs image pairs with annotated semantic and visual correspondences based on existing subject-driven image generation datasets, and design a contrastive architecture to separate the two feature types. Leveraging the disentangled representations, we propose a new metric, Visual Semantic Matching (VSM), that quantifies visual inconsistencies in subject-driven image generation. Empirical results show that our approach outperforms global feature-based metrics such as CLIP, DINO, and vision--language models in quantifying visual inconsistencies while also enabling spatial localization of inconsistent regions. To our knowledge, this is the first method that supports both quantification and localization of inconsistencies in subject-driven generation, offering a valuable tool for advancing this task.

Large Language Model (LLM)-based multi-agent systems show promise for automating real-world tasks but struggle to transfer across domains due to their domain-specific nature. Current approaches face two critical shortcomings: they require complete architectural redesign and full retraining of all components when applied to new domains. We introduce **Workforce**, a hierarchical multi-agent framework that decouples strategic planning from specialized execution through a modular architecture comprising: *(i)* a *domain-agnostic* **Planner** for task decomposition, *(ii)* a **Coordinator** for subtask management, and *(iii)* specialized **Workers** with *domain-specific* tool-calling capabilities. This decoupling enables cross-domain transferability during both inference and training phases: During inference, Workforce seamlessly adapts to new domains by adding or modifying worker agents; For training, we introduce **Optimized Workforce Learning (OWL)**, which improves generalization across domains by optimizing a domain-agnostic planner with reinforcement learning from real-world feedback. To validate our approach, we evaluate Workforce on the GAIA benchmark, covering various realistic, multi-domain agentic tasks. Experimental results demonstrate Workforce achieves open-source state-of-the-art performance (**69.70%**), outperforming commercial systems like OpenAI's Deep Research by **2.34%**. More notably, our OWL-trained 32B model achieves **52.73%** accuracy (**+16.37%**) and demonstrates performance comparable to GPT-4o on challenging tasks. To summarize, by enabling scalable generalization and modular domain transfer, our work establishes a foundation for the next generation of general-purpose AI assistants. *Our code is available at [Anonymous URL](https://anonymous.4open.science/r/annonymous-owl/), and our data is available at [Anonymous URL](https://huggingface.co/anonymous21016).*

In this paper, we introduce Online Multimodal Conversational Response Generation (OMCRG), a novel task designed to produce synchronized verbal and non-verbal listener feedback online, based on the speaker's multimodal inputs. OMCRG captures natural dyadic interactions and introduces new challenges in aligning generated audio with listeners' facial responses. To tackle these challenges, we incorporate text as an intermediate modality to connect audio and facial responses. We propose OmniResponse, a Multimodal Large Language Model (MLLM) that autoregressively generates accurate multimodal listener responses. OmniResponse leverages a pretrained LLM enhanced with two core components: Chrono-Text Markup, which precisely timestamps generated text tokens, and TempoVoice, a controllable online text-to-speech (TTS) module that outputs speech synchronized with facial responses. To advance OMCRG research, we offer ResponseNet, a dataset of 696 detailed dyadic interactions featuring synchronized split-screen videos, multichannel audio, transcripts, and annotated facial behaviors. Comprehensive evaluations on ResponseNet demonstrate that OmniResponse outperforms baseline models in terms of semantic speech content, audio-visual synchronization, and generation quality. Our dataset, code, and models are publicly available at https://omniresponse.github.io/.

Inspired by Minsky's Society of Mind, Schmidhuber's Learning to Think, and other more recent works, this paper proposes and advocates for the concept of natural language-based societies of mind (NLSOMs). We imagine these societies as consisting of a collection of multimodal neural networks, including large language models, which engage in a “mindstorm” to solve problems using a shared natural language interface. Here, we work to identify and discuss key questions about the social structure, governance, and economic principles for NLSOMs, emphasizing their impact on the future of AI. Our demonstrations with NLSOMs-which feature up to 129 agents-show their effectiveness in various tasks, including visual question answering, image captioning, and prompt generation for text-to-image synthesis.

We present Ego-Exo4D, a diverse, large-scale multimodal multiview video dataset and benchmark challenge. Ego-Exo4D centers around simultaneously-captured egocentric and exocentric video of skilled human activities (e.g., sports, music, dance, bike repair). 740 participants from 13 cities worldwide performed these activities in 123 different natural scene contexts, yielding long-form captures from 1 to 42 minutes each and 1,286 hours of video combined. The multimodal nature of the dataset is unprecedented: the video is accompanied by multichannel audio, eye gaze, 3D point clouds, camera poses, IMU, and multiple paired language descriptions—including a novel “expert commentary” done by coaches and teachers and tailored to the skilled-activity domain. To push the frontier of first-person video understanding of skilled human activity, we also present a suite of benchmark tasks and their annotations, including fine-grained activity understanding, proficiency estimation, cross-view translation, and 3D hand/body pose. All resources are open sourced to fuel new research in the community. https://ego-exo4d-data.org/

We introduce Ego4D, a massive-scale egocentric video dataset and benchmark suite. It offers 3,670 hours of daily-life activity video spanning hundreds of scenarios (household, outdoor, workplace, leisure, etc.) captured by 931 unique camera wearers from 74 worldwide locations and 9 different countries. The approach to collection is designed to uphold rigorous privacy and ethics standards, with consenting participants and robust de-identification procedures where relevant. Ego4D dramatically expands the volume of diverse egocentric video footage publicly available to the research community. Portions of the video are accompanied by audio, 3D meshes of the environment, eye gaze, stereo, and/or synchronized videos from multiple egocentric cameras at the same event. Furthermore, we present a host of new benchmark challenges centered around understanding the first-person visual experience in the past (querying an episodic memory), present (analyzing hand-object manipulation, audio-visual conversation, and social interactions), and future (forecasting activities). By publicly sharing this massive annotated dataset and benchmark suite, we aim to push the frontier of first-person perception.

While Transformers have achieved impressive success in natural language processing and computer vision, their performance on 3D point clouds is relatively poor. This is mainly due to the limitation of Transformers: a demanding need for extensive training data. Unfortunately, in the realm of 3D point clouds, the availability of large datasets is a challenge, exacerbating the issue of training Transformers for 3D tasks. In this work, we solve the data issue of point cloud Transformers from two perspectives: (i) introducing more inductive bias to reduce the dependency of Transformers on data, and (ii) relying on cross-modality pretraining. More specifically, we first present Progressive Point Patch Embedding and present a new point cloud Transformer model namely PViT. PViT shares the same backbone as Transformer but is shown to be less hungry for data, enabling Transformer to achieve performance comparable to the state-of-the-art. Second, we formulate a simple yet effective pipeline dubbed Pix4Point that allows harnessing Transformers pretrained in the image domain to enhance downstream point cloud understanding. This is achieved through a modality-agnostic Transformer backbone with the help of a tokenizer and decoder specialized in the different domains. Pretrained on a large number of widely available images, significant gains of PViT are observed in the tasks of 3D point cloud classification, part segmentation, and semantic segmentation on ScanObjectNN, ShapeNetPart, and S3DIS, respectively. Our code and models are available at https: //github.com/guochengqian/Pix4Point.

Continual Learning is a step towards lifelong intelligence where models continuously learn from recently collected data without forgetting previous knowledge. Existing continual learning approaches mostly focus on image classification in the class-incremental setup with clear task boundaries and unlimited computational budget. This work explores the problem of Online Domain-Incremental Continual Segmentation (ODICS), where the model is continually trained over batches of densely labeled images from different domains, with limited computation and no information about the task boundaries. ODICS arises in many practical applications. In autonomous driving, this may correspond to the realistic scenario of training a segmentation model over time on a sequence of cities. We analyze several existing continual learning methods and show that they perform poorly in this setting despite working well in class-incremental segmentation. We propose SimCS, a parameter-free method complementary to existing ones that uses simulated data to regularize continual learning. Experiments show that SimCS provides consistent improvements when combined with different CL methods.

Large pretrained models are increasingly crucial in modern computer vision tasks. These models are typically used in downstream tasks by end-to-end finetuning, which is highly memory-intensive for tasks with high-resolution data, e.g., video understanding, small object detection, and point cloud analysis. In this paper, we propose Dynamic Reversible Dual-Residual Networks, or Dr2Net, a novel family of network architectures that acts as a surrogate net-work to finetune a pretrained model with substantially re-duced memory consumption. Dr2 Net contains two types of residual connections, one maintaining the residual struc-ture in the pretrained models, and the other making the network reversible. Due to its reversibility, intermediate activations, which can be reconstructed from output, are cleared from memory during training. We use two co-efficients on either type of residual connections respec-tively, and introduce a dynamic training strategy that seam-lessly transitions the pretrained model to a reversible net-work with much higher numerical precision. We evaluate Dr2Net on various pretrained models and various tasks, and show that it can reach comparable performance to con-ventional finetuning but with significantly less memory us-age. Code will be available at https://github.com/coolbay/Dr2Net.

Movie trailers are an essential tool for promoting films and attracting audiences. However, the process of creating trailers can be time-consuming and expensive. To stream-line this process, we propose an automatic trailer generation framework that generates plausible trailers from a full movie by automating shot selection and composition. Our approach draws inspiration from machine translation techniques and models the movies and trailers as sequences of shots, thus formulating the trailer generation problem as a sequence-to-sequence task. We introduce Trailer Generation Transformer (TGT), a deep-learning framework utilizing an encoder-decoder architecture. TGT movie encoder is tasked with contextualizing each movie shot representation via self-attention, while the autoregressive trailer de-coder predicts the feature representation of the next trailer shot, accounting for the relevance of shots' temporal order in trailers. Our TGT significantly outperforms previous methods on a comprehensive suite of metrics.

We present Ego-Exo4D, a diverse, large-scale multi-modal multiview video dataset and benchmark challenge. Ego-Exo4D centers around simultaneously-captured ego-centric and exocentric video of skilled human activities (e.g., sports, music, dance, bike repair). 740 participants from 13 cities worldwide performed these activities in 123 different natural scene contexts, yielding long-form captures from 1 to 42 minutes each and 1,286 hours of video combined. The multimodal nature of the dataset is un-precedented: the video is accompanied by multichannel audio, eye gaze, 3D point clouds, camera poses, IMU, and multiple paired language descriptions-including a novel “expert commentary” done by coaches and teachers and tailored to the skilled-activity domain. To push the frontier of first-person video understanding of skilled human activity, we also present a suite of benchmark tasks and their annotations, including fine-grained activity understanding, proficiency estimation, cross-view translation, and 3D hand/body pose. All resources are open sourced to fuel new research in the community.

Advancements in 3D Gaussian Splatting have significantly accelerated 3D reconstruction and generation. However, it may require a large number of Gaussians, which creates a substantial memory footprint. This paper introduces GES (Generalized Exponential Splatting), a novel representation that employs Generalized Exponential Function (GEF) to model 3D scenes, requiring far fewer parti-cles to represent a scene and thus significantly outperforming Gaussian Splatting methods in efficiency with a plug-and-play replacement ability for Gaussian-based utilities. GES is validated theoretically and empirically in both principled 1D setup and realistic 3D scenes. It is shown to represent signals with sharp edges more accurately, which are typically challenging for Gaussians due to their inherent low-pass characteristics. Our empirical analysis demonstrates that GEF outperforms Gaussians in fitting natural-occurring signals (e.g. squares, triangles, parabolic signals), thereby reducing the need for extensive splitting operations that increase the memory footprint of Gaussian Splatting. With the aid of a frequency-modulated loss, GES achieves competitive performance in novel-view syn-thesis benchmarks while requiring less than half the memory storage of Gaussian Splatting and increasing the rendering speed by up to 39%. The code is available on the project website https://abdullahamdi.com/ges.

We present SPAD, a novel approach for creating con-sistent multi-view images from text prompts or single images. To enable multi-view generation, we repurpose a pre-trained 2D diffusion model by extending its self-attention layers with cross-view interactions, and fine-tune it on a high quality subset of Objaverse. We find that a naive extension of the self-attention proposed in prior work (e.g., MV-Dream) leads to content copying between views. Therefore, we explicitly constrain the cross-view attention based on epipolar geometry. To further enhance 3D consistency, we utilize Plücker coordinates derived from camera rays and inject them as positional encoding. This enables SPAD to reason over spatial proximity in 3D well. Compared to concurrent works that can only generate views at fixed azimuth and elevation (e.g., MVDream, SyncDreamer), SPAD offers full camera control and achieves state-of-the-art results in novel view synthesis on unseen objects from the Objaverse and Google Scanned Objects datasets. Finally, we demon-strate that text-to-3D generation using SPAD prevents the multi-face Janus issue.

Recently, temporal action detection (TAD) has seen significant performance improvement with end-to-end training. However, due to the memory bottleneck, only models with limited scales and limited data volumes can afford end-to-end training, which inevitably restricts TAD performance. In this paper, we reduce the memory consumption for end-to-end training, and manage to scale up the TAD backbone to 1 billion parameters and the input video to 1,536 frames, leading to significant detection performance. The key to our approach lies in our proposed temporal-informative adapter (TIA), which is a novel lightweight module that reduces training memory. Using TIA, we free the humongous backbone from learning to adapt to the TAD task by only updating the parameters in TIA. TIA also leads to better TAD representation by temporally aggregating context from adjacent frames throughout the backbone. We evaluate our model across four representative datasets. Owing to our efficient design, we are able to train end-to-end on VideoMAEv2-giant and achieve 75.4% mAP on THUMOS14, being the first end-to-end model to outper-form the best feature-based methods. Code is available at https://github.com/sming256/AdaTAD.

The modern surge in camera usage alongside widespread computer vision technology applications poses significant privacy and security concerns. Current artificial intelligence (AI) technologies aid in recognizing relevant events and assisting in daily tasks in homes, offices, hospitals, etc. The need to access or process personal information for these purposes raises privacy concerns. While software-level solutions like face de-identification provide a good privacy/utility tradeo-ff, they present vulnerabilities to sniffing attacks. In this paper, we propose a hardware-level face de-identification method to solve this vulnerability. Specifically, our approach first learns an optical encoder along with a regression model to obtain a face heatmap while hiding the face identity from the source image. We also propose an anonymization framework that generates a new face using the privacy-preserving image, face heatmap, and a reference face image from a public dataset as input. We validate our approach with extensive simulations and hardware experiments.

Visual prompting of large vision language models such as CLIP exhibits intriguing zero-shot capabilities. A manually drawn red circle, commonly used for highlighting, can guide CLIP's attention to the surrounding region, to identify specific objects within an image. Without precise object proposals, however, it is insufficient for localization. Our novel, simple yet effective approach, i.e., Differentiable Visual Prompting, enables CLIP to zero-shot localize: given an image and a text prompt describing an object, we first pick a rendered ellipse from uniformly distributed anchor ellipses on the image grid via visual prompting, then use three loss functions to tune the ellipse coefficients to encap-sulate the target region gradually. This yields promising ex-perimental results for referring expression comprehension without precisely specified object proposals. In addition, we systematically present the limitations of visual prompting inherent in CLIP and discuss potential solutions.

Semantic segmentation is crucial in remote sensing, where high-resolution satellite images are segmented into meaningful regions. Recent advancements in deep learning have significantly improved satellite image segmentation. However, most of these methods are typically trained in fully supervised settings that require high-quality pixel-level annotations, which are expensive and time-consuming to obtain. In this work, we present a weakly supervised learning algorithm to train semantic segmentation algorithms that only rely on query point annotations instead of full mask labels. Our proposed approach performs accurate semantic segmentation and improves efficiency by significantly reducing the cost and time required for manual annotation. Specifically, we generate superpixels and extend the query point labels into those superpixels that group similar meaningful semantics. Then, we train semantic segmentation models supervised with images partially labeled with the superpixel pseudo-labels. We benchmark our weakly supervised training approach on an aerial image dataset and different semantic segmentation architectures, showing that we can reach competitive performance compared to fully supervised training while reducing the annotation effort. The code of our proposed approach is publicly available at: https://github.com/santiago2205/LSSQPS.

Chronic wounds pose an ongoing health concern globally, largely due to the prevalence of conditions such as diabetes and leprosy’s disease. The standard method of monitoring these wounds involves visual inspection by healthcare professionals, a practice that could present challenges for patients in remote areas with inadequate transportation and healthcare infrastructure. This has led to the development of algorithms designed for the analysis and follow-up of wound images, which perform image-processing tasks such as classification, detection, and segmentation. However, the effectiveness of these algorithms heavily depends on the availability of comprehensive and varied wound image data, which is usually scarce. This paper introduces the CO2Wounds-V2 dataset, an extended collection of RGB wound images from leprosy patients with their corresponding semantic segmentation annotations, aiming to enhance the development and testing of image-processing algorithms in the medical field.

We present ``Magic123'', a two-stage coarse-to-fine approach for high-quality, textured 3D mesh generation from a single image in the wild using *both 2D and 3D priors*. In the first stage, we optimize a neural radiance field to produce a coarse geometry. In the second stage, we adopt a memory-efficient differentiable mesh representation to yield a high-resolution mesh with a visually appealing texture. In both stages, the 3D content is learned through reference-view supervision and novel-view guidance by a joint 2D and 3D diffusion prior. We introduce a trade-off parameter between the 2D and 3D priors to control the details and 3D consistencies of the generation. Magic123 demonstrates a significant improvement over previous image-to-3D techniques, as validated through extensive experiments on diverse synthetic and real-world images.

Video activity localization aims at understanding the semantic content in long, untrimmed videos and retrieving actions of interest. The retrieved action with its start and end locations can be used for highlight generation, temporal action detection, etc. Unfortunately, learning the exact boundary location of activities is highly challenging because temporal activities are continuous in time, and there are often no clear-cut transitions between actions. Moreover, the definition of the start and end of events is subjective, which may confuse the model. To alleviate the boundary ambiguity, we propose to study the video activity localization problem from a denoising perspective. Specifically, we propose an encoder-decoder model named DenosieLoc. During training, a set of temporal spans is randomly generated from the ground truth with a controlled noise scale. Then, we attempt to reverse this process by boundary denoising, allowing the localizer to predict activities with precise boundaries and resulting in faster convergence speed. Experiments show that DenosieLoc advances several video activity understanding tasks. For example, we observe a gain of +12.36% average mAP on the QV-Highlights dataset. Moreover, DenosieLoc achieves state-of-the-art performance on the MAD dataset but with much fewer predictions than others.

We propose and study a realistic Continual Learning (CL) setting where learning algorithms are granted a restricted computational budget per time step while training. We apply this setting to large-scale semi-supervised Continual Learning scenarios with sparse label rate. Previous proficient CL methods perform very poorly in this challenging setting. Overfitting to the sparse labeled data and insufficient computational budget are the two main culprits for such a poor performance. Our new setting encourages learning methods to effectively and efficiently utilize the unlabeled data during training. To that end, we propose a simple but highly effective baseline, DietCL, which utilizes both unlabeled and labeled data jointly. DietCL meticulously allocates computational budget for both types of data. We validate our baseline, at scale, on several datasets, e.g., CLOC, ImageNet10K, and CGLM, under constraint budget setup. DietCL outperforms, by a large margin, all existing supervised CL algorithms as well as more recent continual semi-supervised methods. Our extensive analysis and ablations demonstrate that DietCL is stable under a full spectrum of label sparsity, computational budget and various other ablations.

This paper proposes a novel online evaluation protocol for Test Time Adaptation (TTA) methods, which penalizes slower methods by providing them with fewer samples for adaptation. TTA methods leverage unlabeled data at test time to adapt to distribution shifts. Though many effective methods have been proposed, their impressive performance usually comes at the cost of significantly increased computation budgets. Current evaluation protocols overlook the effect of this extra computation cost, affecting their real-world applicability. To address this issue, we propose a more realistic evaluation protocol for TTA methods, where data is received in an online fashion from a constant-speed data stream, thereby accounting for the method’s adaptation speed. We apply our proposed protocol to benchmark several TTA methods on multiple datasets and scenarios. Extensive experiments shows that, when accounting for inference speed, simple and fast approaches can outperform more sophisticated but slower methods. For example, SHOT from 2020, outperforms the state-of-the-art method SAR from 2023 under our online setting. Our results reveal the importance of developing practical TTA methods that are both accurate and efficient.

Relieving the reliance of neural network training on a global back-propagation (BP) has emerged as a notable research topic due to the biological implausibility and huge memory consumption caused by BP. Among the existing solutions, local learning optimizes gradient-isolated modules of a neural network with local errors and has been proved to be effective even on large-scale datasets. However, the reconciliation among local errors has never been investigated. In this paper, we first theoretically study non-greedy layer-wise training and show that the convergence cannot be assured when the local gradient in a module w.r.t. its input is not reconciled with the local gradient in the previous module w.r.t. its output. Inspired by the theoretical result, we further propose a local training strategy that successively regularizes the gradient reconciliation between neighboring modules without breaking gradient isolation or introducing any learnable parameters. Our method can be integrated into both local-BP and BP-free settings. In experiments, we achieve significant performance improvements compared to previous methods. Particularly, our method for CNN and Transformer architectures on ImageNet is able to attain a competitive performance with global BP, saving more than 40% memory consumption.

While diffusion models have shown impressive performance in 2D image/video generation, diffusion-based Text-to-Multi-view-Video (T2MVid) generation remains underexplored. The new challenges posed by T2MVid generation lie in the lack of massive captioned multi-view videos and the complexity of modeling such multi-dimensional distribution. To this end, we propose a novel diffusion-based pipeline that generates high-quality multi-view videos centered around a dynamic 3D object from text. Specifically, we factor the T2MVid problem into viewpoint-space and time components. Such factorization allows us to combine and reuse layers of advanced pre-trained multi-view image and 2D video diffusion models to ensure multi-view consistency as well as temporal coherence for the generated multi-view videos, largely reducing the training cost. We further introduce alignment modules to align the latent spaces of layers from the pre-trained multi-view and the 2D video diffusion models, addressing the reused layers' incompatibility that arises from the domain gap between 2D and multi-view data. In support of this and future research, we further contribute a captioned multi-view video dataset. Experimental results demonstrate that our method generates high-quality multi-view videos, exhibiting vivid motions, temporal coherence, and multi-view consistency, given a variety of text prompts.

Large Language Model (LLM) agents have been increasingly adopted as simulation tools to model humans in social science and role-playing applications. However, one fundamental question remains: can LLM agents really simulate human behavior? In this paper, we focus on one critical and elemental behavior in human interactions, trust, and investigate whether LLM agents can simulate human trust behavior. We first find that LLM agents generally exhibit trust behavior, referred to as agent trust, under the framework of Trust Games, which are widely recognized in behavioral economics. Then, we discover that GPT-4 agents manifest high behavioral alignment with humans in terms of trust behavior, indicating the feasibility of simulating human trust behavior with LLM agents. In addition, we probe the biases of agent trust and differences in agent trust towards other LLM agents and humans. We also explore the intrinsic properties of agent trust under conditions including external manipulations and advanced reasoning strategies. Our study provides new insights into the behaviors of LLM agents and the fundamental analogy between LLMs and humans beyond value alignment. We further illustrate broader implications of our discoveries for applications where trust is paramount.

Current parameter-efficient fine-tuning (PEFT) methods build adapters widely agnostic of the context of downstream task to learn, or the context of important knowledge to maintain. As a result, there is often a performance gap compared to full-parameter fine-tuning, and meanwhile the fine-tuned model suffers from catastrophic forgetting of the pre-trained world knowledge. In this paper, we propose **CorDA**, a Context-oriented Decomposition Adaptation method that builds learnable **task-aware adapters** from weight decomposition oriented by the context of downstream task or the world knowledge to maintain. Concretely, we collect a few data samples, and perform singular value decomposition for each linear layer of a pre-trained LLM multiplied by the covariance matrix of the input activation using these samples. The inverse of the covariance matrix is multiplied with the decomposed components to reconstruct the original weights. By doing so, the context of the representative samples is captured through deciding the factorizing orientation. Our method enables two options, the **knowledge-preserved adaptation** and the **instruction-previewed adaptation**. For the former, we use question-answering samples to obtain the covariance matrices, and use the decomposed components with the smallest $r$ singular values to initialize a learnable adapter, with the others frozen such that the world knowledge is better preserved. For the latter, we use the instruction data from the fine-tuning task, such as math or coding, to orientate the decomposition and train the largest $r$ components that most correspond to the task to learn. We conduct extensive experiments on Math, Code, and Instruction Following tasks. Our knowledge-preserved adaptation not only achieves better performance than LoRA on fine-tuning tasks, but also mitigates the forgetting of world knowledge. Our instruction-previewed adaptation is able to further enhance the fine-tuning performance to be comparable with full fine-tuning, surpassing the state-of-the-art PEFT methods such as LoRA, DoRA, and PiSSA.

We present a novel approach for digitizing real-world objects by estimating their geometry, material properties, and environmental lighting from a set of posed images with fixed lighting. Our method incorporates into Neural Radiance Field (NeRF) pipelines the split sum approximation used with image-based lighting for real-time physically based rendering. We propose modeling the scene's lighting with a single scene-specific MLP representing pre-integrated image-based lighting at arbitrary resolutions. We accurately model pre-integrated lighting by exploiting a novel regularizer based on efficient Monte Carlo sampling. Additionally, we propose a new method of supervising self-occlusion predictions by exploiting a similar regularizer based on Monte Carlo sampling. Experimental results demonstrate the efficiency and effectiveness of our approach in estimating scene geometry, material properties, and lighting. Our method attains state-of-the-art relighting quality after only ${\sim}1$ hour of training in a single NVIDIA A100 GPU.

Manually reading and logging gauge data is time-inefficient, and the effort increases according to the number of gauges available. We present a pipeline that automates the reading of analog gauges. We propose a two-stage CNN pipeline that identifies the key structural components of an analog gauge and outputs an angular reading. To facilitate the training of our approach, a synthetic dataset is generated thus obtaining a set of realistic analog gauges with their corresponding annotation. To validate our proposal, an additional real-world dataset was collected with 4.813 manually curated images. When compared against state-of-the-art methodologies, our method shows a significant improvement of 4.55◦in the average error, which is a 52% relative improvement. The resources for this project will be made available at: https://github.com/fuankarion/automatic-gauge-reading.

AR/VR applications promise to provide people with a genuine feeling of mutual presence when communicating via their personalized avatars. While realistic avatars are essential in various social settings, the vast possibilities of a virtual world can also generate interest in using stylized avatars for other purposes. We introduce StyleAvatar, the first method for semantic stylization of animatable head avatars. StyleAvatar directly stylizes the avatar representation, rather than stylizing its renders. Specifically, given a model generating the avatar, StyleAvatar first disentangles geometry and texture manipulations, and then stylizes the avatar by fine-tuning a subset of the model’s weights. Our method has multiple virtues, including the ability to describe styles using images or text, preserving the avatar’s animatable capacity, providing control over identity preservation, and disentangling texture and geometry modifications. Experiments have shown that our approach consistently works across skin tones, challenging hair styles, extreme views, and diverse facial expressions.1

Unmanned aerial vehicles (UAVs) are widely used for image acquisition in various applications, and object detection is a crucial task for UAV imagery analysis. However, training accurate object detectors requires a large amount of annotated data, which can be expensive and time-consuming. To address this issue, we propose an active learning framework for single-stage object detectors in UAV images. First, we introduce Diverse Uncertainty Aggregation (DUA), a novel uncertainty aggregation method that aims to select images with a more diverse variety of object classes with high uncertainties. Second, we address the problem of class imbalance by adjusting the uncertainty calculation based on the performance of each class. Third, we illustrate how reducing the number of images for labeling does not necessarily lead to a lower labeling cost. Evaluation of our approach on a common UAV dataset shows that we can perform similarly (within 0.02 0.5mAP) to using the whole dataset while using only 25% of the images and 32% of the labeled objects. It also outperforms Random Selection and some other aggregation methods. Evaluation on VOC2012 show also consistent results utilizing only 25% of the labeling cost to reach a performance within 0.1 0.5mAP of using the whole dataset. Our results suggest that our proposed active learning framework can effectively reduce the annotation cost while improving the performance of singlestage object detectors in UAV image settings. The code is available on: https://github.com/asmayamani/DUA

The training and inference of Graph Neural Networks (GNNs) are costly when scaling up to large-scale graphs. Graph Lottery Ticket (GLT) has presented the first attempt to accelerate GNN inference on large-scale graphs by jointly pruning the graph structure and the model weights. Though promising, GLT encounters robustness and generalization issues when deployed in real-world scenarios, which are also long-standing and critical problems in deep learning ideology. In real-world scenarios, the distribution of unseen test data is typically diverse. We attribute the failures on out-of-distribution (OOD) data to the incapability of discerning causal patterns, which remain stable amidst distribution shifts. In traditional spase graph learning, the model performance deteriorates dramatically as the graph/network sparsity exceeds a certain high level. Worse still, the pruned GNNs are hard to generalize to unseen graph data due to limited training set at hand. To tackle these issues, we propose the Resilient Graph Lottery Ticket (RGLT) to find more robust and generalizable GLT in GNNs. Concretely, we reactivate a fraction of weights/edges by instantaneous gradient information at each pruning point. After sufficient pruning, we conduct environmental interventions to extrapolate potential test distribution. Finally, we perform last several rounds of model averages to further improve generalization. We provide multiple examples and theoretical analyses that underpin the universality and reliability of our proposal. Further, RGLT has been experimentally verified across various independent identically distributed (IID) and out-of-distribution (OOD) graph benchmarks.

In the field of visual scene understanding, deep neural networks have made impressive advancements in various core tasks like segmentation, tracking, and detection. However, most approaches operate on the close-set assumption, meaning that the model can only identify pre-defined categories that are present in the training set. Recently, open vocabulary settings were proposed due to the rapid progress of vision language pre-training. These new approaches seek to locate and recognize categories beyond the annotated label space. The open vocabulary approach is more general, practical, and effective than weakly supervised and zero-shot settings. This paper thoroughly reviews open vocabulary learning, summarizing and analyzing recent developments in the field. In particular, we begin by juxtaposing open vocabulary learning with analogous concepts such as zero-shot learning, open-set recognition, and out-of-distribution detection. Subsequently, we examine several pertinent tasks within the realms of segmentation and detection, encompassing long-tail problems, few-shot, and zero-shot settings. As a foundation for our method survey, we first elucidate the fundamental principles of detection and segmentation in close-set scenarios. Next, we examine various contexts where open vocabulary learning is employed, pinpointing recurring design elements and central themes. This is followed by a comparative analysis of recent detection and segmentation methodologies in commonly used datasets and benchmarks. Our review culminates with a synthesis of insights, challenges, and discourse on prospective research trajectories. To our knowledge, this constitutes the inaugural exhaustive literature review on open vocabulary learning.

Generative Adversarial Networks (GANs) have shown compelling results in various tasks and applications in recent years. However, mode collapse remains a critical problem in GANs. In this paper, we propose a novel training pipeline to address the mode collapse issue of GANs. Different from existing methods, we propose to generalize the discriminator as feature embedding and maximize the entropy of distributions in the embedding space learned by the discriminator. Specifically, two regularization terms, i.e., Deep Local Linear Embedding (DLLE) and Deep Isometric feature Mapping (DIsoMap), are introduced to encourage the discriminator to learn the structural information embedded in the data, such that the embedding space learned by the discriminator can be well-formed. Based on the well-learned embedding space supported by the discriminator, a non-parametric entropy estimator is designed to efficiently maximize the entropy of embedding vectors, playing as an approximation of maximizing the entropy of the generated distribution. By improving the discriminator and maximizing the distance of the most similar samples in the embedding space, our pipeline effectively reduces the mode collapse without sacrificing the quality of generated samples. Extensive experimental results show the effectiveness of our method which outperforms the GAN baseline, MaF-GAN on CelebA (9.13 vs. 12.43 in FID) and surpasses the recent state-of-the-art energy-based model on the ANIMEFACE dataset (2.80 vs. 2.26 in Inception score).

Temporal action localization (TAL) requires long-form reasoning to predict actions of various durations and complex content. Given limited GPU memory, training TAL end to end (i.e., from videos to predictions) on long videos is a significant challenge. Most methods can only train on pre-extracted features without optimizing them for the localization problem, consequently limiting localization performance. In this work, to extend the potential in TAL networks, we propose a novel end-to-end method Re2TAL, which rewires pretrained video backbones for reversible TAL. Re2TAL builds a backbone with reversible modules, where the input can be recovered from the output such that the bulky intermediate activations can be cleared from memory during training. Instead of designing one single type of reversible module, we propose a network rewiring mechanism, to transform any module with a residual connection to a reversible module without changing any parameters. This provides two benefits: (1) a large variety of reversible networks are easily obtained from existing and even future model designs, and (2) the reversible models require much less training effort as they reuse the pre-trained parameters of their original non-reversible versions. Re2TAL, only using the RGB modality, reaches 37.01% average mAP on ActivityNet-v1.3, a new state-of-the-art record, and mAP 64.9% at tIoU=0.5 on THUMOS-14, outperforming all other RGB-only methods. Code is available at https://github.com/coolbay/Re2TAL.

Current evaluations of Continual Learning (CL) methods typically assume that there is no constraint on training time and computation. This is an unrealistic assumption for any real-world setting, which motivates us to propose: a practical real-time evaluation of continual learning, in which the stream does not wait for the model to complete training before revealing the next data for predictions. To do this, we evaluate current CL methods with respect to their computational costs. We conduct extensive experiments on CLOC, a large-scale dataset containing 39 million time-stamped images with geolocation labels. We show that a simple baseline outperforms state-of-the-art CL methods under this evaluation, questioning the applicability of existing methods in realistic settings. In addition, we explore various CL components commonly used in the literature, including memory sampling strategies and regularization approaches. We find that all considered methods fail to be competitive against our simple baseline. This surprisingly suggests that the majority of existing CL literature is tailored to a specific class of streams that is not practical. We hope that the evaluation we provide will be the first step towards a paradigm shift to consider the computational cost in the development of online continual learning methods.

Due to the outstanding capability for data generation, Generative Adversarial Networks (GANs) have attracted considerable attention in unsupervised learning. However, training GANs is difficult, since the training distribution is dynamic for the discriminator, leading to unstable image representation. In this paper, we address the problem of training GANs from a novel perspective, i.e., robust image classification. Motivated by studies on robust image representation, we propose a simple yet effective module, namely AdaptiveMix, for GANs, which shrinks the regions of training data in the image representation space of the discriminator. Considering it is intractable to directly bound feature space, we propose to construct hard samples and narrow down the feature distance between hard and easy samples. The hard samples are constructed by mixing a pair of training images. We evaluate the effectiveness of our AdaptiveMix with widely-used and state-of-the-art GAN architectures. The evaluation results demonstrate that our AdaptiveMix can facilitate the training of GANs and effectively improve the image quality of generated samples. We also show that our AdaptiveMix can be further applied to image classification and Out-Of-Distribution (OOD) detection tasks, by equipping it with state-of-the-art methods. Extensive experiments on seven publicly available datasets show that our method effectively boosts the performance of baselines. The code is publicly available at https://github.com/WentianZhang-ML/AdaptiveMix.

Video steganography is the art of unobtrusively concealing secret data in a cover video and then recovering the secret data through a decoding protocol at the receiver end. Although several attempts have been made, most of them are limited to low-capacity and fixed steganography. To rectify these weaknesses, we propose a Large-capacity and Flexible Video Steganography Network (LF-VSN) in this paper. For large-capacity, we present a reversible pipeline to perform multiple videos hiding and recovering through a single invertible neural network (INN). Our method can hide/recover 7 secret videos in/from 1 cover video with promising performance. For flexibility, we propose a key-controllable scheme, enabling different receivers to recover particular secret videos from the same cover video through specific keys. Moreover, we further improve the flexibility by proposing a scalable strategy in multiple videos hiding, which can hide variable numbers of secret videos in a cover video with a single model and a single training session. Extensive experiments demonstrate that with the significant improvement of the video steganography performance, our proposed LF-VSN has high security, large hiding capacity, and flexibility. The source code is available at https://github.com/MC-E/LF-VSN.

Continual Learning (CL) aims to sequentially train models on streams of incoming data that vary in distribution by preserving previous knowledge while adapting to new data. Current CL literature focuses on restricted access to previously seen data, while imposing no constraints on the computational budget for training. This is unreasonable for applications in-the-wild, where systems are primarily constrained by computational and time budgets, not storage. We revisit this problem with a large-scale benchmark and analyze the performance of traditional CL approaches in a compute-constrained setting, where effective memory samples used in training can be implicitly restricted as a consequence of limited computation. We conduct experiments evaluating various CL sampling strategies, distillation losses, and partial fine-tuning on two large-scale datasets, namely ImageNet2K and Continual Google Landmarks V2 in data incremental, class incremental, and time incremental settings. Through extensive experiments amounting to a total of over 1500 GPU-hours, we find that, under compute-constrained setting, traditional CL approaches, with no exception, fail to outperform a simple minimal baseline that samples uniformly from memory. Our conclusions are consistent in a different number of stream time steps, e.g., 20 to 200, and under several computational budgets. This suggests that most existing CL methods are particularly too computationally expensive for realistic budgeted deployment. Code for this project is available at: https://github.com/drimpossible/BudgetCL.

Modern machine learning pipelines are limited due to data availability, storage quotas, privacy regulations, and expensive annotation processes. These constraints make it difficult or impossible to train and update large-scale models on such dynamic annotated sets. Continual learning directly approaches this problem, with the ultimate goal of devising methods where a deep neural network effectively learns relevant patterns for new (unseen) classes, without significantly altering its performance on previously learned ones. In this paper, we address the problem of continual learning for video data. We introduce PIVOT, a novel method that leverages extensive knowledge in pre-trained models from the image domain, thereby reducing the number of trainable parameters and the associated forgetting. Unlike previous methods, ours is the first approach that effectively uses prompting mechanisms for continual learning without any in-domain pre-training. Our experiments show that PIVOT improves state-of-the-art methods by a significant 27% on the 20-task ActivityNet setup.

Temporal video segmentation is the get-to- go automatic video analysis, which decomposes a long-form video into smaller components for the following-up understanding tasks. Recent works have studied several levels of granularity to segment a video, such as shot, event, and scene. Those segmentations can help compare the semantics in the corresponding scales, but lack a wider view of larger temporal spans, especially when the video is complex and structured. Therefore, we present two abstractive levels of temporal segmentations and study their hierarchy to the existing fine-grained levels. Accordingly, we collect NewsNet, the largest news video dataset consisting of 1,000 videos in over 900 hours, associated with several tasks for hierarchical temporal video segmentation. Each news video is a collection of stories on different topics, represented as aligned audio, visual, and textual data, along with extensive frame-wise annotations in four granularities. We assert that the study on NewsNet can advance the understanding of complex structured video and benefit more areas such as short-video creation, personalized advertisement, digital instruction, and education. Our dataset and code is publicly available at https://github.com/NewsNet-Benchmark/NewsNet.

This paper deals with the problem of localizing objects in image and video datasets from visual exemplars. In particular, we focus on the challenging problem of egocentric visual query localization. We first identify grave implicit biases in current query-conditioned model design and visual query datasets. Then, we directly tackle such biases at both frame and object set levels. Concretely, our method solves these issues by expanding limited annotations and dynamically dropping object proposals during training. Additionally, we propose a novel transformer-based module that allows for object-proposal set context to be considered while incorporating query information. We name our module Conditioned Contextual Transformer or CocoFormer. Our experiments show the proposed adaptations improve egocentric query detection, leading to a better visual query localization system in both 2D and 3D configurations. Thus, we can improve frame-level detection performance from 26.28% to 31.26% in AP, which correspondingly improves the VQ2D and VQ3D localization scores by significant margins. Our improved context-aware query object detector ranked first and second respectively in the VQ2D and VQ3D tasks in the 2nd Ego4D challenge. In addition to this, we showcase the relevance of our proposed model in the Few-Shot Detection (FSD) task, where we also achieve SOTA results. Our code is available at https://github.com/facebookresearch/vq2d_cvpr.

The recent introduction of the large-scale, long-form MAD and Ego4D datasets has enabled researchers to investigate the performance of current state-of-the-art methods for video grounding in the long-form setup, with interesting findings: current grounding methods alone fail at tackling this challenging task and setup due to their inability to process long video sequences. In this paper, we propose a method for improving the performance of natural language grounding in long videos by identifying and pruning out non-describable windows. We design a guided grounding framework consisting of a Guidance Model and a base grounding model. The Guidance Model emphasizes describable windows, while the base grounding model ana-lyzes short temporal windows to determine which segments accurately match a given language query. We offer two designs for the Guidance Model: Query-Agnostic and Query-Dependent, which balance efficiency and accuracy. Experiments demonstrate that our proposed method outperforms state-of-the-art models by 4.1% in MAD and 4.52% in Ego4D (NLQ), respectively. Code, data and MAD’s audio features necessary to reproduce our experiments are available at: https://github.com/waybarrios/guidance-based-video-grounding.

Semantic segmentation is a crucial task in computer vision that involves segmenting images into semantically meaningful regions at the pixel level. However, existing approaches often rely on expensive human annotations as supervision for model training, limiting their scalability to large, unlabeled datasets. To address this challenge, we present ZeroSeg, a novel method that leverages the existing pretrained vision-language (VL) model (e.g. CLIP vision encoder [39]) to train open-vocabulary zero-shot semantic segmentation models. Although acquired extensive knowledge of visual concepts, it is non-trivial to exploit knowledge from these VL models to the task of semantic segmentation, as they are usually trained at an image level. ZeroSeg overcomes this by distilling the visual concepts learned by VL models into a set of segment tokens, each summarizing a localized region of the target image. We evaluate ZeroSeg on multiple popular segmentation benchmarks, including PASCAL VOC 2012, PASCAL Context, and COCO, in a zero-shot manner Our approach achieves state-of-the-art performance when compared to other zero-shot segmentation methods under the same training data, while also performing competitively compared to strongly supervised methods. Finally, we also demonstrated the effectiveness of ZeroSeg on open-vocabulary segmentation, through both human studies and qualitative visualizations. The code is publicly available at https://github.com/facebookresearch/ZeroSeg

The "pre-training → downstream adaptation" presents both new opportunities and challenges for Continual Learning (CL). Although the recent state-of-the-art in CL is achieved through Parameter-Efficient-Tuning (PET) adaptation paradigm, only prompt has been explored, limiting its application to Transformers only. In this paper, we position prompting as one instantiation of PET, and propose a unified CL framework with general PET, dubbed as Learning-Accumulation-Ensemble (LAE). PET, e.g., using Adapter, LoRA, or Prefix, can adapt a pre-trained model to downstream tasks with fewer parameters and resources. Given a PET method, our LAE framework incorporates it for CL with three novel designs. 1) Learning: the pre-trained model adapts to the new task by tuning an online PET module, along with our adaptation speed calibration to align different PET modules, 2) Accumulation: the task-specific knowledge learned by the online PET module is accumulated into an offline PET module through momentum update, 3) Ensemble: During inference, we respectively construct two experts with online/offline PET modules (which are favored by the novel/historical tasks) for prediction ensemble. We show that LAE is compatible with a battery of PET methods and gains strong CL capability. For example, LAE with Adaptor PET surpasses the prior state-of-the-art by 1.3% and 3.6% in last-incremental accuracy on CIFAR100 and ImageNet-R datasets, respectively. Code is available at https://github.com/gqk/LAE.

We revisit the common practice of evaluating adaptation of Online Continual Learning (OCL) algorithms through the metric of online accuracy, which measures the accuracy of the model on the immediate next few samples. However, we show that this metric is unreliable, as even vacuous blind classifiers, which do not use input images for prediction, can achieve unrealistically high online accuracy by exploiting spurious label correlations in the data stream. Our study reveals that existing OCL algorithms can also achieve high online accuracy, but perform poorly in retaining useful information, suggesting that they unintentionally learn spurious label correlations. To address this issue, we propose a novel metric for measuring adaptation based on the accuracy on the near-future samples, where spurious correlations are removed. We benchmark existing OCL approaches using our proposed metric on large-scale datasets under various computational budgets and find that better generalization can be achieved by retaining and reusing past seen information. We believe that our proposed metric can aid in the development of truly adaptive OCL methods. We provide code to reproduce our results at https://github.com/drimpossible/EvalOCL.

Recent work on deep reinforcement learning (DRL) has pointed out that algorithmic information about good policies can be extracted from offline data which lack explicit information about executed actions [45], [46], [30]. For example, videos of humans or robots may convey a lot of implicit information about rewarding action sequences, but a DRL machine that wants to profit from watching such videos must first learn by itself to identify and recognize relevant states/actions/rewards. Without relying on ground-truth annotations, our new method called Deep State Identifier learns to predict returns from episodes encoded as videos. Then it uses a kind of mask-based sensitivity analysis to extract/identify important critical states. Extensive experiments showcase our method’s potential for understanding and improving agent behavior. The source code and the generated datasets are available at https://github.com/AI-Initiative-KAUST/VideoRLCS.

With the recent advances in video and 3D understanding, novel 4D spatio-temporal methods fusing both concepts have emerged. Towards this direction, the Ego4D Episodic Memory Benchmark proposed a task for Visual Queries with 3D Localization (VQ3D). Given an egocentric video clip and an image crop depicting a query object, the goal is to localize the 3D position of the center of that query object with respect to the camera pose of a query frame. Current methods tackle the problem of VQ3D by unprojecting the 2D localization results of the sibling task Visual Queries with 2D Localization (VQ2D) into 3D predictions. Yet, we point out that the low number of camera poses caused by camera re-localization from previous VQ3D methods severally hinders their overall success rate. In this work, we formalize a pipeline (we dub EgoLoc) that better entangles 3D multiview geometry with 2D object retrieval from egocentric videos. Our approach involves estimating more robust camera poses and aggregating multi-view 3D displacements by leveraging the 2D detection confidence, which enhances the success rate of object queries and leads to a significant improvement in the VQ3D baseline performance. Specifically, our approach achieves an overall success rate of up to 87.12%, which sets a new state-of-the-art result in the VQ3D task1. We provide a comprehensive empirical analysis of the VQ3D task and existing solutions, and highlight the remaining challenges in VQ3D. The code is available at https://github.com/Wayne-Mai/EgoLoc.

This paper proposes a novel approach for rendering a pre-trained Neural Radiance Field (NeRF) in real-time on resource-constrained devices. We introduce Re-ReND, a method enabling Real-time Rendering of NeRFs across Devices. Re-ReND is designed to achieve real-time performance by converting the NeRF into a representation that can be efficiently processed by standard graphics pipelines. The proposed method distills the NeRF by extracting the learned density into a mesh, while the learned color information is factorized into a set of matrices that represent the scene’s light field. Factorization implies the field is queried via inexpensive MLP-free matrix multiplications, while using a light field allows rendering a pixel by querying the field a single time—as opposed to hundreds of queries when employing a radiance field. Since the proposed representation can be implemented using a fragment shader, it can be directly integrated with standard rasterization frameworks. Our flexible implementation can render a NeRF in real-time with low memory requirements and on a wide range of resource-constrained devices, including mobiles and AR/VR headsets. Notably, we find that Re-ReND can achieve over a 2.6-fold increase in rendering speed versus the state-of-the-art without perceptible losses in quality.

We propose a novel approach to animate human hair in a still portrait photo. Existing work has largely studied the animation of fluid elements such as water and fire. However, hair animation for a real image remains underexplored, which is a challenging problem, due to the high complexity of hair structure and dynamics. Considering the complexity of hair structure, we innovatively treat hair wisp extraction as an instance segmentation problem, where a hair wisp is referred to as an instance. With advanced instance segmentation networks, our method extracts meaningful and natural hair wisps. Furthermore, we propose a wisp-aware animation module that animates hair wisps with pleasing motions without noticeable artifacts. The extensive experiments show the superiority of our method. Our method provides the most pleasing and compelling viewing experience in the qualitative experiments, and outperforms state-of-the-art still-image animation methods by a large margin in the quantitative evaluation. Project url: https://nevergiveu.github.io/AutomaticHairBlowing/

Recently, conditional diffusion models have gained popularity in numerous applications due to their exceptional generation ability. However, many existing methods are training-required. They need to train a time-dependent classifier or a condition-dependent score estimator, which increases the cost of constructing conditional diffusion models and is inconvenient to transfer across different conditions. Some current works aim to overcome this limitation by proposing training-free solutions, but most can only be applied to a specific category of tasks and not to more general conditions. In this work, we propose a training-Free conditional Diffusion Model (FreeDoM) used for various conditions. Specifically, we leverage off-the-shelf pretrained networks, such as a face detection model, to construct time-independent energy functions, which guide the generation process without requiring training. Furthermore, because the construction of the energy function is very flexible and adaptable to various conditions, our proposed FreeDoM has a broader range of applications than existing training-free methods. FreeDoM is advantageous in its simplicity, effectiveness, and low cost. Experiments demonstrate that FreeDoM is effective for various conditions and suitable for diffusion models of diverse data domains, including image and latent code domains. Code is available at https://github.com/vvictoryuki/FreeDoM.

The rapid advancement of chat-based language models has led to remarkable progress in complex task-solving. However, their success heavily relies on human input to guide the conversation, which can be challenging and time-consuming. This paper explores the potential of building scalable techniques to facilitate autonomous cooperation among communicative agents, and provides insight into their “cognitive” processes. To address the challenges of achieving autonomous cooperation, we propose a novel communicative agent framework named role-playing . Our approach involves using inception prompting to guide chat agents toward task completion while maintaining consistency with human intentions. We showcase how role-playing can be used to generate conversational data for studying the behaviors and capabilities of a society of agents, providing a valuable resource for investigating conversational language models. In particular, we conduct comprehensive studies on instruction-following cooperation in multi-agent settings. Our contributions include introducing a novel communicative agent framework, offering a scalable approach for studying the cooperative behaviors and capabilities of multi-agent systems, and open-sourcing our library to support research on communicative agents and beyond: https://github.com/camel-ai/camel.

Training Generative Adversarial Networks (GANs) remains a challenging problem. The discriminator trains the generator by learning the distribution of real/generated data. However, the distribution of generated data changes throughout the training process, which is difficult for the discriminator to learn. In this paper, we propose a novel method for GANs from the viewpoint of online continual learning. We observe that the discriminator model, trained on historically generated data, often slows down its adaptation to the changes in the new arrival generated data, which accordingly decreases the quality of generated results. By treating the generated data in training as a stream, we propose to detect whether the discriminator slows down the learning of new knowledge in generated data. Therefore, we can explicitly enforce the discriminator to learn new knowledge fast. Particularly, we propose a new discriminator, which automatically detects its retardation and then dynamically masks its features, such that the discriminator can adaptively learn the temporally-vary distribution of generated data. Experimental results show our method outperforms the state-of-the-art approaches.

This work tackles the problem of characterizing and understanding the decision boundaries of neural networks with piecewise linear non-linearity activations. We use tropical geometry, a new development in the area of algebraic geometry, to characterize the decision boundaries of a simple network of the form (Affine, ReLU, Affine). Our main finding is that the decision boundaries are a subset of a tropical hypersurface, which is intimately related to a polytope formed by the convex hull of two zonotopes. The generators of these zonotopes are functions of the network parameters. This geometric characterization provides new perspectives to three tasks. (i) We propose a new tropical perspective to the lottery ticket hypothesis, where we view the effect of different initializations on the tropical geometric representation of a network's decision boundaries. (ii) Moreover, we propose new tropical based optimization reformulations that directly influence the decision boundaries of the network for the task of network pruning. (iii) At last, we discuss the reformulation of the generation of adversarial attacks in a tropical sense. We demonstrate that one can construct adversaries in a new tropical setting by perturbing a specific set of decision boundaries by perturbing a set of parameters in the network.

Convolutional neural networks (CNNs) have been very successful at solving a variety of computer vision tasks such as object classification and detection, semantic segmentation, activity understanding, to name just a few. One key enabling factor for their great performance has been the ability to train very deep networks. Despite their huge success in many tasks, CNNs do not work well with non-euclidean data, which is prevalent in many real-world applications. Graph Convolutional Networks (GCNs) offer an alternative that allows for non-Eucledian data input to a neural network. While GCNs already achieve encouraging results, they are currently limited to architectures with a relatively small number of layers, primarily due to vanishing gradients during training. This work transfers concepts such as residual/dense connections and dilated convolutions from CNNs to GCNs in order to successfully train very deep GCNs. We show the benefit of using deep GCNs (with as many as 112 layers) experimentally across various datasets and tasks. Specifically, we achieve very promising performance in part segmentation and semantic segmentation on point clouds and in node classification of protein functions across biological protein-protein interaction (PPI) graphs. We believe that the insights in this work will open avenues for future research on GCNs and their application to further tasks not explored in this paper. The source code for this work is available at https://github.com/lightaime/deep_gcns_torch and https://github.com/lightaime/deep_gcns for PyTorch and TensorFlow implementations respectively.

Graph Neural Networks (GNNs) have been drawing significant attention to representation learning on graphs. Recent works developed frameworks to train very deep GNNs and showed impressive results in tasks like point cloud learning and protein interaction prediction. In this work, we study the performance of such deep models in large-scale graphs. In particular, we look at the effect of adequately choosing an aggregation function on deep models. We find that GNNs are very sensitive to the choice of aggregation functions (e.g. mean, max, and sum) when applied to different datasets. We systematically study and propose to alleviate this issue by introducing a novel class of aggregation functions named Generalized Aggregation Functions. The proposed functions extend beyond commonly used aggregation functions to a wide range of new permutation-invariant functions. Generalized Aggregation Functions are fully differentiable, where their parameters can be learned in an end-to-end fashion to yield a suitable aggregation function for each task. We show that equipped with the proposed aggregation functions, deep residual GNNs outperform state-of-the-art in several benchmarks from Open Graph Benchmark (OGB) across tasks and domains.

The task of situation recognition aims to solve the visual reasoning problem with the ability to predict the activity happening (salient action) in an image and the nouns of all associated semantic roles playing in the activity. This poses severe challenges due to long-tailed data distributions and local class ambiguities. Prior works only propagate the local noun-level features on one single image without utilizing global information. We propose a Knowledge-aware Global Reasoning (KGR) framework to endow neural networks with the capability of adaptive global reasoning over nouns by exploiting diverse statistical knowledge. Our KGR is a local-global architecture, which consists of a local encoder to generate noun features using local relations and a global encoder to enhance the noun features via global reasoning supervised by an external global knowledge pool. The global knowledge pool is created by counting the pairwise relationships of nouns in the dataset. In this paper, we design an action-guided pairwise knowledge as the global knowledge pool based on the characteristic of the situation recognition task. Extensive experiments have shown that our KGR not only achieves state-of-the-art results on a large-scale situation recognition benchmark, but also effectively solves the long-tailed problem of noun classification by our global knowledge.

Point cloud architecture design has become a crucial problem for deep learning in 3D. Several efforts have been made to manually design architectures targeting high accuracy in point cloud tasks such as classification, segmentation, and detection. Recent progress in automatic Neural Architecture Search (NAS) minimizes the human effort in network design and optimizes architectures for high performance. However, those efforts fail to consider crucial factors such as latency during inference, which is of high importance in time-critical and hardware-bounded applications like self-driving cars, robot navigation, and mobile applications. In this paper, we introduce a new NAS framework, dubbed LC-NAS, that searches for point cloud architectures constrained to a target latency. We implement a novel latency constraint formulation for the trade-off between accuracy and latency in our architecture search. Contrary to previous works, our latency loss enables us to find the best architecture with latency near a specific target value, which is crucial when the end task is to be deployed in a limited hardware setting. Extensive experiments show that LC-NAS is able to find state-of-the-art architectures for point cloud classification in ModelNet40 with a minimal computational cost. We also show how our searched architectures achieve any desired latency with a reasonably low drop in accuracy. Finally, we show how our searched architectures easily transfer to the part segmentation task on PartNet, where we achieve state-of-the-art results with significantly lower latency.

Deep neural networks are vulnerable to input deformations in the form of vector fields of pixel displacements and to other parameterized geometric deformations e.g. translations, rotations, etc. Current input deformation certification methods either (i) do not scale to deep networks on large input datasets, or (ii) can only certify a specific class of deformations, e.g. only rotations. We reformulate certification in randomized smoothing setting for both general vector field and parameterized deformations and propose DeformRS-VF and DeformRS-Par, respectively. Our new formulation scales to large networks on large input datasets. For instance, DeformRS-Par certifies rich deformations, covering translations, rotations, scaling, affine deformations, and other visually aligned deformations such as ones parameterized by Discrete-Cosine-Transform basis. Extensive experiments on MNIST, CIFAR10, and ImageNet show competitive performance of DeformRS-Par achieving a certified accuracy of 39\% against perturbed rotations in the set [-10 degree, 10 degree] on ImageNet.

Deep neural networks are vulnerable to small input perturbations known as adversarial attacks. Inspired by the fact that these adversaries are constructed by iteratively minimizing the confidence of a network for the true class label, we propose the anti-adversary layer, aimed at countering this effect. In particular, our layer generates an input perturbation in the opposite direction of the adversarial one and feeds the classifier a perturbed version of the input. Our approach is training-free and theoretically supported. We verify the effectiveness of our approach by combining our layer with both nominally and robustly trained models and conduct large-scale experiments from black-box to adaptive attacks on CIFAR10, CIFAR100, and ImageNet. Our layer significantly enhances model robustness while coming at no cost on clean accuracy.

We propose a novel scene flow estimation approach to capture and infer 3D motions from point clouds. Estimating 3D motions for point clouds is challenging, since a point cloud is unordered and its density is significantly non-uniform. Such unstructured data poses difficulties in matching corresponding points between point clouds, leading to inaccurate flow estimation. We propose a novel architecture named Sparse Convolution-Transformer Network (SCTN) that equips the sparse convolution with the transformer. Specifically, by leveraging the sparse convolution, SCTN transfers irregular point cloud into locally consistent flow features for estimating spatially consistent motions within an object/local object part. We further propose to explicitly learn point relations using a point transformer module, different from exiting methods. We show that the learned relation-based contextual information is rich and helpful for matching corresponding points, benefiting scene flow estimation. In addition, a novel loss function is proposed to adaptively encourage flow consistency according to feature similarity. Extensive experiments demonstrate that our proposed approach achieves a new state of the art in scene flow estimation. Our approach achieves an error of 0.038 and 0.037 (EPE3D) on FlyingThings3D and KITTI Scene Flow respectively, which significantly outperforms previous methods by large margins.

We introduce Ego4D, a massive-scale egocentric video dataset and benchmark suite. It offers 3,670 hours of dailylife activity video spanning hundreds of scenarios (household, outdoor, workplace, leisure, etc.) captured by 931 unique camera wearers from 74 worldwide locations and 9 different countries. The approach to collection is designed to uphold rigorous privacy and ethics standards, with consenting participants and robust de-identification procedures where relevant. Ego4D dramatically expands the volume of diverse egocentric video footage publicly available to the research community. Portions of the video are accompanied by audio, 3D meshes of the environment, eye gaze, stereo, and/or synchronized videos from multiple egocentric cameras at the same event. Furthermore, we present a host of new benchmark challenges centered around understanding the first-person visual experience in the past (querying an episodic memory), present (analyzing hand-object manipulation, audio-visual conversation, and social interactions), and future (forecasting activities). By publicly sharing this massive annotated dataset and benchmark suite, we aim to push the frontier of first-person perception. Project page: https://ego4d-data.org/

Data augmentation helps neural networks generalize better by enlarging the training set, but it remains an open question how to effectively augment graph data to enhance the performance of GNNs (Graph Neural Networks). While most existing graph regularizers focus on manipulating graph topological structures by adding/removing edges, we offer a method to augment node features for better performance. We propose FLAG (Free Large-scale Adversarial Augmentation on Graphs), which iteratively augments node features with gradient-based adversarial perturbations during training. By making the model invariant to small fluctuations in input data, our method helps models generalize to out-of-distribution samples and boosts model performance at test time. FLAG is a general-purpose approach for graph data, which universally works in node classification, link prediction, and graph classification tasks. FLAG is also highly flexible and scalable, and is deployable with arbitrary GNN backbones and large-scale datasets. We demon-strate the efficacy and stability of our method through ex-tensive experiments and ablation studies. We also provide intuitive observations for a deeper understanding of our method. We open source our implementation at https://github.com/devnkong/FLAG.

Hyperspectral imaging has attracted significant attention to identify spectral signatures for image classification and automated pattern recognition in computer vision. State-of-the-art implementations of snapshot hyperspectral imaging rely on bulky, non-integrated, and expensive optical elements, including lenses, spectrometers, and filters. These macroscopic components do not allow fast data processing for, e.g. real-time and high-resolution videos. This work introduces Hyplex™, a new integrated architecture addressing the limitations discussed above. Hyplex™ is a CMOS-compatible, fast hyperspectral camera that replaces bulk optics with nanoscale metasurfaces inversely designed through artificial intelligence. Hyplex™ does not require spectrometers but makes use of conventional monochrome cameras, opening up the possibility for real-time and high-resolution hyperspectral imaging at inexpensive costs. Hyplex™ exploits a model-driven optimization, which connects the physical metasurfaces layer with modern visual computing approaches based on end-to-end training. We design and implement a prototype version of Hyplex™ and compare its performance against the state-of-the-art for typical imaging tasks such as spectral reconstruction and semantic segmentation. In all benchmarks, Hyplex™ reports the smallest reconstruction error. We additionally present what is, to the best of our knowledge, the largest publicly available labeled hyperspectral dataset for semantic segmentation.11Dataset available on https://github.com/makamoa/hyplex.

3D computer vision models are commonly used in security-critical applications such as autonomous driving and surgical robotics. Emerging concerns over the robustness of these models against real-world deformations must be addressed practically and reliably. In this work, we propose 3DeformRS, a method to certify the robustness of point cloud Deep Neural Networks (DNNs) against real-world deformations. We developed 3DeformRS by building upon recent work that generalized Randomized Smoothing (RS) from pixel-intensity perturbations to vector-field deformations. In particular, we specialized RS to certify DNNs against parameterized deformations (e.g. rotation, twisting), while enjoying practical computational costs. We leverage the virtues of 3DeformRS to conduct a comprehensive empirical study on the certified robustness of four representative point cloud DNNs on two datasets and against seven different deformations. Compared to previous approaches for certifying point cloud DNNs, 3DeformRS is fast, scales well with point cloud size, and provides comparable-to-better certificates. For instance, when certifying a plain PointNet against a 3° z-rotation on 1024-point clouds, 3DeformRS grants a certificate 3× larger and 20× faster than previous work11Code:https://github.com/gaperezsa/3DeformRS.

The recent and increasing interest in video-language research has driven the development of large-scale datasets that enable data-intensive machine learning techniques. In comparison, limited effort has been made at assessing the fitness of these datasets for the video-language grounding task. Recent works have begun to discover significant limitations in these datasets, suggesting that state-of-the-art techniques commonly overfit to hidden dataset biases. In this work, we present MAD (Movie Audio Descriptions), a novel benchmark that departs from the paradigm of augmenting existing video datasets with text annotations and focuses on crawling and aligning available audio descriptions of mainstream movies. MAD contains over 384, 000 natural language sentences grounded in over 1, 200 hours of videos and exhibits a significant reduction in the currently diagnosed biases for video-language grounding datasets. MAD's collection strategy enables a novel and more challenging version of video-language grounding, where short temporal moments (typically seconds long) must be accurately grounded in diverse long-form videos that can last up to three hours. We have released MAD's data and baselines code at https://github.com/Soldelli/MAD.

We propose a novel few-shot action recognition framework, STRM, which enhances class-specific feature discrim-inability while simultaneously learning higher-order temporal representations. The focus of our approach is a novel spatio-temporal enrichment module that aggregates spatial and temporal contexts with dedicated local patch-level and global frame-level feature enrichment sub-modules. Local patch-level enrichment captures the appearance-based characteristics of actions. On the other hand, global framelevel enrichment explicitly encodes the broad temporal context, thereby capturing the relevant object features over time. The resulting spatio-temporally enriched representations are then utilized to learn the relational matching between query and support action sub-sequences. We further introduce a query-class similarity classifier on the patch-level enriched features to enhance class-specific feature dis-criminability by reinforcing the feature learning at different stages in the proposed framework. Experiments are performed on four few-shot action recognition benchmarks: Kinetics, SSv2, HMDB51 and UCF101. Our extensive ablation study reveals the benefits of the proposed contributions. Furthermore, our approach sets a new state-of-the-art on all four benchmarks. On the challenging SSv2 benchmark, our approach achieves an absolute gain of 3.5% in classification accuracy, as compared to the best existing method in the literature. Our code and models are available at https://github.com/Anirudh257/strm.

Continual learning (CL) is under-explored in the video domain. The few existing works contain splits with imbalanced class distributions over the tasks, or study the problem in unsuitable datasets. We introduce vCLIMB, a novel video continual learning benchmark. vCLIMB is a standardized test-bed to analyze catastrophic forgetting of deep models in video continual learning. In contrast to previous work, we focus on class incremental continual learning with models trained on a sequence of disjoint tasks, and distribute the number of classes uniformly across the tasks. We perform in-depth evaluations of existing CL methods in vCLIMB, and observe two unique challenges in video data. The selection of instances to store in episodic memory is performed at the frame level. Second, untrimmed training data influences the effectiveness of frame sampling strategies. We address these two challenges by proposing a temporal consistency regularization that can be applied on top of memory-based continual learning methods. Our approach significantly improves the baseline, by up to 24% on the untrimmed continual learning task. The code of our benchmark can be found at: https://vclimb.netlify.app/.

Imaging is usually a mixture problem of incomplete color sampling, noise degradation, and limited resolution. This mixture problem is typically solved by a sequential solution that applies demosaicing (DM), denoising (DN), and super-resolution (SR) sequentially in a fixed and predefined pipeline (execution order of tasks), DM→DN→SR. The most recent work on image processing focuses on developing more sophisticated architectures to achieve higher image quality. Little attention has been paid to the design of the pipeline, and it is still not clear how significant the pipeline is to image quality. In this work, we comprehensively study the effects of pipelines on the mixture problem of learning-based DN, DM, and SR, in both sequential and joint solutions. On the one hand, in sequential solutions, we find that the pipeline has a non-trivial effect on the resulted image quality. Our suggested pipeline DN→SR→DM yields consistently better performance than other sequential pipelines in various experimental settings and benchmarks. On the other hand, in joint solutions, we propose an end-to-end Trinity Pixel Enhancement NETwork (TENet) that achieves the state-of-the-art performance for the mixture problem. We further present a novel and simple method that can integrate a certain pipeline into a given end-to-end network by providing intermediate supervision using a detachable head. Extensive experiments show that an end-to-end network with the proposed pipeline can attain only a consistent but insignificant improvement. Our work indicates that the investigation of pipelines is applicable in sequential solutions, but is not very necessary in end-to-end networks.

Video-Language Pretraining (VLP), which aims to learn transferable representation to advance a wide range of video-text downstream tasks, has recently received increasing attention. Best performing works rely on large-scale, 3rd-person video-text datasets, such as HowTo100M. In this work, we exploit the recently released Ego4D dataset to pioneer Egocentric VLP along three directions. (i) We create EgoClip, a 1st-person video-text pretraining dataset comprising 3.8M clip-text pairs well-chosen from Ego4D, covering a large variety of human daily activities. (ii) We propose a novel pretraining objective, dubbed EgoNCE, which adapts video-text contrastive learning to the egocentric domain by mining egocentric-aware positive and negative samples. (iii) We introduce EgoMCQ, a development benchmark that is close to EgoClip and hence can support effective validation and fast exploration of our design decisions in EgoClip and EgoNCE. Furthermore, we demonstrate strong performance on five egocentric downstream tasks across three datasets: video-text retrieval on EPIC-KITCHENS-100; action recognition on Charades-Ego; natural language query, moment query, and object state change classification on Ego4D challenge benchmarks. The dataset and code are available at https://github.com/showlab/EgoVLP.

PointNet++ is one of the most influential neural architectures for point cloud understanding. Although the accuracy of PointNet++ has been largely surpassed by recent networks such as PointMLP and Point Transformer, we find that a large portion of the performance gain is due to improved training strategies, i.e. data augmentation and optimization techniques, and increased model sizes rather than architectural innovations. Thus, the full potential of PointNet++ has yet to be explored. In this work, we revisit the classical PointNet++ through a systematic study of model training and scaling strategies, and offer two major contributions. First, we propose a set of improved training strategies that significantly improve PointNet++ performance. For example, we show that, without any change in architecture, the overall accuracy (OA) of PointNet++ on ScanObjectNN object classification can be raised from 77.9% to 86.1%, even outperforming state-of-the-art PointMLP. Second, we introduce an inverted residual bottleneck design and separable MLPs into PointNet++ to enable efficient and effective model scaling and propose PointNeXt, the next version of PointNets. PointNeXt can be flexibly scaled up and outperforms state-of-the-art methods on both 3D classification and segmentation tasks. For classification, PointNeXt reaches an overall accuracy of 87.7 on ScanObjectNN, surpassing PointMLP by 2.3%, while being 10x faster in inference. For semantic segmentation, PointNeXt establishes a new state-of-the-art performance with 74.9% mean IoU on S3DIS (6-fold cross-validation), being superior to the recent Point Transformer. The code and models are available at https://github.com/guochengqian/pointnext.

Randomized smoothing is a recent technique that achieves state-of-art performance in training certifiably robust deep neural networks. While the smoothing family of distributions is often connected to the choice of the norm used for certification, the parameters of these distributions are always set as global hyper parameters independent from the input data on which a network is certified. In this work, we revisit Gaussian randomized smoothing and show that the variance of the Gaussian distribution can be optimized at each input so as to maximize the certification radius for the construction of the smooth classifier. Since the data dependent classifier does not directly enjoy sound certification with existing approaches, we propose a memory-enhanced data dependent smooth classifier that is certifiable by construction. This new approach is generic, parameter-free, and easy to implement. In fact, we show that our data dependent framework can be seamlessly incorporated into 3 randomized smoothing approaches, leading to consistent improved certified accuracy. When this framework is used in the training routine of these approaches followed by a data dependent certification, we achieve 9% and 6% improvement over the certified accuracy of the strongest baseline for a radius of 0.5 on CIFAR10 and ImageNet.

Video grounding aims at localizing the temporal moment related to the given language description, which is very helpful to many cross-modal content understanding applications like visual question answering and sentence-video search. Existing approaches usually directly regress the temporal boundaries of an event described by a query sentence in the video sequence. This direct regression manner often encounters a large decision space due to diverse target events and variable video durations, leading to inaccurate localization as well as inefficient grounding. This paper presents an efficient framework termed from which to where to facilitate video grounding. The core idea is imitating the reading comprehension process to gradually narrow the decision space, in what we decompose the direct regression into two steps. The “which” step first roughly selects a candidate area by evaluating which video segment in the predefined set is closest to the ground truth. To this end, we formulate this step into a multi-choice reading comprehension problem and propose a criterion to select the best-matched segment. In this way, the excessive decision space is effectively reduced. The “where” step aims to precisely regress the temporal boundary of the selected video segment from the shrunk decision space. We thus introduce a triple-span representation for each candidate video segment to use the regional context for better boundary regression. The “which” and “where” steps can be combined into a unified framework and learned end-to-end, leading to an efficient video grounding system. Extensive experiments on Charades-STA, ActivityNet-Captions, and TACoS benchmarks clearly demonstrate the effectiveness of our framework.

In this paper, we propose a novel framework to translate a portrait photo-face into an anime appearance. Different from existing translation methods which do not designate specific styles, we aim to synthesize anime-faces which are style-consistent with a given reference anime-face. However, unlike typical translation tasks, such anime-face translation is particularly challenging due to the large and complex variations of appearances among anime-faces. Existing methods often fail to transfer the styles of reference anime-faces to the generated anime-faces, or introduce noticeable artifacts/distortions in the local shapes of their generated anime-faces. We propose a novel GAN-based anime-face translator, called AniGAN, to synthesize high-quality anime-faces. Specifically, a new generator architecture is proposed to simultaneously transfer color/texture styles and transform local facial shapes into anime-like counterparts based on the style of a reference anime-face, while preserving the global structure of the source photo-face. New normalization functions are designed for the generator to further improve local shape transformation and color/texture style transfer. Besides, we propose a double-branch discriminator to learn domain-specific distributions through individual branches and learn cross-domain shared distributions via shared layers, helping generate visually pleasing anime-faces and effectively mitigate artifacts/distortions. Extensive experiments on benchmark datasets qualitatively and quantitatively demonstrate the superiority of our method over state-of-the-art methods.

The effectiveness of learning-based point cloud upsampling pipelines heavily relies on the upsampling modules and feature extractors used therein. For the point upsampling module, we propose a novel model called NodeShuffle, which uses a Graph Convolutional Network (GCN) to better encode local point information from point neighborhoods. NodeShuffle is versatile and can be incorporated into any point cloud upsampling pipeline. Extensive experiments show how NodeShuffle consistently improves state-of-the-art upsampling methods. For feature extraction, we also propose a new multi-scale point feature extractor, called Inception DenseGCN. By aggregating features at multiple scales, this feature extractor enables further performance gain in the final upsampled point clouds. We combine Inception DenseGCN with NodeShuffle into a new point upsampling pipeline called PU-GCN. PU-GCN sets new state-of-art performance with much fewer parameters and more efficient inference. Our code is publicly available at https://github.com/guochengqian/PU-GCN.

Temporal language grounding in videos aims to localize the temporal span relevant to the given query sentence.Previous methods treat it either as a boundary regression task or a span extraction task.This paper will formulate temporal language grounding into video reading comprehension and propose a Relation-aware Network (RaNet) to address it.This framework aims to select a video moment choice from the predefined answer set with the aid of coarse-and-fine choice-query interaction and choice-choice relation construction.A choicequery interactor is proposed to match the visual and textual information simultaneously in sentence-moment and token-moment levels, leading to a coarse-and-fine cross-modal interaction.Moreover, a novel multi-choice relation constructor is introduced by leveraging graph convolution to capture the dependencies among video moment choices for the best choice selection.Extensive experiments on ActivityNet-Captions, TACoS, and Charades-STA demonstrate the effectiveness of our solution.Codes will be available at https: //github.com/Huntersxsx/RaNet.

A high quality disparity remapping method that preserves 2D shapes and 3D structures, and adjusts disparities of important objects in stereo image pairs is proposed. It is formulated as a constrained optimization problem, whose solution is challenging, since we need to meet multiple requirements of disparity remapping simultaneously. The one-stage optimization process either degrades the quality of important objects or introduces serious distortions in background regions. To address this challenge, we propose a two-stage warping process to solve it. In the first stage, we develop a warping model that finds the optimal warping grids for important objects to fulfill multiple requirements of disparity remapping. In the second stage, we derive another warping model to refine warping results in less important regions by eliminating serious distortions in shape, disparity and 3D structure. The superior performance of the proposed method is demonstrated by experimental results.

Active speaker detection requires a mindful integration of multi-modal cues. Current methods focus on modeling and fusing short-term audiovisual features for individual speakers, often at frame level. We present a novel approach to active speaker detection that directly addresses the multi-modal nature of the problem and provides a straightforward strategy, where independent visual features (speakers) in the scene are assigned to a previously detected speech event. Our experiments show that a small graph data structure built from local information can approximate an instantaneous audio-visual assignment problem. More-over, the temporal extension of this initial graph achieves a new state-of-the-art performance on the AVA-ActiveSpeaker dataset with a mAP of 88.8%.

Multi-view projection methods have demonstrated their ability to reach state-of-the-art performance on 3D shape recognition. Those methods learn different ways to aggregate information from multiple views. However, the camera view-points for those views tend to be heuristically set and fixed for all shapes. To circumvent the lack of dynamism of current multi-view methods, we propose to learn those view-points. In particular, we introduce the Multi-View Transformation Network (MVTN) that regresses optimal view-points for 3D shape recognition, building upon advances in differentiable rendering. As a result, MVTN can be trained end-to-end along with any multi-view network for 3D shape classification. We integrate MVTN in a novel adaptive multi-view pipeline that can render either 3D meshes or point clouds. MVTN exhibits clear performance gains in the tasks of 3D shape classification and 3D shape retrieval with-out the need for extra training supervision. In these tasks, MVTN achieves state-of-the-art performance on ModelNet40, ShapeNet Core55, and the most recent and realistic ScanObjectNN dataset (up to 6% improvement). Interestingly, we also show that MVTN can provide network robustness against rotation and occlusion in the 3D domain. The code is available at https://github.com/ajhamdi/MVTN.

Video content creation keeps growing at an incredible pace; yet, creating engaging stories remains challenging and requires non-trivial video editing expertise. Many video editing components are astonishingly hard to automate primarily due to the lack of raw video materials. This paper focuses on a new task for computational video editing, namely the task of raking cut plausibility. Our key idea is to leverage content that has already been edited to learn fine-grained audiovisual patterns that trigger cuts. To do this, we first collected a data source of more than 10K videos, from which we extract more than 255K cuts. We devise a model that learns to discriminate between real and artificial cuts via contrastive learning. We set up a new task and a set of baselines to benchmark video cut generation. We observe that our proposed model outperforms the baselines by large margins. To demonstrate our model in real-world applications, we conduct human studies in a collection of unedited videos. The results show that our model does a better job at cutting than random and alternative baselines.

Many video analysis tasks require temporal localization for the detection of content changes. However, most existing models developed for these tasks are pre-trained on general video action classification tasks. This is due to large scale annotation of temporal boundaries in untrimmed videos being expensive. Therefore, no suitable datasets exist that enable pre-training in a manner sensitive to temporal boundaries. In this paper for the first time, we investigate model pre-training for temporal localization by introducing a novel boundary-sensitive pretext (BSP) task. Instead of relying on costly manual annotations of temporal boundaries, we propose to synthesize temporal boundaries in existing video action classification datasets. By defining different ways of synthesizing boundaries, BSP can then be simply conducted in a self-supervised manner via the classification of the boundary types. This enables the learning of video representations that are much more transferable to downstream temporal localization tasks. Extensive experiments show that the proposed BSP is superior and complementary to the existing action classification-based pre-training counterpart, and achieves new state-of-the-art performance on several temporal localization tasks. Please visit our website for more details https://frostinassiky.github.io/bsp.

Temporal action localization (TAL) in videos is a challenging task, especially due to the large variation in action temporal scales. Short actions usually occupy a major proportion in the datasets, but tend to have the lowest performance. In this paper, we confront the challenge of short actions and propose a multi-level cross-scale solution dubbed as video self-stitching graph network (VSGN). We have two key components in VSGN: video self-stitching (VSS) and cross-scale graph pyramid network (xGPN). In VSS, we focus on a short period of a video and magnify it along the temporal dimension to obtain a larger scale. We stitch the original clip and its magnified counterpart in one input sequence to take advantage of the complementary properties of both scales. The xGPN component further exploits the cross-scale correlations by a pyramid of cross-scale graph networks, each containing a hybrid module to aggregate features from across scales as well as within the same scale. Our VSGN not only enhances the feature representations, but also generates more positive anchors for short actions and more short training samples. Experiments demonstrate that VSGN obviously improves the localization performance of short actions as well as achieving the state-of-the-art overall performance on THUMOS-14 and ActivityNet-v1.3. VSGN code is available at https://github.com/coolbay/VSGN.

Deep graph neural networks (GNNs) have achieved excellent results on various tasks on increasingly large graph datasets with millions of nodes and edges. However, memory complexity has become a major obstacle when training deep GNNs for practical applications due to the immense number of nodes, edges, and intermediate activations. To improve the scalability of GNNs, prior works propose smart graph sampling or partitioning strategies to train GNNs with a smaller set of nodes or sub-graphs. In this work, we study reversible connections, group convolutions, weight tying, and equilibrium models to advance the memory and parameter efficiency of GNNs. We find that reversible connections in combination with deep network architectures enable the training of overparameterized GNNs that significantly outperform existing methods on multiple datasets. Our models RevGNN-Deep (1001 layers with 80 channels each) and RevGNN-Wide (448 layers with 224 channels each) were both trained on a single commodity GPU and achieve an ROC-AUC of 87.74 $\pm$ 0.13 and 88.14 $\pm$ 0.15 on the ogbn-proteins dataset. To the best of our knowledge, RevGNN-Deep is the deepest GNN in the literature by one order of magnitude.

Access to 3D point cloud representations has been widely facilitated by LiDAR sensors embedded in various mobile devices. This has led to an emerging need for fast and accurate point cloud processing techniques. In this paper, we revisit and dive deeper into PointNet++, one of the most influential yet under-explored networks, and develop faster and more accurate variants of the model. We first present a novel Separable Set Abstraction (SA) module that disentangles the vanilla SA module used in PointNet++ into two separate learning stages: (1) learning channel correlation and (2) learning spatial correlation. The Separable SA module is significantly faster than the vanilla version, yet it achieves comparable performance. We then introduce a new Anisotropic Reduction function into our Separable SA module and propose an Anisotropic Separable SA (ASSA) module that substantially increases the network's accuracy. We later replace the vanilla SA modules in PointNet++ with the proposed ASSA modules, and denote the modified network as ASSANet. Extensive experiments on point cloud classification, semantic segmentation, and part segmentation show that ASSANet outperforms PointNet++ and other methods, achieving much higher accuracy and faster speeds. In particular, ASSANet outperforms PointNet++ by $7.4$ mIoU on S3DIS Area 5, while maintaining $1.6 \times $ faster inference speed on a single NVIDIA 2080Ti GPU. Our scaled ASSANet variant achieves $66.8$ mIoU and outperforms KPConv, while being more than $54 \times$ faster.

Most existing temporal action localization (TAL) methods rely on a transfer learning pipeline: by first optimizing a video encoder on a large action classification dataset (i.e., source domain), followed by freezing the encoder and training a TAL head on the action localization dataset (i.e., target domain). This results in a task discrepancy problem for the video encoder – trained for action classification, but used for TAL. Intuitively, joint optimization with both the video encoder and TAL head is a strong baseline solution to this discrepancy. However, this is not operable for TAL subject to the GPU memory constraints, due to the prohibitive computational cost in processing long untrimmed videos. In this paper, we resolve this challenge by introducing a novel low-fidelity (LoFi) video encoder optimization method. Instead of always using the full training configurations in TAL learning, we propose to reduce the mini-batch composition in terms of temporal, spatial, or spatio-temporal resolution so that jointly optimizing the video encoder and TAL head becomes operable under the same memory conditions of a mid-range hardware budget. Crucially, this enables the gradients to flow backwards through the video encoder conditioned on a TAL supervision loss, favourably solving the task discrepancy problem and providing more effective feature representations. Extensive experiments show that the proposed LoFi optimization approach can significantly enhance the performance of existing TAL methods. Encouragingly, even with a lightweight ResNet18 based video encoder in a single RGB stream, our method surpasses two-stream (RGB + optical-flow) ResNet50 based alternatives, often by a good margin. Our code is publicly available at https://github.com/saic-fi/lofiactionlocalization.

Video action detectors are usually trained using datasets with fully-supervised temporal annotations. Building such datasets is an expensive task. To alleviate this problem, recent methods have tried to leverage weak labeling, where videos are untrimmed and only a video-level label is available. In this paper, we propose RefineLoc, a novel weaklysupervised temporal action localization method. RefineLoc uses an iterative refinement approach by estimating and training on snippet-level pseudo ground truth at every iteration. We show the benefit of this iterative approach and present an extensive analysis of five different pseudo ground truth generators. We show the effectiveness of our model on two standard action datasets, ActivityNet v1.2 and THUMOS14. RefineLoc shows competitive results with the stateof-the-art in weakly-supervised temporal localization. Additionally, our iterative refinement process is able to significantly improve the performance of two state-of-the-art methods, setting a new state-of-the-art on THUMOS14.

Abstract We propose a framework to generate customized summarizations of visual data collections, such as collections of images, materials, 3D shapes, and 3D scenes. We assume that the elements in the visual data collections can be mapped to a set of vectors in a feature space, in which a fitness score for each element can be defined, and we pose the problem of customized summarizations as selecting a subset of these elements. We first describe the design choices a user should be able to specify for modeling customized summarizations and propose a corresponding user interface. We then formulate the problem as a constrained optimization problem with binary variables and propose a practical and fast algorithm based on the alternating direction method of multipliers (ADMM). Our results show that our problem formulation enables a wide variety of customized summarizations, and that our solver is both significantly faster than state‐of‐the‐art commercial integer programming solvers and produces better solutions than fast relaxation‐based solvers.

Viewing various stereo images under different viewing conditions has escalated the need for effective object-level remapping techniques. In this paper, we propose a new object spatial mapping scheme, which adjusts the depth and size of the selected object to match user preference and viewing conditions. Existing warping-based methods often distort the shape of important objects or cannot faithfully adjust the depth/size of the selected object due to improper warping such as local rotations. In this paper, by explicitly reducing the transformation freedom degree of warping, we propose an optimization model based on axis-aligned warping for object spatial remapping. The proposed axis-aligned warping based optimization model can simultaneously adjust the depths and sizes of selected objects to their target values without introducing severe shape distortions. Moreover, we propose object consistency constraints to ensure the size/shape of parts inside a selected object to be consistently adjusted. Such constraints improve the size/shape adjustment performance while remaining robust to some extent to incomplete object extraction. Experimental results demonstrate that the proposed method achieves high flexibility and effectiveness in adjusting the size and depth of objects compared with existing methods.

In real-world scenarios (i.e., in the wild), pedestrians are often far from the camera (i.e., small scale), and they often gather together and occlude with each other (i.e., heavily occluded). However, detecting these small-scale and heavily occluded pedestrians remains a challenging problem for the existing pedestrian detection methods. We argue that these problems arise because of two factors: 1) insufficient resolution of feature maps for handling small-scale pedestrians and 2) lack of an effective strategy for extracting body part information that can directly deal with occlusion. To solve the above-mentioned problems, in this article, we propose a key-point-guided super-resolution network (coined KGSNet) for detecting these small-scale and heavily occluded pedestrians in the wild. Specifically, to address factor 1), a super-resolution network is first trained to generate a clear super-resolution pedestrian image from a small-scale one. In the super-resolution network, we exploit key points of the human body to guide the super-resolution network to recover fine details of the human body region for easier pedestrian detection. To address factor 2), a part estimation module is proposed to encode the semantic information of different human body parts where four semantic body parts (i.e., head and upper/middle/bottom body) are extracted based on the key points. Finally, based on the generated clear super-resolved pedestrian patches padded with the extracted semantic body part images at the image level, a classification network is trained to further distinguish pedestrians/backgrounds from the inputted proposal regions. Both proposed networks (i.e., super-resolution network and classification network) are optimized in an alternating manner and trained in an end-to-end fashion. Extensive experiments on the challenging CityPersons data set demonstrate the effectiveness of the proposed method, which achieves superior performance over previous state-of-the-art methods, especially for those small-scale and heavily occluded instances. Beyond this, we also achieve state-of-the-art performance (i.e., 3.89% MR-2on the reasonable subset) on the Caltech data set.

We consider the problem of unconstrained minimization of a smooth objective function in ℝn in a setting where only function evaluations are possible. While importance sampling is one of the most popular techniques used by machine learning practitioners to accelerate the convergence of their models when applicable, there is not much existing theory for this acceleration in the derivative-free setting. In this paper, we propose the first derivative free optimization method with importance sampling and derive new improved complexity results on non-convex, convex and strongly convex functions. We conduct extensive experiments on various synthetic and real LIBSVM datasets confirming our theoretical results. We test our method on a collection of continuous control tasks on MuJoCo environments with varying difficulty. Experiments show that our algorithm is practical for high dimensional continuous control problems where importance sampling results in a significant sample complexity improvement.

One major factor impeding more widespread adoption of deep neural networks (DNNs) is their lack of robustness, which is essential for safety-critical applications such as autonomous driving. This has motivated much recent work on adversarial attacks for DNNs, which mostly focus on pixel-level perturbations void of semantic meaning. In contrast, we present a general framework for adversarial attacks on trained agents, which covers semantic perturbations to the environment of the agent performing the task as well as pixel-level attacks. To do this, we re-frame the adversarial attack problem as learning a distribution of parameters that always fools the agent. In the semantic case, our proposed adversary (denoted as BBGAN) is trained to sample parameters that describe the environment with which the black-box agent interacts, such that the agent performs its dedicated task poorly in this environment. We apply BBGAN on three different tasks, primarily targeting aspects of autonomous navigation: object detection, self-driving, and autonomous UAV racing. On these tasks, BBGAN can generate failure cases that consistently fool a trained agent.

Current methods for active speaker detection focus on modeling audiovisual information from a single speaker. This strategy can be adequate for addressing single-speaker scenarios, but it prevents accurate detection when the task is to identify who of many candidate speakers are talking. This paper introduces the Active Speaker Context, a novel representation that models relationships between multiple speakers over long time horizons. Our new model learns pairwise and temporal relations from a structured ensemble of audiovisual observations. Our experiments show that a structured feature ensemble already benefits active speaker detection performance. We also find that the proposed Active Speaker Context improves the state-of-the-art on the AVA-ActiveSpeaker dataset achieving an mAP of 87.1%. Moreover, ablation studies verify that this result is a direct consequence of our long-term multi-speaker analysis.

In video understanding, action spotting consists in temporally localizing human-induced events annotated with single timestamps. In this paper, we propose a novel loss function that specifically considers the temporal context naturally present around each action, rather than focusing on the single annotated frame to spot. We benchmark our loss on a large dataset of soccer videos, SoccerNet, and achieve an improvement of 12.8% over the baseline. We show the generalization capability of our loss for generic activity proposals and detection on ActivityNet, by spotting the beginning and the end of each activity. Furthermore, we provide an extended ablation study and display challenging cases for action spotting in soccer videos. Finally, we qualitatively illustrate how our loss induces a precise temporal understanding of actions and show how such semantic knowledge can be used for automatic highlights generation.

Architecture design has become a crucial component of successful deep learning. Recent progress in automatic neural architecture search (NAS) shows a lot of promise. However, discovered architectures often fail to generalize in the final evaluation. Architectures with a higher validation accuracy during the search phase may perform worse in the evaluation. Aiming to alleviate this common issue, we introduce sequential greedy architecture search (SGAS), an efficient method for neural architecture search. By dividing the search procedure into sub-problems, SGAS chooses and prunes candidate operations in a greedy fashion. We apply SGAS to search architectures for Convolutional Neural Networks (CNN) and Graph Convolutional Networks (GCN). Extensive experiments show that SGAS is able to find state-of-the-art architectures for tasks such as image classification, point cloud classification and node classification in protein-protein interaction graphs with minimal computational cost.

Temporal action detection is a fundamental yet challenging task in video understanding. Video context is a critical cue to effectively detect actions, but current works mainly focus on temporal context, while neglecting semantic context as well as other important context properties. In this work, we propose a graph convolutional network (GCN) model to adaptively incorporate multi-level semantic context into video features and cast temporal action detection as a sub-graph localization problem. Specifically, we formulate video snippets as graph nodes, snippet-snippet correlations as edges, and actions associated with context as target sub-graphs. With graph convolution as the basic operation, we design a GCN block called GCNeXt, which learns the features of each node by aggregating its context and dynamically updates the edges in the graph. To localize each sub-graph, we also design an SGAlign layer to embed each sub-graph into the Euclidean space. Extensive experiments show that G-TAD is capable of finding effective video context without extra supervision and achieves state-of-the-art performance on two detection benchmarks. On ActivityNet-1.3 it obtains an average mAP of 34.09%; on THUMOS14 it reaches 51.6% at [email protected] when combined with a proposal processing method. The code has been made available at https://github.com/frostinassiky/gtad.

Although deep convolutional neural networks (CNNs) have achieved great success in computer vision tasks, its real-world application is still impeded by its voracious demand of computational resources. Current works mostly seek to compress the network by reducing its parameters or parameter-incurred computation, neglecting the influence of the input image on the system complexity. Based on the fact that input images of a CNN contain substantial redundancy, in this paper, we propose a unified framework, dubbed as ThumbNet, to simultaneously accelerate and compress CNN models by enabling them to infer on one thumbnail image. We provide three effective strategies to train ThumbNet. In doing so, ThumbNet learns an inference network that performs equally well on small images as the original-input network on large images. With ThumbNet, not only do we obtain the thumbnail-input inference network that can drastically reduce computation and memory requirements, but also we obtain an image downscaler that can generate thumbnail images for generic classification tasks. Extensive experiments show the effectiveness of ThumbNet, and demonstrate that the thumbnail-input inference network learned by ThumbNet can adequately retain the accuracy of the original-input network even when the input images are downscaled 16 times.

Visual and audio modalities are highly correlated, yet they contain different information. Their strong correlation makes it possible to predict the semantics of one from the other with good accuracy. Their intrinsic differences make cross-modal prediction a potentially more rewarding pretext task for self-supervised learning of video and audio representations compared to within-modality learning. Based on this intuition, we propose Cross-Modal Deep Clustering (XDC), a novel self-supervised method that leverages unsupervised clustering in one modality (e.g., audio) as a supervisory signal for the other modality (e.g., video). This cross-modal supervision helps XDC utilize the semantic correlation and the differences between the two modalities. Our experiments show that XDC outperforms single-modality clustering and other multi-modal variants. XDC achieves state-of-the-art accuracy among self-supervised methods on multiple video and audio benchmarks. Most importantly, our video model pretrained on large-scale unlabeled data significantly outperforms the same model pretrained with full-supervision on ImageNet and Kinetics for action recognition on HMDB51 and UCF101. To the best of our knowledge, XDC is the first self-supervised learning method that outperforms large-scale fully-supervised pretraining for action recognition on the same architecture.

In popular TV programs (such as CSI), a very low-resolution face image of a person, who is not even looking at the camera in many cases, is digitally super-resolved to a degree that suddenly the person's identity is made visible and recognizable. Of course, we suspect that this is merely a cinematographic special effect and such a magical transformation of a single image is not technically possible. Or, is it? In this paper, we push the boundaries of super-resolving (hallucinating to be more accurate) a tiny, non-frontal face image to understand how much of this is possible by leveraging the availability of large datasets and deep networks. To this end, we introduce a novel Transformative Adversarial Neural Network (TANN) to jointly frontalize very-low resolution (i.e., 16 × 16 pixels) out-of-plane rotated face images (including profile views) and aggressively super-resolve them (8×), regardless of their original poses and without using any 3D information. TANN is composed of two components: a transformative upsampling network which embodies encoding, spatial transformation and deconvolutional layers, and a discriminative network that enforces the generated high-resolution frontal faces to lie on the same manifold as real frontal face images. We evaluate our method on a large set of synthesized non-frontal face images to assess its reconstruction performance. Extensive experiments demonstrate that TANN generates both qualitatively and quantitatively superior results achieving over 4 dB improvement over the state-of-the-art.

Single Photon Avalanche Photodiodes (SPADs) have recently received a lot of attention in imaging and vision applications due to their excellent performance in low-light conditions, as well as their ultra-high temporal resolution. Unfortunately, like many evolving sensor technologies, image sensors built around SPAD technology currently suffer from a low pixel count. In this work, we investigate a simple, low-cost, and compact optical coding camera design that supports high-resolution image reconstructions from raw measurements with low pixel counts. We demonstrate this approach for regular intensity imaging, depth imaging, as well transient imaging. Our method uses an end-to-end framework to simultaneously optimize the optical design and a reconstruction network for obtaining super-resolved images from raw measurements. The optical design space is that of an engineered point spread function (implemented with diffractive optics), which can be considered an optimized anti-aliasing filter to preserve as much high-resolution information as possible despite imaging with a low pixel count, low fill-factor SPAD array. We further investigate a deep network for reconstruction. The effectiveness of this joint design and reconstruction approach is demonstrated for a range of different applications, including high-speed imaging, and time of flight depth imaging, as well as transient imaging. While our work specifically focuses on low-resolution SPAD sensors, similar approaches should prove effective for other emerging image sensor technologies with low pixel counts and low fill-factors.

Deep neural networks have been playing an essential role in many computer vision tasks including Visual Question Answering (VQA). Until recently, the study of their accuracy was the main focus of research but now there is a trend toward assessing the robustness of these models against adversarial attacks by evaluating their tolerance to varying noise levels. In VQA, adversarial attacks can target the image and/or the proposed main question and yet there is a lack of proper analysis of the later. In this work, we propose a flexible framework that focuses on the language part of VQA that uses semantically relevant questions, dubbed basic questions, acting as controllable noise to evaluate the robustness of VQA models. We hypothesize that the level of noise is negatively correlated to the similarity of a basic question to the main question. Hence, to apply noise on any given main question, we rank a pool of basic questions based on their similarity by casting this ranking task as a LASSO optimization problem. Then, we propose a novel robustness measure Rscore and two largescale basic question datasets (BQDs) in order to standardize robustness analysis for VQA models.

Point clouds are challenging to process due to their sparsity, therefore autonomous vehicles rely more on appearance attributes than pure geometric features. However, 3D LIDAR perception can provide crucial information for urban navigation in challenging light or weather conditions. In this paper, we investigate the versatility of Shape Completion for 3D Object Tracking in LIDAR point clouds. We design a Siamese tracker that encodes model and candidate shapes into a compact latent representation. We regularize the encoding by enforcing the latent representation to decode into an object model shape. We observe that 3D object tracking and 3D shape completion complement each other. Learning a more meaningful latent representation shows better discriminatory capabilities, leading to improved tracking performance. We test our method on the KITTI Tracking set using car 3D bounding boxes. Our model reaches a 76.94% Success rate and 81.38% Precision for 3D Object Tracking, with the shape completion regularization leading to an improvement of 3% in both metrics.

We propose a novel method for instance label segmentation of dense 3D voxel grids. We target volumetric scene representations, which have been acquired with depth sensors or multi-view stereo methods and which have been processed with semantic 3D reconstruction or scene completion methods. The main task is to learn shape information about individual object instances in order to accurately separate them, including connected and incompletely scanned objects. We solve the 3D instance-labeling problem with a multi-task learning strategy. The first goal is to learn an abstract feature embedding, which groups voxels with the same instance label close to each other while separating clusters with different instance labels from each other. The second goal is to learn instance information by densely estimating directional information of the instance's center of mass for each voxel. This is particularly useful to find instance boundaries in the clustering post-processing step, as well as, for scoring the segmentation quality for the first goal. Both synthetic and real-world experiments demonstrate the viability and merits of our approach. In fact, it achieves state-of-the-art performance on the ScanNet 3D instance segmentation benchmark.

Convolutional Neural Networks (CNNs) achieve impressive performance in a wide variety of fields. Their success benefited from a massive boost when very deep CNN models were able to be reliably trained. Despite their merits, CNNs fail to properly address problems with non-Euclidean data. To overcome this challenge, Graph Convolutional Networks (GCNs) build graphs to represent non-Euclidean data, borrow concepts from CNNs, and apply them in training. GCNs show promising results, but they are usually limited to very shallow models due to the vanishing gradient problem. As a result, most state-of-the-art GCN models are no deeper than 3 or 4 layers. In this work, we present new ways to successfully train very deep GCNs. We do this by borrowing concepts from CNNs, specifically residual/dense connections and dilated convolutions, and adapting them to GCN architectures. Extensive experiments show the positive effect of these deep GCN frameworks. Finally, we use these new concepts to build a very deep 56-layer GCN, and show how it significantly boosts performance (+3.7% mIoU over state-of-the-art) in the task of point cloud semantic segmentation. We believe that the community can greatly benefit from this work, as it opens up many opportunities for advancing GCN-based research.

Color correction and color transfer methods have gained a lot of attention in the past few years to circumvent color degradation that may occur due to various sources. In this paper, we propose a novel simple yet powerful strategy to profoundly enhance color distorted underwater images. The proposed approach combines both local and global information through a simple yet powerful affine transform model. Local and global information are carried through local color mapping and color covariance mapping between an input and some reference source, respectively. Several experiments on degraded underwater images demonstrate that the proposed method performs favourably to all other methods including ones that are tailored to correcting underwater images by explicit noise modelling.

This paper revisits the integer programming (IP) problem, which plays a fundamental role in many computer vision and machine learning applications. The literature abounds with many seminal works that address this problem, some focusing on continuous approaches (e.g., linear program relaxation), while others on discrete ones (e.g., min-cut). However, since many of these methods are designed to solve specific IP forms, they cannot adequately satisfy the simultaneous requirements of accuracy, feasibility, and scalability. To this end, we propose a novel and versatile framework called $\ell _p$ℓp-box ADMM, which is based on two main ideas. (1) The discrete constraint is equivalently replaced by the intersection of a box and an $\ell _p$ℓp-norm sphere. (2) We infuse this equivalence into the Alternating Direction Method of Multipliers (ADMM) framework to handle the continuous constraints separately and to harness its attractive properties. More importantly, the ADMM update steps can lead to manageable sub-problems in the continuous domain. To demonstrate its efficacy, we apply it to an optimization form that occurs often in computer vision and machine learning, namely binary quadratic programming (BQP). In this case, the ADMM steps are simple, computationally efficient. Moreover, we present the theoretic analysis about the global convergence of the $\ell _p$ℓp-box ADMM through adding a perturbation with the sufficiently small factor $\epsilon$ε to the original IP problem. Specifically, the globally converged solution generated by $\ell _p$ℓp-box ADMM for the perturbed IP problem will be close to the stationary and feasible point of the original IP problem within $O(\epsilon)$O(ε). We demonstrate the applicability of $\ell _p$ℓp-box ADMM on three important applications: MRF energy minimization, graph matching, and clustering. Results clearly show that it significantly outperforms existing generic IP solvers both in runtime and objective. It also achieves very competitive performance to state-of-the-art methods designed specifically for these applications.

Total Variation (TV) is an effective and popular prior model in the field of regularization-based image processing. This paper focuses on total variation for removing impulse noise in image restoration. This type of noise frequently arises in data acquisition and transmission due to many reasons, e.g., a faulty sensor or analog-to-digital converter errors. Removing this noise is an important task in image restoration. State-of-the-art methods such as Adaptive Outlier Pursuit(AOP)[1], which is based on TV with$\ell _{02}$-norm data fidelity, only give sub-optimal performance. In this paper, we propose a new sparse optimization method, called$\ell _0TV$-PADMM, which solves the TV-based restoration problem with$\ell _0$-norm data fidelity. To effectively deal with the resulting non-convex non-smooth optimization problem, we first reformulate it as an equivalent biconvex Mathematical Program with Equilibrium Constraints (MPEC), and then solve it using a proximal Alternating Direction Method of Multipliers (PADMM). Our$\ell _0TV$-PADMM method finds a desirable solution to the original$\ell _0$-norm optimization problem and is proven to be convergent under mild conditions. We apply$\ell _0TV$-PADMM to the problems of image denoising and deblurring in the presence of impulse noise. Our extensive experiments demonstrate that$\ell _0TV$-PADMM outperforms state-of-the-art image restoration methods.

Face detection techniques have been developed for decades, and one of remaining open challenges is detecting small faces in unconstrained conditions. The reason is that tiny faces are often lacking detailed information and blurring. In this paper, we proposed an algorithm to directly generate a clear high-resolution face from a blurry small one by adopting a generative adversarial network (GAN). Toward this end, the basic GAN formulation achieves it by super-resolving and refining sequentially (e.g. SR-GAN and cycle-GAN). However, we design a novel network to address the problem of super-resolving and refining jointly. We also introduce new training losses to guide the generator network to recover fine details and to promote the discriminator network to distinguish real vs. fake and face vs. non-face simultaneously. Extensive experiments on the challenging dataset WIDER FACE demonstrate the effectiveness of our proposed method in restoring a clear high-resolution face from a blurry small one, and show that the detection performance outperforms other state-of-the-art methods.

The outstanding performance of deep neural networks (DNNs), for the visual recognition task in particular, has been demonstrated on several large-scale benchmarks. This performance has immensely strengthened the line of research that aims to understand and analyze the driving reasons behind the effectiveness of these networks. One important aspect of this analysis has recently gained much attention, namely the reaction of a DNN to noisy input. This has spawned research on developing adversarial input attacks as well as training strategies that make DNNs more robust against these attacks. To this end, we derive in this paper exact analytic expressions for the first and second moments (mean and variance) of a small piecewise linear (PL) network (Affine, ReLU, Affine) subject to general Gaussian input. We experimentally show that these expressions are tight under simple linearizations of deeper PL-DNNs, especially popular architectures in the literature (e.g. LeNet and AlexNet). Extensive experiments on image classification show that these expressions can be used to study the behaviour of the output mean of the logits for each class, the interclass confusion and the pixel-level spatial noise sensitivity of the network. Moreover, we show how these expressions can be used to systematically construct targeted and non-targeted adversarial attacks.

In this work we propose a new automatic image annotation model, dubbed diverse and distinct image annotation (D2IA). The generative model D2IA is inspired by the ensemble of human annotations, which create semantically relevant, yet distinct and diverse tags. In D2IA, we generate a relevant and distinct tag subset, in which the tags are relevant to the image contents and semantically distinct to each other, using sequential sampling from a determinantal point process (DPP) model. Multiple such tag subsets that cover diverse semantic aspects or diverse semantic levels of the image contents are generated by randomly perturbing the DPP sampling process. We leverage a generative adversarial network (GAN) model to train D2IA. Extensive experiments including quantitative and qualitative comparisons, as well as human subject studies, on two benchmark datasets demonstrate that the proposed model can produce more diverse and distinct tags than the state-of-the-arts.

Weakly-supervised object detection has attracted much attention lately, since it does not require bounding box annotations for training. Although significant progress has also been made, there is still a large gap in performance between weakly-supervised and fully-supervised object detection. Recently, some works use pseudo ground-truths which are generated by a weakly-supervised detector to train a supervised detector. Such approaches incline to find the most representative parts of objects, and only seek one ground-truth box per class even though many same-class instances exist. To overcome these issues, we propose a weakly-supervised to fully-supervised framework, where a weakly-supervised detector is implemented using multiple instance learning. Then, we propose a pseudo ground-truth excavation (PGE) algorithm to find the pseudo ground-truth of each instance in the image. Moreover, the pseudo ground-truth adaptation (PGA) algorithm is designed to further refine the pseudo ground-truths from PGE. Finally, we use these pseudo ground-truths to train a fully-supervised detector. Extensive experiments on the challenging PASCAL VOC 2007 and 2012 benchmarks strongly demonstrate the effectiveness of our framework. We obtain 52.4% and 47.8% mAP on VOC2007 and VOC2012 respectively, a significant improvement over previous state-of-the-art methods.

With the aim of developing a fast yet accurate algorithm for compressive sensing (CS) reconstruction of natural images, we combine in this paper the merits of two existing categories of CS methods: the structure insights of traditional optimization-based methods and the speed of recent network-based ones. Specifically, we propose a novel structured deep network, dubbed ISTA-Net, which is inspired by the Iterative Shrinkage-Thresholding Algorithm (ISTA)for optimizing a general ℓ1norm CS reconstruction model. To cast ISTA into deep network form, we develop an effective strategy to solve the proximal mapping associated with the sparsity-inducing regularizer using nonlinear transforms. All the parameters in ISTA-Net (e.g. nonlinear transforms, shrinkage thresholds, step sizes, etc.) are learned end-to-end, rather than being hand-crafted. Moreover, considering that the residuals of natural images are more compressible, an enhanced version of ISTA-Net in the residual domain, dubbed ISTA-Net+, is derived to further improve CS reconstruction. Extensive CS experiments demonstrate that the proposed ISTA-Nets outperform existing state-of-the-art optimization-based and networkbased CS methods by large margins, while maintaining fast computational speed. Our source codes are available: http://jianzhang.tech/projects/ISTA-Net.

Binary optimization is a central problem in mathematical optimization and its applications are abundant. To solve this problem, we propose a new class of continuous optimization techniques, which is based on Mathematical Programming with Equilibrium Constraints (MPECs). We first reformulate the binary program as an equivalent augmented biconvex optimization problem with a bilinear equality constraint, then we propose an exact penalty method to solve it. The resulting algorithm seeks a desirable solution to the original problem via solving a sequence of linear programming convex relaxation subproblems. In addition, we prove that the penalty function, induced by adding the complementarity constraint to the objective, is exact, i.e., it has the same local and global minima with those of the original binary program when the penalty parameter is over some threshold. The convergence of the algorithm can be guaranteed, since it essentially reduces to block coordinate descent in the literature. Finally, we demonstrate the effectiveness of our method on the problem of dense subgraph discovery. Extensive experiments show that our method outperforms existing techniques, such as iterative hard thresholding and linear programming relaxation.

In this paper, we revisit the LASSO sparse representation problem, which has been studied and used in a variety of different areas, ranging from signal processing and information theory to computer vision and machine learning. In the vision community, it found its way into many important applications, including face recognition, tracking, super resolution, image denoising, to name a few. Despite advances in efficient sparse algorithms, solving large-scale LASSO problems remains a challenge. To circumvent this difficulty, people tend to downsample and subsample the problem (e.g. via dimensionality reduction) to maintain a manageable sized LASSO, which usually comes at the cost of losing solution accuracy. This paper proposes a novel circulant reformulation of the LASSO that lifts the problem to a higher dimension, where ADMM can be efficiently applied to its dual form. Because of this lifting, all optimization variables are updated using only basic element-wise operations, the most computationally expensive of which is a 1D FFT. In this way, there is no need for a linear system solver nor matrix-vector multiplication. Since all operations in our FFTLasso method are element-wise, the subproblems are completely independent and can be trivially parallelized (e.g. on a GPU). The attractive computational properties of FFTLasso are verified by extensive experiments on synthetic and real data and on the face recognition task. They demonstrate that FFTLasso scales much more effectively than a state-of-the-art solver.

Our paper presents a new approach for temporal detection of human actions in long, untrimmed video sequences. We introduce Single-Stream Temporal Action Proposals (SST), a new effective and efficient deep architecture for the generation of temporal action proposals. Our network can run continuously in a single stream over very long input video sequences, without the need to divide input into short overlapping clips or temporal windows for batch processing. We demonstrate empirically that our model outperforms the state-of-the-art on the task of temporal action proposal generation, while achieving some of the fastest processing speeds in the literature. Finally, we demonstrate that using SST proposals in conjunction with existing action classifiers results in improved state-of-the-art temporal action detection performance.

Despite the recent advances in large-scale video analysis, action detection remains as one of the most challenging unsolved problems in computer vision. This snag is in part due to the large volume of data that needs to be analyzed to detect actions in videos. Existing approaches have mitigated the computational cost, but still, these methods lack rich high-level semantics that helps them to localize the actions quickly. In this paper, we introduce a Semantic Cascade Context (SCC) model that aims to detect action in long video sequences. By embracing semantic priors associated with human activities, SCC produces high-quality class-specific action proposals and prune unrelated activities in a cascade fashion. Experimental results in ActivityNet unveils that SCC achieves state-of-the-art performance for action detection while operating at real time.

Correlation filter (CF) based trackers have recently gained a lot of popularity due to their impressive performance on benchmark datasets, while maintaining high frame rates. A significant amount of recent research focuses on the incorporation of stronger features for a richer representation of the tracking target. However, this only helps to discriminate the target from background within a small neighborhood. In this paper, we present a framework that allows the explicit incorporation of global context within CF trackers. We reformulate the original optimization problem and provide a closed form solution for single and multi-dimensional features in the primal and dual domain. Extensive experiments demonstrate that this framework significantly improves the performance of many CF trackers with only a modest impact on frame rate.

In this work we study the task of image annotation, of which the goal is to describe an image using a few tags. Instead of predicting the full list of tags, here we target for providing a short list of tags under a limited number (e.g., 3), to cover as much information as possible of the image. The tags in such a short list should be representative and diverse. It means they are required to be not only corresponding to the contents of the image, but also be different to each other. To this end, we treat the image annotation as a subset selection problem based on the conditional determinantal point process (DPP) model, which formulates the representation and diversity jointly. We further explore the semantic hierarchy and synonyms among the candidate tags, and require that two tags in a semantic hierarchy or in a pair of synonyms should not be selected simultaneously. This requirement is then embedded into the sampling algorithm according to the learned conditional DPP model. Besides, we find that traditional metrics for image annotation (e.g., precision, recall and F1 score) only consider the representation, but ignore the diversity. Thus we propose new metrics to evaluate the quality of the selected subset (i.e., the tag list), based on the semantic hierarchy and synonyms. Human study through Amazon Mechanical Turk verifies that the proposed metrics are more close to the humans judgment than traditional metrics. Experiments on two benchmark datasets show that the proposed method can produce more representative and diverse tags, compared with existing image annotation methods.

Composite function minimization captures a wide spectrum of applications in both computer vision and machine learning. It includes bound constrained optimization and cardinality regularized optimization as special cases. This paper proposes and analyzes a new Matrix Splitting Method (MSM) for minimizing composite functions. It can be viewed as a generalization of the classical Gauss-Seidel method and the Successive Over-Relaxation method for solving linear systems in the literature. Incorporating a new Gaussian elimination procedure, the matrix splitting method achieves state-of-the-art performance. For convex problems, we establish the global convergence, convergence rate, and iteration complexity of MSM, while for non-convex problems, we prove its global convergence. Finally, we validate the performance of our matrix splitting method on two particular applications: nonnegative matrix factorization and cardinality regularized sparse coding. Extensive experiments show that our method outperforms existing composite function minimization techniques in term of both efficiency and efficacy.

Convolutional sparse coding (CSC) has gained attention for its successful role as a reconstruction and a classification tool in the computer vision and machine learning community. Current CSC methods can only reconstruct singlefeature 2D images independently. However, learning multidimensional dictionaries and sparse codes for the reconstruction of multi-dimensional data is very important, as it examines correlations among all the data jointly. This provides more capacity for the learned dictionaries to better reconstruct data. In this paper, we propose a generic and novel formulation for the CSC problem that can handle an arbitrary order tensor of data. Backed with experimental results, our proposed formulation can not only tackle applications that are not possible with standard CSC solvers, including colored video reconstruction (5D- tensors), but it also performs favorably in reconstruction with much fewer parameters as compared to naive extensions of standard CSC to multiple features/channels.

In this paper, we present a technique that places 3D bounding boxes around objects in an RGB-D scene. Our approach makes best use of the 2D information to quickly reduce the search space in 3D, benefiting from state-of-the-art 2D object detection techniques. We then use the 3D information to orient, place, and score bounding boxes around objects. We independently estimate the orientation for every object, using previous techniques that utilize normal information. Object locations and sizes in 3D are learned using a multilayer perceptron (MLP). In the final step, we refine our detections based on object class relations within a scene. When compared to state-of-the-art detection methods that operate almost entirely in the sparse 3D domain, extensive experiments on the well-known SUN RGB-D dataset [29] show that our proposed method is much faster (4.1s per image) in detecting 3D objects in RGB-D images and performs better (3 mAP higher) than the state-of-the-art method that is 4.7 times slower and comparably to the method that is two orders of magnitude slower. This work hints at the idea that 2D-driven object detection in 3D should be further explored, especially in cases where the 3D input is sparse.

Convolutional sparse coding (CSC) plays an essential role in many computer vision applications ranging from image compression to deep learning. In this work, we spot the light on a new application where CSC can effectively serve, namely line drawing analysis. The process of drawing a line drawing can be approximated as the sparse spatial localization of a number of typical basic strokes, which in turn can be cast as a non-standard CSC model that considers the line drawing formation process from parametric curves. These curves are learned to optimize the fit between the model and a specific set of line drawings. Parametric representation of sketches is vital in enabling automatic sketch analysis, synthesis and manipulation. A couple of sketch manipulation examples are demonstrated in this work. Consequently, our novel method is expected to provide a reliable and automatic method for parametric sketch description. Through experiments, we empirically validate the convergence of our method to a feasible solution.

In multi-label learning, there are two main challenges: missing labels and class imbalance (CIB). The former assumes that only a partial set of labels are provided for each training instance while other labels are missing. CIB is observed from two perspectives: first, the number of negative labels of each instance is much larger than its positive labels; second, the rate of positive instances (i.e. the number of positive instances divided by the total number of instances) of different classes are significantly different. Both missing labels and CIB lead to significant performance degradation. In this work, we propose a new method to handle these two challenges simultaneously. We formulate the problem as a constrained submodular minimization that is composed of a submodular objective function that encourages label consistency and smoothness, as well as, class cardinality bound constraints to handle class imbalance. We further present a convex approximation based on the Lovasz extension of submodular functions, leading to a linear program, which can be efficiently solved by the alternative direction method of multipliers (ADMM). Experimental results on several benchmark datasets demonstrate the improved performance of our method over several state-of-the-art methods.

Semi-definite rank minimization problems model a wide range of applications in both signal processing and machine learning fields. This class of problem is NP-hard in general. In this paper, we propose a proximal Alternating Direction Method (ADM) for the well-known semi-definite rank regularized minimization problem. Specifically, we first reformulate this NP-hard problem as an equivalent biconvex MPEC (Mathematical Program with Equilibrium Constraints), and then solve it using proximal ADM, which involves solving a sequence of structured convex semi-definite subproblems to find a desirable solution to the original rank regularized optimization problem. Moreover, based on the Kurdyka-Lojasiewicz inequality, we prove that the proposed method always converges to a KKT stationary point under mild conditions. We apply the proposed method to the widely studied and popular sensor network localization problem. Our extensive experiments demonstrate that the proposed algorithm outperforms state-of-the-art low-rank semi-definite minimization algorithms in terms of solution quality.

In this paper, we present a part-based sparse tracker in a particle filter framework where both the motion and appearance model are formulated in 3D. The motion model is adaptive and directed according to a simple yet powerful occlusion handling paradigm, which is intrinsically fused in the motion model. Also, since 3D trackers are sensitive to synchronization and registration noise in the RGB and depth streams, we propose automated methods to solve these two issues. Extensive experiments are conducted on a popular RGBD tracking benchmark, which demonstrate that our tracker can achieve superior results, outperforming many other recent and state-of-the-art RGBD trackers.

In many large-scale video analysis scenarios, one is interested in localizing and recognizing human activities that occur in short temporal intervals within long untrimmed videos. Current approaches for activity detection still struggle to handle large-scale video collections and the task remains relatively unexplored. This is in part due to the computational complexity of current action recognition approaches and the lack of a method that proposes fewer intervals in the video, where activity processing can be focused. In this paper, we introduce a proposal method that aims to recover temporal segments containing actions in untrimmed videos. Building on techniques for learning sparse dictionaries, we introduce a learning framework to represent and retrieve activity proposals. We demonstrate the capabilities of our method in not only producing high quality proposals but also in its efficiency. Finally, we show the positive impact our method has on recognition performance when it is used for action detection, while running at 10FPS.

Sparse representation has been introduced to visual tracking by finding the best target candidate with minimal reconstruction error within the particle filter framework. However, most sparse representation based trackers have high computational cost, less than promising tracking performance, and limited feature representation. To deal with the above issues, we propose a novel circulant sparse tracker (CST), which exploits circulant target templates. Because of the circulant structure property, CST has the following advantages: (1) It can refine and reduce particles using circular shifts of target templates. (2) The optimization can be efficiently solved entirely in the Fourier domain. (3) High dimensional features can be embedded into CST to significantly improve tracking performance without sacrificing much computation time. Both qualitative and quantitative evaluations on challenging benchmark sequences demonstrate that CST performs better than all other sparse trackers and favorably against state-of-the-art methods.

In this paper, we propose a persistent, robust and autonomous object tracking system for unmanned aerial vehicles (UAVs) called Persistent Aerial Tracking (PAT). A computer vision and control strategy is applied to a diverse set of moving objects (e.g. humans, animals, cars, boats, etc.) integrating multiple UAVs with a stabilized RGB camera. A novel strategy is employed to successfully track objects over a long period, by `handing over the camera' from one UAV to another. We evaluate several state-of-the-art trackers on the VIVID aerial video dataset and additional sequences that are specifically tailored to low altitude UAV target tracking. Based on the evaluation, we select the leading tracker and improve upon it by optimizing for both speed and performance, integrate the complete system into an off-the-shelf UAV, and obtain promising results showing the robustness of our solution in real-world aerial scenarios.

Abstract Are simple strokes unique to the artist or designer who renders them? If so, can this idea be used to identify authorship or to classify artistic drawings? Also, could training methods be devised to develop particular styles? To answer these questions, we propose the Stroke Authorship Recognition (SAR) approach, a novel method that distinguishes the authorship of 2D digitized drawings. SAR converts a drawing into a histogram of stroke attributes that is discriminative of authorship. We provide extensive classification experiments on a large variety of data sets, which validate SAR's ability to distinguish unique authorship of artists and designers. We also demonstrate the usefulness of SAR in several applications including the detection of fraudulent sketches, the training and monitoring of artists in learning a particular new style and the first quantitative way to measure the quality of automatic sketch synthesis tools.

In this paper, we formulate particle filter-based object tracking as an exclusive sparse learning problem that exploits contextual information. To achieve this goal, we propose the context-aware exclusive sparse tracker (CEST) to model particle appearances as linear combinations of dictionary templates that are updated dynamically. Learning the representation of each particle is formulated as an exclusive sparse representation problem, where the overall dictionary is composed of multiple group dictionaries that can contain contextual information. With context, CEST is less prone to tracker drift. Interestingly, we show that the popular L1 tracker is a special case of our CEST formulation. The proposed learning problem is efficiently solved using an accelerated proximal gradient method that yields a sequence of closed form updates. To make the tracker much faster, we reduce the number of learning problems to be solved by using the dual problem to quickly and systematically rank and prune particles in each frame. We test our CEST tracker on challenging benchmark sequences that involve heavy occlusion, drastic illumination changes, and large pose variations. Experimental results show that CEST consistently outperforms state-of-the-art trackers.

One of the cornerstone principles of deep models is their abstraction capacity, i.e. their ability to learn abstract concepts from `simpler' ones. Through extensive experiments, we characterize the nature of the relationship between abstract concepts (specifically objects in images) learned by popular and high performing convolutional networks (conv-nets) and established mid-level representations used in computer vision (specifically semantic visual attributes). We focus on attributes due to their impact on several applications, such as object description, retrieval and mining, and active (and zero-shot) learning. Among the findings we uncover, we show empirical evidence of the existence of Attribute Centric Nodes (ACNs) within a conv-net, which is trained to recognize objects (not attributes) in images. These special conv-net nodes (1) collectively encode information pertinent to visual attribute representation and discrimination, (2) are unevenly and sparsely distribution across all layers of the conv-net, and (3) play an important role in conv-net based object recognition.

This paper proposes a new framework for estimating the Manhattan Frame (MF) of an indoor scene from a single RGB-D image. Our technique formulates this problem as the estimation of a rotation matrix that best aligns the normals of the captured scene to a canonical world axes. By introducing sparsity constraints, our method can simultaneously estimate the scene MF, the surfaces in the scene that are best aligned to one of three coordinate axes, and the outlier surfaces that do not align with any of the axes. To test our approach, we contribute a new set of annotations to determine ground truth MFs in each image of the popular NYUv2 dataset. We use this new benchmark to experimentally demonstrate that our method is more accurate, faster, more reliable and more robust than the methods used in the literature. We further motivate our technique by showing how it can be used to address the RGB-D SLAM problem in indoor scenes by incorporating it into and improving the performance of a popular RGB-D SLAM method.

In spite of many dataset efforts for human action recognition, current computer vision algorithms are still severely limited in terms of the variability and complexity of the actions that they can recognize. This is in part due to the simplicity of current benchmarks, which mostly focus on simple actions and movements occurring on manually trimmed videos. In this paper we introduce ActivityNet, a new large-scale video benchmark for human activity understanding. Our benchmark aims at covering a wide range of complex human activities that are of interest to people in their daily living. In its current version, ActivityNet provides samples from 203 activity classes with an average of 137 untrimmed videos per class and 1.41 activity instances per video, for a total of 849 video hours. We illustrate three scenarios in which ActivityNet can be used to compare algorithms for human activity understanding: untrimmed video classification, trimmed activity classification and activity detection.

Total Variation (TV) is an effective and popular prior model in the field of regularization-based image processing. This paper focuses on TV for image restoration in the presence of impulse noise. This type of noise frequently arises in data acquisition and transmission due to many reasons, e.g. a faulty sensor or analog-to-digital converter errors. Removing this noise is an important task in image restoration. State-of-the-art methods such as Adaptive Outlier Pursuit(AOP) [42], which is based on TV with ℓ02-norm data fidelity, only give sub-optimal performance. In this paper, we propose a new method, called ℓ0TV -PADMM, which solves the TV-based restoration problem with ℓ0-norm data fidelity. To effectively deal with the resulting non-convex non-smooth optimization problem, we first reformulate it as an equivalent MPEC (Mathematical Program with Equilibrium Constraints), and then solve it using a proximal Alternating Direction Method of Multipliers (PADMM). Our ℓ0TV -PADMM method finds a desirable solution to the original ℓ0-norm optimization problem and is proven to be convergent under mild conditions. We apply ℓ0TV -PADMM to the problems of image denoising and deblurring in the presence of impulse noise. Our extensive experiments demonstrate that ℓ0TV -PADMM outperforms state-of-the-art image restoration methods.

Sparse representation has been applied to visual tracking by finding the best target candidate with minimal reconstruction error by use of target templates. However, most sparse representation based trackers only consider holistic or local representations and do not make full use of the intrinsic structure among and inside target candidates, thereby making the representation less effective when similar objects appear or under occlusion. In this paper, we propose a novel Structural Sparse Tracking (SST) algorithm, which not only exploits the intrinsic relationship among target candidates and their local patches to learn their sparse representations jointly, but also preserves the spatial layout structure among the local patches inside each target candidate. We show that our SST algorithm accommodates most existing sparse trackers with the respective merits. Both qualitative and quantitative evaluations on challenging benchmark image sequences demonstrate that the proposed SST algorithm performs favorably against several state-of-the-art methods.

Recent studies on image memorability have shed light on what distinguishes the memorability of different images and the intrinsic and extrinsic properties that make those images memorable. However, a clear understanding of the memorability of specific objects inside an image remains elusive. In this paper, we provide the first attempt to answer the question: what exactly is remembered about an image? We augment both the images and object segmentations from the PASCAL-S dataset with ground truth memorability scores and shed light on the various factors and properties that make an object memorable (or forgettable) to humans. We analyze various visual factors that may influence object memorability (e.g. color, visual saliency, and object categories). We also study the correlation between object and image memorability and find that image memorability is greatly affected by the memorability of its most memorable object. Lastly, we explore the effectiveness of deep learning and other computational approaches in predicting object memorability in images. Our efforts offer a deeper understanding of memorability in general thereby opening up avenues for a wide variety of applications.

In this paper, we address the problem of computing an intrinsic decomposition of the colors of a surface into an albedo and a shading term. The surface is reconstructed from a single or multiple RGB-D images of a static scene obtained from different views. We thereby extend and improve existing works in the area of intrinsic image decomposition. In a variational framework, we formulate the problem as a minimization of an energy composed of two terms: a data term and a regularity term. The first term is related to the image formation process and expresses the relation between the albedo, the surface normals, and the incident illumination. We use an affine shading model, a combination of a Lambertian model, and an ambient lighting term. This model is relevant for Lambertian surfaces. When available, multiple views can be used to handle view-dependent non-Lambertian reflections. The second term contains an efficient combination of l2and l1-regularizers on the illumination vector field and albedo respectively. Unlike most previous approaches, especially Retinex-like techniques, these terms do not depend on the image gradient or texture, thus reducing the mixing shading/reflectance artifacts and leading to better results. The obtained non-linear optimization problem is efficiently solved using a cyclic block coordinate descent algorithm. Our method outperforms a range of state-of-the-art algorithms on a popular benchmark dataset.

This work focuses on the problem of multi-label learning with missing labels (MLML), which aims to label each test instance with multiple class labels given training instances that have an incomplete/partial set of these labels (i.e. some of their labels are missing). To handle missing labels, we propose a unified model of label dependencies by constructing a mixed graph, which jointly incorporates (i) instance-level similarity and class co-occurrence as undirected edges and (ii) semantic label hierarchy as directed edges. Unlike most MLML methods, We formulate this learning problem transductively as a convex quadratic matrix optimization problem that encourages training label consistency and encodes both types of label dependencies (i.e. undirected and directed edges) using quadratic terms and hard linear constraints. The alternating direction method of multipliers (ADMM) can be used to exactly and efficiently solve this problem. To evaluate our proposed method, we consider two popular applications (image and video annotation), where the label hierarchy can be derived from Wordnet. Experimental results show that our method achieves a significant improvement over state-of-the-art methods in performance and robustness to missing labels.

In this paper, we develop a novel framework for action recognition in videos. The framework is based on automatically learning the discriminative trajectory groups that are relevant to an action. Different from previous approaches, our method does not require complex computation for graph matching or complex latent models to localize the parts. We model a video as a structured bag of trajectory groups with latent class variables. We model action recognition problem in a weakly supervised setting and learn discriminative trajectory groups by employing multiple instance learning (MIL) based Support Vector Machine (SVM) using pre-computed kernels. The kernels depend on the spatio-temporal relationship between the extracted trajectory groups and their associated features. We demonstrate both quantitatively and qualitatively that the classification performance of our proposed method is superior to baselines and several state-of-the-art approaches on three challenging standard benchmark datasets.

Abstract In this paper, we study the problem of automatic camera placement for computer graphics and computer vision applications. We extend the problem formulations of previous work by proposing a novel way to incorporate visibility constraints and camera‐to‐camera relationships. For example, the placement solution can be encouraged to have cameras that image the same important locations from different viewing directions, which can enable reconstruction and surveillance tasks to perform better. We show that the general camera placement problem can be formulated mathematically as a convex binary quadratic program (BQP) under linear constraints. Moreover, we propose an optimization strategy with a favorable trade‐off between speed and solution quality. Our solution is almost as fast as a greedy treatment of the problem, but the quality is significantly higher, so much so that it is comparable to exact solutions that take orders of magnitude more computation time. Because it is computationally attractive, our method also allows users to explore the space of solutions for variations in input parameters. To evaluate its effectiveness, we show a range of 3D results on real‐world floorplans (garage, hotel, mall, and airport).

Abstract We propose a new framework to reconstruct building details by automatically assembling 3D templates on coarse textured building models. In a preprocessing step, we generate an initial coarse model to approximate a point cloud computed using Structure from Motion and Multi View Stereo, and we model a set of 3D templates of facade details. Next, we optimize the initial coarse model to enforce consistency between geometry and appearance (texture images). Then, building details are reconstructed by assembling templates on the textured faces of the coarse model. The 3D templates are automatically chosen and located by our optimization‐based template assembly algorithm that balances image matching and structural regularity. In the results, we demonstrate how our framework can enrich the details of coarse models using various data sets.

In this paper, we explore the use of a semi-supervised manifold alignment method for domain adaptation in the context of human body and head pose estimation in videos. We build upon an existing state-of-the-art system that leverages on external labelled datasets for the body and head features, and on the unlabelled test data with weak velocity labels to do a coupled estimation of the body and head pose. While this previous approach showed promising results, the learning of the underlying manifold structure of the features in the train and target data and the need to align them were not explored despite the fact that the pose features between two datasets may vary according to the scene, e.g. due to different camera point of view or perspective. In this paper, we propose to use a semi-supervised manifold alignment method to bring the train and target samples closer within the resulting embedded space. To this end, we consider an adaptation set from the target data and rely on (weak) labels, given for example by the velocity direction whenever they are reliable. These labels, along with the training labels are used to bias the manifold distance within each manifold and to establish correspondences for alignment.

We address the problem of modeling and classifying American Football offense \nteams’ plays in video, a challenging example of group activity analysis. Automatic play \nclassification will allow coaches to infer patterns and tendencies of opponents more ef- \nficiently, resulting in better strategy planning in a game. We define a football play as a \nunique combination of player trajectories. To this end, we develop a framework that uses \nplayer trajectories as inputs to MedLDA, a supervised topic model. The joint maximiza- \ntion of both likelihood and inter-class margins of MedLDA in learning the topics allows \nus to learn semantically meaningful play type templates, as well as, classify different \nplay types with 70% average accuracy. Furthermore, this method is extended to analyze \nindividual player roles in classifying each play type. We validate our method on a large \ndataset comprising 271 play clips from real-world football games, which will be made \npublicly available for future comparisons.

In this paper, we propose a low-rank sparse coding (LRSC) method that exploits local structure information among features in an image for the purpose of image-level classification. LRSC represents densely sampled SIFT descriptors, in a spatial neighborhood, collectively as low-rank, sparse linear combinations of code words. As such, it casts the feature coding problem as a low-rank matrix learning problem, which is different from previous methods that encode features independently. This LRSC has a number of attractive properties. (1) It encourages sparsity in feature codes, locality in codebook construction, and low-rankness for spatial consistency. (2) LRSC encodes local features jointly by considering their low-rank structure information, and is computationally attractive. We evaluate the LRSC by comparing its performance on a set of challenging benchmarks with that of 7 popular coding and other state-of-the-art methods. Our experiments show that by representing local features jointly, LRSC not only outperforms the state-of-the-art in classification accuracy but also improves the time complexity of methods that use a similar sparse linear representation model for feature coding.

In this paper, we formulate object tracking in a particle filter framework as a multi-task sparse learning problem, which we denote as Multi-Task Tracking (MTT). Since we model particles as linear combinations of dictionary templates that are updated dynamically, learning the representation of each particle is considered a single task in MTT. By employing popular sparsity-inducing ℓp, qmixed norms (p ∈ {2, ∞} and q = 1), we regularize the representation problem to enforce joint sparsity and learn the particle representations together. As compared to previous methods that handle particles independently, our results demonstrate that mining the interdependencies between particles improves tracking performance and overall computational complexity. Interestingly, we show that the popular L1tracker [15] is a special case of our MTT formulation (denoted as the L11tracker) when p = q = 1. The learning problem can be efficiently solved using an Accelerated Proximal Gradient (APG) method that yields a sequence of closed form updates. As such, MTT is computationally attractive. We test our proposed approach on challenging sequences involving heavy occlusion, drastic illumination changes, and large pose variations. Experimental results show that MTT methods consistently outperform state-of-the-art trackers.

Multiple player tracking is one of the main building blocks needed in a sports video analysis system. In an uncalibrated camera setting, robust mutli-object tracking can be very difficult due to a number of reasons including the presence of noise, occlusion, fast camera motion, low-resolution image capture, varying viewpoints and illumination changes. To address the problem of multi-object tracking in sports videos, we go beyond the video frame domain and make use of information in a homography transform domain that is denoted the homography field domain. We propose a novel particle filter based tracking algorithm that uses both object appearance information (e.g. color and shape) in the image domain and cross-domain contextual information in the field domain to improve object tracking. In the field domain, the effect of fast camera motion is significantly alleviated since the underlying homography transform from each frame to the field domain can be accurately estimated. We use contextual trajectory information (intra-trajectory and inter-trajectory context) to further improve the prediction of object states within an particle filter framework. Here, intra-trajectory contextual information is based on history tracking results in the field domain, while inter-trajectory contextual information is extracted from a compiled trajectory dataset based on tracks computed from videos depicting the same sport. Experimental results on real world sports data show that our system is able to effectively and robustly track a variable number of targets regardless of background clutter, camera motion and frequent mutual occlusion between targets.

Given a sequence of observable features of a linear dynamical system (LDS), we propose the problem of finding a representation of the LDS which is sparse in terms of a given dictionary of LDSs. Since LDSs do not belong to Euclidean space, traditional sparse coding techniques do not apply. We propose a probabilistic framework and an efficient MAP algorithm to learn this sparse code. Since dynamic textures (DTs) can be modeled as LDSs, we validate our framework and algorithm by applying them to the problems of DT representation and DT recognition. In the case of occlusion, we show that this sparse coding scheme outperforms conventional DT recognition methods.

Given the video of a still background occluded by a fluid dynamic texture (FDT), this paper addresses the problem of separating the video sequence into its two constituent layers. One layer corresponds to the video of the unoccluded background, and the other to that of the dynamic texture, as it would appear if viewed against a black background. The model of the dynamic texture is unknown except that it represents fluid flow. We present an approach that uses the image motion information to simultaneously obtain a model of the dynamic texture and separate it from the background which is required to be still. Previous methods have considered occluding layers whose dynamics follows simple motion models (e.g. periodic or 2D parametric motion). FDTs considered in this paper exhibit complex stochastic motion. We consider videos showing an FDT layer (e.g. pummeling smoke or heavy rain) in front of a static background layer (e.g. brick building). We propose a novel method for simultaneously separating these two layers and learning a model for the FDT. Due to the fluid nature of the DT, we are required to learn a model for both the spatial appearance and the temporal variations (due to changes in density) of the FDT, along with a valid estimate of the background. We model the frames of a sequence as being produced by a continuous HMM, characterized by transition probabilities based on the Navier-Stokes equations for fluid dynamics, and by generation probabilities based on the convex matting of the FDT with the background. We learn the FDT appearance, the FDT temporal variations, and the background by maximizing their joint probability using interactive conditional modes (ICM). Since the learned model is generative, it can be used to synthesize new videos with different backgrounds and density variations. Experiments on videos that we compiled demonstrate the performance of our method.

Computational representation of perceived image quality is a fundamental problem in computer vision and image processing, which has assumed increased importance with the growing role of images and video in human-computer interaction. It is well-known that the commonly used Peak Signal-to-noise ratio (PSNR), although analysis-friendly, falls far short of this need. We propose a perceptual image quality measure (IQM) in terms of an image's region structure. Given a reference image and its "distorted" version, we propose a "full-reference" IQM, called segmentation-based perceptual image quality as sessment (SPIQA), which quantifies this quality reduction, while minimizing the disparity between human judgment and automated prediction of image quality. One novel feature of SPIQA is that it enables the use of inter- and intra- region attributes in a way that closely resembles how the human visual system (HVS) perceives distortion. Experimental results over a number of images and distortion types demonstrate SPIQA's performance benefits.

This paper presents a model of spatiotemporal variations in a dynamic texture (DT) sequence. Most recent work on DT modelling represents images in a DT sequence as the responses of a linear dynamical system (LDS) to noise. Despite its merits, this model has limitations because it attempts to model temporal variations in pixel intensities which do not take advantage of global motion coherence. We propose a model that relates texture dynamics to the variation of the Fourier phase, which captures the relationships among the motions of all pixels (i.e. global motion) within the texture, as well as the appearance of the texture. Unlike LDS, our model does not require segmentation or cropping during the training stage, which allows it to handle DT sequences containing a static background. We test the performance of this model on recognition and synthesis of DT's. Experiments with a dataset that we have compiled demonstrate that our phase based model outperforms LDS.

Temporal or dynamic textures (DT's) are video sequences that are spatially repetitive and temporally stationary. DT's are temporal analogs of the well known spatial still image texture. Examples of DT's include moving water, foliage, smoke, clouds, etc. We present a new DT model that can efficiently compress DT sequences. Our proposed method compactly represents the spatiotemporal properties of a DT by modelling its varying Fourier phase content, which can be shown to be the major determinant of both its dynamics and appearance. This is possible because this method combines both temporal and spatial properties in a compact spectral framework. Making use of the benefits inherent to working in the frequency domain, this model provides a significant improvement in DT compression, which can be used to improve the performance of MPEG-2 encoding. We will present experimental evidence that validates this method for a variety of complex sequences, while also comparing it to the most recent DT representational model that is based on modelling a DT as a linear dynamical system (LDS).

While market-based approaches, such as TraderBots, have shown much promise for efficient coordination of multirobot teams, the cost estimation mechanism and its impact on solution efficiency has not been investigated. This paper provides a first analysis of the cost estimation process in the TraderBots approach applied to a distributed sensing task. In the presented implementation, path costs are estimated using the D* path planning algorithm with optimistic costing of unknown map cells. The reported results show increased team efficiency when cost estimates reflect different environmental and mission characteristics. Thus, this paper demonstrates that market-based approaches can improve team efficiency if cost estimates take into account environmental and mission characteristics. These findings encourage future research on applying learning techniques for online modification of cost estimation and in market-based coordination.