Region representation learning plays a pivotal role in urban computing by extracting meaningful features from unlabeled urban data. Analogous to how perceived facial age reflects an individual's health, the visual appearance of a city serves as its "portrait", encapsulating latent socio-economic and environmental characteristics. Recent studies have explored leveraging Large Language Models (LLMs) to incorporate textual knowledge into imagery-based urban region representation learning. However, two major challenges remain: i) difficulty in aligning fine-grained visual features with long captions, and ii) suboptimal knowledge incorporation due to noise in LLM-generated captions. To address these issues, we propose a novel pre-training framework called UrbanLN that improves Urban region representation learning through Long-text awareness and Noise suppression. Specifically, we introduce an information-preserved stretching interpolation strategy that aligns long captions with fine-grained visual semantics in complex urban scenes. To effectively mine knowledge from LLM-generated captions and filter out noise, we propose a dual-level optimization strategy. At the data level, a multi-model collaboration pipeline automatically generates diverse and reliable captions without human intervention. At the model level, we employ a momentum-based self-distillation mechanism to generate stable pseudo-targets, facilitating robust cross-modal learning under noisy conditions. Extensive experiments across four real-world cities and various downstream tasks demonstrate the superior performance of our UrbanLN.

The rapid advancement of Internet of Vehicles technology has led to massive growth in vehicular data generation, imposing strict computational demands for latency-sensitive applications. By offloading computational tasks to Road-side Units (RSUs), Vehicular Edge Computing (VEC) offers an efficient solution for those latency-sensitive applications. However, current task offloading schemes generally ignore the time-varying topology caused by vehicle mobility, which poses a risk of task interruption. Moreover, existing task offloading models primarily focus on serial tasks processing and fail to adequately account for the relationships among parallel tasks, leading to inefficient resource utilization and potential latency accumulation in multi-task VEC scenarios. To this end, we propose a mobility-awareCollaborativeTaskOffloading scheme forParallel tasks (CoTOP) in VEC. Integrated with vehicle mobility detection, a collaborative task offloading model is designed based on deep reinforcement learning, achieving effective coordination among RSUs to reduce task processing latency. Additionally, a task prioritization algorithm is incorporated to optimize resource allocation. Experimental results show that CoTOP significantly outperforms existing schemes in terms of task processing latency, energy consumption and completion ratio.

Text-to-image diffusion models significantly enhance the efficiency of artistic creation with high-fidelity image generation. However, in typical application scenarios like comic book production, they can neither place each subject into its expected spot nor maintain the consistent appearance of each subject across images. For these issues, we pioneer a novel task, Layout-to-Consistent-Image (L2CI) generation, which produces consistent and compositional images in accordance with the given layout conditions and text prompts. To accomplish this challenging task, we present a new formalization of dual energy guidance with optimization in a dual semantic-latent space and thus propose a training-free pipeline, SpotActor, which features a layout-conditioned optimizing stage and a consistent sampling stage. In the optimizing stage, we innovate a nuanced layout energy function to mimic the attention activations with a sigmoid-like objective. While in the sampling stage, we design Regional Interconnection Self-Attention (RISA) and Semantic Fusion Cross-Attention (SFCA) mechanisms that allow mutual interactions across images. To evaluate the performance, we present ActorBench, a specified benchmark with hundreds of reasonable prompt-box pairs stemming from object detection datasets. Comprehensive experiments are conducted to demonstrate the effectiveness of our method. The results prove that SpotActor fulfills the expectations of this task and showcases the potential for practical applications with superior layout alignment, subject consistency, prompt conformity and background diversity.

Video action recognition involves interpreting both global context and specific details to accurately identify actions. While previous models are effective at capturing spatiotemporal features, they often lack a focused representation of key action details. To address this, we introduce FocusVideo, a framework designed for refining video action recognition through integrated global and local feature learning. Inspired by human visual cognition theory, our approach balances the focus on both broad contextual changes and action-specific details, minimizing the influence of irrelevant background noise. We first employ learnable action queries to selectively emphasize action-relevant regions without requiring region-specific labels. Next, these queries are learned by a local action streaming branch that enables progressive query propagation. Moreover, we introduce a parameter-free feature interaction mechanism for effective multi-scale interaction between global and local features with minimal additional overhead. Extensive experiments demonstrate that FocusVideo achieves state-of-the-art performance across multiple action recognition datasets, validating its effectiveness and robustness in handling action-relevant details.

Previous methods often utilize CLIP semantic classifiers with class names for zero-shot object counting. However, they ignore crucial density and depth knowledge for counting tasks. Thus, we propose a density-aware and depth-aware prompt counting model, which captures density information via learning density-aware prompts based on density-aware contrastive loss and incorporates depth guidance with predefined depth-aware prompts. To facilitate the training process, we design two strategies for standard counting loss and the contrastive loss, where the former prioritizes larger and sparser objects initially, gradually focusing on smaller and denser objects, and the latter adopts coarse-to-fine density learning. Besides, we construct a dataset named LVIS-372 with more real-world scenarios and balanced instance distribution compared to existing ones. Finally, the experimental results demonstrate the effectiveness of our proposed method.

Diffusion models have demonstrated significant potential for generating high-quality images, audio, and videos. However, their iterative inference process entails substantial computational costs, limiting practical applications. Recently, researchers have introduced accelerated sampling methods that enable diffusion models to generate samples with far fewer timesteps than those used during training. Nonetheless, as the number of sampling steps decreases, the prediction errors significantly degrade the quality of generated outputs. Additionally, the exposure bias in diffusion models further amplifies these errors. To address these challenges, we leverage a manifold hypothesis to explore the exposure bias problem in depth. Based on this geometric perspective, we propose a manifold constraint that effectively reduces exposure bias during accelerated sampling of diffusion models. Notably, our method involves no additional training and requires only minimal hyperparameter tuning. Extensive experiments demonstrate the effectiveness of our approach, achieving a FID score of 15.60 with 10-step SDXL on MS-COCO, surpassing the baseline by a reduction of 2.57 in FID.

Integrating natural language instructions and visual perception with decision-making is a critical challenge for embodied agents. Existing methods often struggle to balance the conciseness of language commands with the richness of video content. To bridge the gap between modalities, we propose extracting key spatiotemporal patterns from video that capture visual saliency and temporal evolution, referred to as dynamic representation. Building on this, we introduce DynaMind, a framework that enhances decision-making through dynamic reasoning. Specifically, we design an adaptive FrameScorer to evaluate video frames based on semantic consistency and visual saliency, assigning each frame an importance score. These scores are used to filter redundant video content and synthesize compact dynamic representations. Leveraging these representations, we predict critical future dynamics and apply a dynamic-guided policy to generate coherent and context-aware actions. Extensive results demonstrate that DynaMind significantly outperforms the baselines across several simulation benchmarks and real-world scenarios.

Text-to-image diffusion models have emerged as powerful tools for high-quality image generation and editing. Many existing approaches rely on text prompts as editing guidance. However, these methods are constrained by the need for manual prompt crafting, which can be time-consuming, introduce irrelevant details, and significantly limit editing performance. In this work, we propose optimizing semantic embeddings guided by attribute classifiers to steer text-to-image models toward desired edits, without relying on text prompts or requiring any training or fine-tuning of the diffusion model. We utilize classifiers to learn precise semantic embeddings at the dataset level. The learned embeddings are theoretically justified as the optimal representation of attribute semantics, enabling disentangled and accurate edits. Experiments further demonstrate that our method achieves high levels of disentanglement and strong generalization across different domains of data. Code is available at https://github.com/Chang-yuanyuan/CASO.

Text-guided video editing (TGVE) has become a recent hotspot due to its entertainment value and practical applications. To reduce overhead, existing methods primarily extend from text-to-image diffusion models and typically involve reconstruction and editing phases. However, challenges persist, particularly in enhancing temporal consistency of a video while adhering to textual alignment requirements. A crucial factor leading to the aforementioned issue is the inadequate and implicit tuning of the attention module within existing methods, which is specifically designed to capture temporal information. In light of this, we introduce VidEvo, a novel one-shot video editing method that leverages explicit cues derived from the original video to enhance temporal modeling. By integrating null-video embedding (NVE) and window-frame attention (WFA) components, VidEvo facilitates the smooth and coherent generation of videos from global and local perspectives simultaneously. To be specific, NVE learns a set of multi-scale temporal embeddings within the visual space during the reconstruction phase. These embeddings are subsequently directly injected into the attention module of the editing phase, explicitly augmenting the temporal consistency of the entire video. On the other hand, WFA enhances local temporal modeling by dynamically optimizing attention mechanisms between adjacent frames, which improves temporal coherence with reduced computational costs. Experimental evaluations show that VidEvo enhances frame-to-frame temporal consistency. Ablation studies confirm NVE and WFA’s effectiveness and their plug-and-play capability with other methods.

Large language models have demonstrated remarkable generalization capabilities and strong performance across various fields. Recent research has highlighted their significant potential in time series forecasting. However, time series data often exhibit complex periodic characteristics, posing a substantial challenge in enabling these models to effectively capture latent patterns. To address this challenge, we propose a novel framework, LLM-TPF, which leverages individuality and commonality fusion to enhance time series forecasting. In the frequency domain, periodic features are extracted to reveal the intrinsic periodicity of the data, while textual prototypes are used to indicate temporal trends. In the time domain, carefully designed prompts are employed to guide the models in comprehending global information. A commonality fusion mechanism further aggregates heterogeneous information across dimensions, and three distinct language models are utilized to independently process different types of information. Extensive real-world experiments demonstrate that LLM-TPF is a powerful tool for time series forecasting, achieving superior performance compared to state-of-the-art specialized models and exhibiting exceptional generalization ability in zero-shot scenarios. Code is available at https://github.com/switchsky/LLM-TPF.

With the development of wearable cardiac ultrasound devices, it is no longer sufficient to solely rely on doctors for diagnosing long-term echocardiogram videos. Automated diagnosis of echocardiogram videos has now become a research hotspot. Existing studies only analyze echocardiogram video through discriminative models, which have limited question-answering capabilities. Therefore, this study innovatively proposes a large language model with cardiac ultrasound diagnostic capabilities—EchoGPT. EchoGPT integrates the robust communication and comprehension capabilities of large language models (LLMs) with the diagnostic prowess of traditional medical models, empowering patients to obtain accurate medical indicator data and comprehend their health conditions through interactive questioning with the model. The model is capable of local deployment on personal computers, effectively safe guarding user privacy. EchoGPT operates through three main components: left ventricle segmentation, left ventricular ejection fraction LVEF prediction, and finetuning of video-text LLMs. Experimental results demonstrate EchoGPT’s superior accuracy in predicting LVEF compared to other models, and positive feedback from professional physicians through questionnaire surveys, validating its potential in practical applications. The demo is available at https://github.com/zhuqh19/EchoGPT.

3D visual grounding allows an embodied agent to understand visual information in real-world 3D environments based on human instructions, which is crucial for embodied intelligence. Existing 3D visual grounding methods typically rely on separate encoders for different modalities (e.g., RGB images, text, and 3D point clouds), resulting in large and complex models that are inefficient to train. While some approaches use pre-trained 2D multi-modal models like CLIP for 3D tasks, they still struggle with aligning point cloud data to 2D encoders. As a result, these methods continue to depend on 3D encoders for feature extraction, further increasing model complexity and training inefficiency. In this paper, we propose a unified 2D pre-trained multi-modal network to process all three modalities (RGB images, text, and point clouds), significantly simplifying the architecture. By leveraging a 2D CLIP bi-modal model with adapter-based fine-tuning, this framework effectively adapts to the tri-modal setting, improving both adaptability and performance across modalities. Our Geometric-Aware 2D-3D Feature Recovery and Fusion (GARF) module is designed to fuse geometric multi-scale features from point clouds and images. We then integrate textual features for final modality fusion and introduce a multi-modal decoder to facilitate deep cross-modal understanding. Together, our method achieves unified feature extraction and fusion across the three modalities, enabling an end-to-end 3D visual grounding model. Compared to the baseline, our method reduces the number of trainable parameters by approximately 58%, while achieving a 6.52% improvement in the 3D detection task and a 6.25% improvement in the 3D visual grounding task.

Although learning 3D manipulation policies from monocular RGB images is lightweight and deployment-friendly, the lack of structural information often leads to inaccurate action estimation. While explicit 3D inputs can mitigate this issue, they typically require additional sensors and introduce data acquisition overhead. An intuitive alternative is to incorporate a pre-trained depth estimator; however, this often incurs substantial inference-time cost. To address this, we propose MonoLift, a tri-level knowledge distillation framework that transfers spatial, temporal, and action-level knowledge from a depth-guided teacher to a monocular RGB student. By jointly distilling geometry-aware features, temporal dynamics, and policy behaviors during training, MonoLift enables the student model to perform 3D-aware reasoning and precise control at deployment using only monocular RGB input. Extensive experiments on both simulated and real-world manipulation tasks show that MonoLift not only outperforms existing monocular approaches but even surpasses several methods that rely on explicit 3D input, offering a resource-efficient and effective solution for vision-based robotic control. The video demonstration is available on our project page: https://robotasy.github.io/MonoLift/.

Filter pruning is a significant feature selection technique to shrink the existing feature fusion schemes (especially on convolution calculation and model size), which helps to develop more efficient feature fusion models while maintaining state-of-the-art performance. In addition, it reduces the storage and computation requirements of deep neural networks (DNNs) and accelerates the inference process dramatically. Existing methods mainly rely on manual constraints such as normalization to select the filters. A typical pipeline comprises two stages: first pruning the original neural network and then fine-tuning the pruned model. However, choosing a manual criterion can be somehow tricky and stochastic. Moreover, directly regularizing and modifying filters in the pipeline suffer from being sensitive to the choice of hyperparameters, thus making the pruning procedure less robust. To address these challenges, we propose to handle the filter pruning issue through one stage: using an attention-based architecture that adaptively fuses the filter selection with filter learning in a unified network. Specifically, we present a pruning method named adding before pruning (ABP) to make the model focus on the filters of higher significance by training instead of man-made criteria such as norm, rank, etc. First, we add an auxiliary attention layer into the original model and set the significance scores in this layer to be binary. Furthermore, to propagate the gradients in the auxiliary attention layer, we design a specific gradient estimator and prove its effectiveness for convergence in the graph flow through mathematical derivation. In the end, to relieve the dependence on the complicated prior knowledge for designing the thresholding criterion, we simultaneously prune and train the filters to automatically eliminate network redundancy with recoverability. Extensive experimental results on the two typical image classification benchmarks, CIFAR-10 and ILSVRC-2012, illustrate that the proposed approach performs favorably against previous state-of-the-art filter pruning algorithms.

The canonical approach to video action recognition dictates a neural network model to do a classic and standard 1-of-N majority vote task. They are trained to predict a fixed set of predefined categories, limiting their transferability on new datasets with unseen concepts. In this article, we provide a new perspective on action recognition by attaching importance to the semantic information of label texts rather than simply mapping them into numbers. Specifically, we model this task as a video-text matching problem within a multimodal learning framework, which strengthens the video representation with more semantic language supervision and enables our model to do zero-shot action recognition without any further labeled data or parameters' requirements. Moreover, to handle the deficiency of label texts and make use of tremendous web data, we propose a new paradigm based on this multimodal learning framework for action recognition, which we dub "pre-train, adapt and fine-tune." This paradigm first learns powerful representations from pre-training on a large amount of web image-text or video-text data. Then, it makes the action recognition task to act more like pre-training problems via adaptation engineering. Finally, it is fine-tuned end-to-end on target datasets to obtain strong performance. We give an instantiation of the new paradigm, ActionCLIP, which not only has superior and flexible zero-shot/few-shot transfer ability but also reaches a top performance on general action recognition task, achieving 83.8% top-1 accuracy on Kinetics-400 with a ViT-B/16 as the backbone. Code is available at https://github.com/sallymmx/ActionCLIP.git.

The point cloud-based 3D single object tracking plays an indispensable role in autonomous driving. However, the application of 3D object tracking in the real world is still challenging due to the inherent sparsity and self-occlusion of point cloud data. Therefore, it is necessary to exploit as much useful information from limited data as we can. Since 3D object tracking is a video-level task, the appearance of objects changes gradually over time, and there is rich spatiotemporal contextual information among historical frames. However, existing methods do not fully utilize this information. To address this, we propose a new method called SCTrack, which utilizes a memory-based paradigm to exploit spatiotemporal contextual information. SCTrack incorporates both long-term and short-term memory banks to store the spatiotemporal features of targets from historical frames. By doing so, the tracker can benefit from the entire video sequence and make more informed predictions. Additionally, SCTrack extracts the mask prior to augmenting the target representation, improving the target-background discriminability. Extensive experiments on KITTI, nuScenes, and Waymo Open datasets verify the effectiveness of our proposed method.

Despite significant progress in Text-to-Image (T2I) generative models, even lengthy and complex text descriptions still struggle to convey detailed controls. In contrast, Layout-to-Image (L2I) generation, aiming to generate realistic and complex scene images from user-specified layouts, has risen to prominence. However, existing methods transform layout information into tokens or RGB images for conditional control in the generative process, leading to insufficient spatial and semantic controllability of individual instances. To address these limitations, we propose a novel Spatial-Semantic Map Guided (SSMG) diffusion model that adopts the feature map, derived from the layout, as guidance. Owing to rich spatial and semantic information encapsulated in well-designed feature maps, SSMG achieves superior generation quality with sufficient spatial and semantic controllability compared to previous works. Additionally, we propose the Relation-Sensitive Attention (RSA) and Location-Sensitive Attention (LSA) mechanisms. The former aims to model the relationships among multiple objects within scenes while the latter is designed to heighten the model's sensitivity to the spatial information embedded in the guidance. Extensive experiments demonstrate that SSMG achieves highly promising results, setting a new state-of-the-art across a range of metrics encompassing fidelity, diversity, and controllability.

Recently, the rise of large-scale vision-language pretrained models like CLIP, coupled with the technology of Parameter-Efficient FineTuning (PEFT), has captured substantial attraction in video action recognition. Nevertheless, prevailing approaches tend to prioritize strong supervised performance at the expense of compromising the models' generalization capabilities during transfer. In this paper, we introduce a novel Multimodal, Multi-task CLIP adapting framework named M2-CLIP to address these challenges, preserving both high supervised performance and robust transferability. Firstly, to enhance the individual modality architectures, we introduce multimodal adapters to both the visual and text branches. Specifically, we design a novel visual TED-Adapter, that performs global Temporal Enhancement and local temporal Difference modeling to improve the temporal representation capabilities of the visual encoder. Moreover, we adopt text encoder adapters to strengthen the learning of semantic label information. Secondly, we design a multi-task decoder with a rich set of supervisory signals, including the original contrastive learning head, a cross-modal classification head, a cross-modal masked language modeling head, and a visual classification head. This multi-task decoder adeptly satisfies the need for strong supervised performance within a multimodal framework. Experimental results validate the efficacy of our approach, demonstrating exceptional performance in supervised learning while maintaining strong generalization in zero-shot scenarios.

While large models have achieved significant progress in computer vision, challenges such as optimization complexity, the intricacy of transformer architectures, computational constraints, and practical application demands highlight the importance of simpler model designs in medical image segmentation. This need is particularly pronounced in mobile medical devices, which require lightweight, deployable models with real-time performance. However, existing lightweight models often suffer from poor robustness across datasets, limiting their widespread adoption. To address these challenges, this paper introduces LV-UNet, a lightweight and vanilla model that leverages pre-trained MobileNetv3-Large backbones and incorporates fusible modules. LV-UNet employs an enhanced deep training strategy and switches to a deployment mode during inference by re-parametrization, significantly reducing parameter count and computational overhead. Experimental results on ISIC 2016, BUSI, CVC-ClinicDB, CVC-ColonDB, and Kvair-SEG datasets demonstrate a better trade-off between performance and the computational load. The code will be released at https://github.com/juntaoJianggavin/LV-UNet.

Multimodal Visual Object Tracking (VOT) has recently gained significant attention due to its robustness. Early research focused on fully fine-tuning RGB-based trackers, which was inefficient and lacked generalized representation due to the scarcity of multimodal data. Therefore, recent studies have utilized prompt tuning to transfer pre-trained RGB-based trackers to multimodal data. However, the modality gap limits pre-trained knowledge recall, and the dominance of the RGB modality persists, preventing the full utilization of information from other modalities. To address these issues, we propose a novel symmetric multimodal tracking framework called SDSTrack. We introduce lightweight adaptation for efficient fine-tuning, which directly transfers the feature extraction ability from RGB to other domains with a small number of trainable parameters and integrates multimodal features in a balanced, symmetric manner. Furthermore, we design a complementary masked patch distillation strategy to enhance the robustness of trackers in complex environments, such as extreme weather, poor imaging, and sensor failure. Extensive experiments demonstrate that SDSTrack outperforms state-of-the-art methods in various multimodal tracking scenarios, including RGB+Depth, RGB+Thermal, and RGB+Event tracking, and exhibits impressive results in extreme conditions. Our source code is available at: https://github.com/hoqolo/SDSTrack.

The conjugate gradient method is a crucial first-order optimization method that generally converges faster than the steepest descent method, and its computational cost is much lower than that of second-order methods. However, while various types of conjugate gradient methods have been studied in Euclidean spaces and on Riemannian manifolds, there is little study for those in distributed scenarios. This paper proposes a decentralized Riemannian conjugate gradient descent (DRCGD) method that aims at minimizing a global function over the Stiefel manifold. The optimization problem is distributed among a network of agents, where each agent is associated with a local function, and the communication between agents occurs over an undirected connected graph. Since the Stiefel manifold is a non-convex set, a global function is represented as a finite sum of possibly non-convex (but smooth) local functions. The proposed method is free from expensive Riemannian geometric operations such as retractions, exponential maps, and vector transports, thereby reducing the computational complexity required by each agent. To the best of our knowledge, DRCGD is the first decentralized Riemannian conjugate gradient algorithm to achieve global convergence over the Stiefel manifold.

Tracking a specific person in 3D scene is gaining momentum due to its numerous applications in robotics. Currently, most 3D trackers focus on driving scenarios with neglected jitter and uncomplicated surroundings, which results in their severe degeneration in complex environments, especially on jolting robot platforms (only 20-60% success rate). To improve the accuracy, a Point-Video-based Transformer Tracking model (PVTrack) is presented for robots. It is the first multi-modal 3D human tracking work that incorporates point clouds together with RGB videos to achieve information complementarity. Moreover, PVTrack proposes the Siamese Point-Video Transformer for feature aggregation to overcome dynamic environments, which captures more target-aware information through the hierarchical attention mechanism adaptively. Considering the violent shaking on robots and rugged terrains, a lateral Human-ware Proposal Network is designed together with an Anti-shake Proposal Compensation module. It alleviates the disturbance caused by complex scenes as well as the particularity of the robot platform. Experiments show that our method achieves state-of-the-art performance on both KITTI/Waymo datasets and a quadruped robot for various indoor and outdoor scenes.

The goal of this paper is to strike a feasible tracking paradigm that can make 3D human trackers applicable on robot platforms and enable more high-level tasks. Till now, two fundamental problems haven’t been adequately addressed. One is the computational cost lightweight enough for robotic deployment, and the other is the easily-influenced accuracy varied greatly in complex real environments. In this paper, a robotic-centric tracking paradigm called MATNet is proposed that directly matches the LiDAR point clouds and RGB videos through end-to-end learning. To improve the low accuracy of human tracking against disturbance, a coarse-to-fine Transformer along with target-ware augmentation is proposed by fusing RGB videos and point clouds through a pyramid encoding and decoding strategy. To better meet the real-time requirement of actual robot deployment, we introduce the parameter-efficient adaptation tuning that greatly shortens the model’s training time. Furthermore, we also propose a five-step Anti-shake Refinement strategy and have added human prior values to overcome the strong shaking on the robot plat-form. Extensive experiments confirm that MATNet significantly outperforms the previous state-of-the-art on both open-source datasets and large-scale robotic datasets.

Text-guided diffusion models have significantly advanced image editing, enabling high-quality and diverse modifications driven by text prompts. However, effective editing requires inverting the source image into a latent space, a process often hindered by prediction errors inherent in DDIM inversion. These errors accumulate during the diffusion process, resulting in inferior content preservation and edit fidelity, especially with conditional inputs. We address these challenges by investigating the primary contributors to error accumulation in DDIM inversion and identify the singularity problem in traditional noise schedules as a key issue. To resolve this, we introduce the *Logistic Schedule*, a novel noise schedule designed to eliminate singularities, improve inversion stability, and provide a better noise space for image editing. This schedule reduces noise prediction errors, enabling more faithful editing that preserves the original content of the source image. Our approach requires no additional retraining and is compatible with various existing editing methods. Experiments across eight editing tasks demonstrate the Logistic Schedule's superior performance in content preservation and edit fidelity compared to traditional noise schedules, highlighting its adaptability and effectiveness. The project page is available at https://lonelvino.github.io/SYE/.

Recent advancements have shown promise in applying traditional Semi-Supervised Learning strategies to the task of Generalized Category Discovery (GCD). Typically, this involves a teacher-student framework in which the teacher imparts knowledge to the student to classify categories, even in the absence of explicit labels. Nevertheless, GCD presents unique challenges, particularly the absence of priors for new classes, which can lead to the teacher's misguidance and unsynchronized learning with the student, culminating in suboptimal outcomes. In our work, we delve into why traditional teacher-student designs falter in generalized category discovery as compared to their success in closed-world semi-supervised learning. We identify inconsistent pattern learning as the crux of this issue and introduce FlipClass—a method that dynamically updates the teacher to align with the student's attention, instead of maintaining a static teacher reference. Our teacher-attention-update strategy refines the teacher's focus based on student feedback, promoting consistent pattern recognition and synchronized learning across old and new classes. Extensive experiments on a spectrum of benchmarks affirm that FlipClass significantly surpasses contemporary GCD methods, establishing new standards for the field.

Text-to-image diffusion models benefit artists with high-quality image generation. Yet their stochastic nature hinders artists from creating consistent images of the same subject. Existing methods try to tackle this challenge and generate consistent content in various ways. However, they either depend on external restricted data or require expensive tuning of the diffusion model. For this issue, we propose a novel one-shot tuning paradigm, termed OneActor. It efficiently performs consistent subject generation solely driven by prompts via a learned semantic guidance to bypass the laborious backbone tuning. We lead the way to formalize the objective of consistent subject generation from a clustering perspective, and thus design a cluster-conditioned model. To mitigate the overfitting challenge shared by one-shot tuning pipelines, we augment the tuning with auxiliary samples and devise two inference strategies: semantic interpolation and cluster guidance. These techniques are later verified to significantly improve the generation quality. Comprehensive experiments show that our method outperforms a variety of baselines with satisfactory subject consistency, superior prompt conformity as well as high image quality. Our method is capable of multi-subject generation and compatible with popular diffusion extensions. Besides, we achieve a $4\times$ faster tuning speed than tuning-based baselines and, if desired, avoid increasing the inference time. Furthermore, our method can be naturally utilized to pre-train a consistent subject generation network from scratch, which will implement this research task into more practical applications. (Project page: https://johnneywang.github.io/OneActor-webpage/)

In this paper, we study Discretized Neural Networks (DNNs) composed of low-precision weights and activations, which suffer from either infinite or zero gradients due to the non-differentiable discrete function during training. Most training-based DNNs in such scenarios employ the standard Straight-Through Estimator (STE) to approximate the gradient w.r.t. discrete values. However, the use of STE introduces the problem of gradient mismatch, arising from perturbations in the approximated gradient. To address this problem, this paper reveals that this mismatch can be interpreted as a metric perturbation in a Riemannian manifold, viewed through the lens of duality theory. Building on information geometry, we construct the Linearly Nearly Euclidean (LNE) manifold for DNNs, providing a background for addressing perturbations. By introducing a partial differential equation on metrics, i.e., the Ricci flow, we establish the dynamical stability and convergence of the LNE metric with the $L^2$-norm perturbation. In contrast to previous perturbation theories with convergence rates in fractional powers, the metric perturbation under the Ricci flow exhibits exponential decay in the LNE manifold. Experimental results across various datasets demonstrate that our method achieves superior and more stable performance for DNNs compared to other representative training-based methods.

Cross-domain generalizable depth estimation aims to estimate the depth of target domains (i.e., real-world) using models trained on the source domains (i.e., synthetic). Previous methods mainly use additional real-world domain datasets to extract depth specific information for cross-domain generalizable depth estimation. Unfortunately, due to the large domain gap, adequate depth specific information is hard to obtain and interference is difficult to remove, which limits the performance. To relieve these problems, we propose a domain generalizable feature extraction network with adaptive guidance fusion (AGDF-Net) to fully acquire essential features for depth estimation at multi-scale feature levels. Specifically, our AGDF-Net first separates the image into initial depth and weak-related depth components with reconstruction and contrary losses. Subsequently, an adaptive guidance fusion module is designed to sufficiently intensify the initial depth features for domain generalizable intensified depth features acquisition. Finally, taking intensified depth features as input, an arbitrary depth estimation network can be used for real-world depth estimation. Using only synthetic datasets, our AGDF-Net can be applied to various real-world datasets (i.e., KITTI, NYUDv2, NuScenes, DrivingStereo and CityScapes) with state-of-the-art performances. Furthermore, experiments with a small amount of real-world data in a semi-supervised setting also demonstrate the superiority of AGDF-Net over state-of-the-art approaches.

To build commercial robots, skid-steering mechanical design is of increased popularity due to its manufacturing simplicity and unique mechanism. However, these also cause significant challenges on software and algorithm design, especially for the pose estimation (i.e., determining the robot’s rotation and position) of skid-steering robots, since they change their orientation with an inevitable skid. To tackle this problem, we propose a probabilistic sliding-window estimator dedicated to skid-steering robots, using measurements from a monocular camera, the wheel encoders, and optionally an inertial measurement unit (IMU). Specifically, we explicitly model the kinematics of skid-steering robots by both track instantaneous centers of rotation (ICRs) and correction factors, which are capable of compensating for the complexity of track-to-terrain interaction, the imperfectness of mechanical design, terrain conditions and smoothness, etc. To prevent performance reduction in robots’ long-term missions, the time- and location- varying kinematic parameters are estimated online along with pose estimation states in a tightly-coupled manner. More importantly, we conduct in-depth observability analysis for different sensors and design configurations in this paper, which provides us with theoretical tools in making the correct choice when building real commercial robots. In our experiments, we validate the proposed method by both simulation tests and real-world experiments, which demonstrate that our method outperforms competing methods by wide margins. Note to Practitioners—This paper was motivated by the problem of long-term pose estimation of the commonly commercial-used skid-steering robots with only low-cost sensors. Skid-steering robots change their orientation with a skid, which poses a significant challenge for pose estimation when using the wheel encoders. We propose to online estimate the robot’s kinematics, which succeeds in compensating for the complexity of track-to-terrain interaction, due to the slippage, the imperfectness of mechanical design, terrain conditions and smoothness. It is critical to estimate the kinematics and poses jointly to prevent performance reduction in robots’ long-term missions. We further theoretically analyze whether the kinematics parameters can be estimated under different sensor configurations, and find out the special degrade motions that make the parameters unobservable.

Semantic scene completion (SSC) aims to predict the semantic occupancy of each voxel in the entire 3D scene from limited observations, which is an emerging and critical task for autonomous driving. Recently, many studies have turned to camera-based SSC solutions due to the richer visual cues and cost-effectiveness of cameras. However, existing methods usually rely on sophisticated and heavy 3D models to process the lifted 3D features directly, which are not discriminative enough for clear segmentation boundaries. In this paper, we adopt the dense-sparse-dense design and propose a one-stage camera-based SSC framework, termed SGN, to propagate semantics from the semantic-aware seed voxels to the whole scene based on spatial geometry cues. Firstly, to exploit depth-aware context and dynamically select sparse seed voxels, we redesign the sparse voxel proposal network to process points generated by depth prediction directly with the coarse-to-fine paradigm. Furthermore, by designing hybrid guidance (sparse semantic and geometry guidance) and effective voxel aggregation for spatial geometry cues, we enhance the feature separation between different categories and expedite the convergence of semantic propagation. Finally, we devise the multi-scale semantic propagation module for flexible receptive fields while reducing the computation resources. Extensive experimental results on the SemanticKITTI and SSCBench-KITTI-360 datasets demonstrate the superiority of our SGN over existing state-of-the-art methods. And even our lightweight version SGN-L achieves notable scores of 14.80% mIoU and 45.45% IoU on SeamnticKITTI validation with only 12.5 M parameters and 7.16 G training memory. Code is available at https://github.com/Jieqianyu/SGN.

Spatial and temporal modeling is one of the most core aspects of few-shot action recognition. Most previous works mainly focus on long-term temporal relation modeling based on high-level spatial representations, without considering the crucial low-level spatial features and short-term temporal relations. Actually, the former feature could bring rich local semantic information, and the latter feature could represent motion characteristics of adjacent frames, respectively. In this paper, we propose SloshNet, a new framework that revisits the spatial and temporal modeling for few-shot action recognition in a finer manner. First, to exploit the low-level spatial features, we design a feature fusion architecture search module to automatically search for the best combination of the low-level and high-level spatial features. Next, inspired by the recent transformer, we introduce a long-term temporal modeling module to model the global temporal relations based on the extracted spatial appearance features. Meanwhile, we design another short-term temporal modeling module to encode the motion characteristics between adjacent frame representations. After that, the final predictions can be obtained by feeding the embedded rich spatial-temporal features to a common frame-level class prototype matcher. We extensively validate the proposed SloshNet on four few-shot action recognition datasets, including Something-Something V2, Kinetics, UCF101, and HMDB51. It achieves favorable results against state-of-the-art methods in all datasets.

Recently, neural implicit surfaces have become popular for multi-view reconstruction. To facilitate practical applications like scene editing and manipulation, some works extend the framework with semantic masks input for the object-compositional reconstruction rather than the holistic perspective. Though achieving plausible disentanglement, the performance drops significantly when processing the indoor scenes where objects are usually partially observed. We propose RICO to address this by regularizing the unobservable regions for indoor compositional reconstruction. Our key idea is to first regularize the smoothness of the occluded background, which then in turn guides the foreground object reconstruction in unobservable regions based on the object-background relationship. Particularly, we regularize the geometry smoothness of occluded background patches. With the improved background surface, the signed distance function and the reversedly rendered depth of objects can be optimized to bound them within the background range. Extensive experiments show our method outperforms other methods on synthetic and real-world indoor scenes and prove the effectiveness of proposed regularizations. The code is available at https://github.com/kyleleey/RICO

Siamese network has been a de facto benchmark framework for 3D LiDAR object tracking with a shared-parametric encoder extracting features from template and search region, respectively. This paradigm relies heavily on an additional matching network to model the cross-correlation/similarity of the template and search region. In this paper, we forsake the conventional Siamese paradigm and propose a novel single-branch framework, SyncTrack, synchronizing the feature extracting and matching to avoid forwarding encoder twice for template and search region as well as introducing extra parameters of matching network. The synchronization mechanism is based on the dynamic affinity of the Transformer, and an in-depth analysis of the relevance is provided theoretically. Moreover, based on the synchronization, we introduce a novel Attentive PointsSampling strategy into the Transformer layers (APST), replacing the random/Farthest Points Sampling (FPS) method with sampling under the supervision of attentive relations between the template and search region. It implies connecting point-wise sampling with the feature learning, beneficial to aggregating more distinctive and geometric features for tracking with sparse points. Extensive experiments on two benchmark datasets (KITTI and NuScenes) show that SyncTrack achieves state-of-the-art performance in realtime tracking.

Class prototype construction and matching are core aspects of few-shot action recognition. Previous methods mainly focus on designing spatiotemporal relation modeling modules or complex temporal alignment algorithms. Despite the promising results, they ignored the value of class prototype construction and matching, leading to unsatisfactory performance in recognizing similar categories in every task. In this paper, we propose GgHM, a new framework with Graph-guided Hybrid Matching. Concretely, we learn task-oriented features by the guidance of a graph neural network during class prototype construction, optimizing the intra- and inter-class feature correlation explicitly. Next, we design a hybrid matching strategy, combining frame-level and tuple-level matching to classify videos with multivariate styles. We additionally propose a learnable dense temporal modeling module to enhance the video feature temporal representation to build a more solid foundation for the matching process. GgHM shows consistent improvements over other challenging baselines on several few-shot datasets, demonstrating the effectiveness of our method. The code will be publicly available at https://github.com/jiazheng-xing/GgHM.

Reliable LiDAR panoptic segmentation (LPS), including both semantic and instance segmentation, is vital for many robotic applications, such as autonomous driving. This work proposes a new LPS framework named PANet to eliminate the dependency on the offset branch and improve the performance on large objects, which are always over-segmented by clustering algorithms. Firstly, we propose a non-learning Sparse Instance Proposal (SIP) module with the “sampling-shifting-grouping” scheme to directly group thing points into instances from the raw point cloud efficiently. More specifically, balanced point sampling is introduced to generate sparse seed points with more uniform point distribution over the distance range. And a shift module, termed bubble shifting, is proposed to shrink the seed points to the clustered centers. Then we utilize the connected component label algorithm to generate instance proposals. Furthermore, an instance aggregation module is devised to integrate potentially fragmented instances, improving the performance of the SIP module on large objects. Extensive experiments show that PANet achieves state-of-the-art performance among published works on the SemanticKITII validation and nuScenes validation for the panoptic segmentation task. Code is available at https://github.com/Jieqianyu/PANet.git.

Semantic scene completion (SSC) jointly predicts the semantics and geometry of the entire 3D scene, which plays an essential role in 3D scene understanding for autonomous driving systems. SSC has achieved rapid progress with the help of semantic context in segmentation. However, how to effectively exploit the relationships between the semantic context in semantic segmentation and geometric structure in scene completion remains under exploration. In this paper, we propose to solve outdoor SSC from the perspective of representation separation and BEV fusion. Specifically, we present the network, named SSC-RS, which uses separate branches with deep supervision to explicitly disentangle the learning procedure of the semantic and geometric representations. And a BEV fusion network equipped with the proposed Adaptive Representation Fusion (ARF) module is presented to aggregate the multi-scale features effectively and efficiently. Due to the low computational burden and powerful representation ability, our model has good generality while running in real-time. Extensive experiments on SemanticKITTI demonstrate our SSC-RS achieves state-of-the-art performance. Code is available at https://github.com/Jieqianyu/SSC-RS.git.

This paper focuses on LiDAR Panoptic Segmentation (LPS), which has attracted more attention recently due to its broad application prospect for autonomous driving and robotics. The mainstream LPS approaches either adopt a top-down strategy relying on 3D object detectors to discover instances or utilize time-consuming heuristic clustering algorithms to group instances in a bottom-up manner. Inspired by the center representation and kernel-based segmentation, we propose a new detection-free and clustering-free framework called CenterLPS, with the center-based instance encoding and decoding paradigm. Specifically, we propose a sparse center proposal network to generate the sparse 3D instance centers, as well as center feature embedding, which can well encode characteristics of instances. Then a center-aware transformer is applied to collect the context between different center feature embedding and around centers. Moreover, we generate the kernel weights based on the enhanced center feature embedding and initialize dynamic convolutions to decode the final instance masks. Finally, a mask fusion module is devised to unify the semantic and instance predictions and improve the panoptic quality. Extensive experiments on SemanticKITTI and nuScenes demonstrate the effectiveness of our proposed center-based framework CenterLPS.

Recent methods for action recognition always apply 3D Convolutional Neural Networks (CNNs) to extract spatiotemporal features and introduce optical flows to present motion features. Although achieving state-of-the-art performance, they are expensive in both time and space. In this paper, we propose to represent both two kinds of features in a unified 2D CNN without any 3D convolution or optical flows calculation. In particular, we first design a channel-wise spatiotemporal module to present the spatiotemporal features and a channel-wise motion module to encode feature-level motion features efficiently. Besides, we provide a distinctive illustration of the two modules from the frequency domain by interpreting them as advanced and learnable versions of frequency components. Second, we combine these two modules and an identity mapping path into one united block that can easily replace the original residual block in the ResNet architecture, forming a simple yet effective network dubbed STM network by introducing very limited extra computation cost and parameters. Third, we propose a novel Twins Training framework for action recognition by incorporating a correlation loss to optimize the inter-class and intra-class correlation and a siamese structure to fully stretch the training data. We extensively validate the proposed STM on both temporal-related datasets (i.e., Something-Something v1 & v2) and scene-related datasets (i.e., Kinetics-400, UCF-101, and HMDB-51). It achieves favorable results against state-of-the-art methods in all the datasets.

In this article, we present a fast real-time tangled memory network that segments the objects effectively and efficiently for semi-supervised video object segmentation (VOS). We propose a tangled reference encoder and a memory bank organization mechanism based on a state estimator to fully utilize the mask features and alleviate memory overhead and computational burden brought by the unlimited memory bank used in many memory-based methods. First, the tangled memory network exploits the mask features that uncover abundant object information like edges and contours but are not fully explored in existing methods. Specifically, a tangled two-stream reference encoder is designed to extract and fuse the features from both RGB frames and the predicted masks. Second, to indicate the quality of the predicted mask and feedback the online prediction state for organizing the memory bank, we devise a target state estimator to learn the IoU score between the predicted mask and ground truth. Moreover, to accelerate the forward process and avoid memory overflow, we use a memory bank of fixed size to store historical features by designing a new efficient memory bank organization mechanism based on the mask state score provided by the state estimator. We conduct comprehensive experiments on the public benchmarks DAVIS and YouTube-VOS, demonstrating that our method obtains competitive results while running at high speed (66 FPS on the DAVIS16-val set).

Video object detection (VOD) focuses on detecting objects for each frame in a video, which is a challenging task due to appearance deterioration in certain video frames. Recent works usually distill crucial information from multiple support frames to improve the reference features, but they only perform at frame level or proposal level that cannot integrate spatial-temporal features sufficiently. To deal with this challenge, we treat VOD as a spatial-temporal hierarchical features interacting process and introduce a Multi-level Spatial-Temporal (MST) feature aggregation framework to fully exploit frame-level, proposal-level, and instance-level information in a unified framework. Specifically, MST first measures context similarity in pixel space to enhance all frame-level features rather than only update reference features. The proposal-level feature aggregation then models object relation to augment reference object proposals. Furthermore, to filter out irrelevant information from other classes and backgrounds, we introduce an instance ID constraint to boost instance-level features by leveraging support object proposal features that belong to the same object. Besides, we propose a Deformable Feature Alignment (DAlign) module before MST to achieve a more accurate pixel-level spatial alignment for better feature aggregation. Extensive experiments are conducted on ImageNet VID and UAVDT datasets that demonstrate the superiority of our method over state-of-the-art (SOTA) methods. Our method achieves 83.3% and 62.1% with ResNet-101 on two datasets, outperforming SOTA MEGA by 0.4% and 2.7%.

This paper focuses on the mask utilization of video object segmentation (VOS). The mask here mains the reference masks in the memory bank, i.e., several chosen high-quality predicted masks, which are usually used with the reference frames together. The reference masks depict the edge and contour features of the target object and indicate the boundary of the target against the background, while the reference frames contain the raw RGB information of the whole image. It is obvious that the reference masks could play a significant role in the VOS, but this is not well explored yet. To tackle this, we propose to investigate the mask advantages of both the encoder and the matcher. For the encoder, we provide a unified codebase to integrate and compare eight different mask-fused encoders. Half of them are inherited or summarized from existing methods, and the other half are devised by ourselves. We find the best configuration from our design and give valuable observations from the comparison. Then, we propose a new mask-enhanced matcher to reduce the background distraction and enhance the locality of the matching process. Combining the mask-fused encoder, mask-enhanced matcher and a standard decoder, we formulate a new architecture named MaskVOS, which sufficiently exploits the mask benefits for VOS. Qualitative and quantitative results demonstrate the effectiveness of our method. We hope our exploration could raise the attention of mask utilization in VOS.

Small object detection remains an unsolved challenge because it is hard to extract the information of small objects with only a few pixels. While scale-level corresponding detection in feature pyramid network alleviates this problem, we find feature coupling of various scales still impairs the performance of small objects. In this paper, we propose an extended feature pyramid network (EFPN) with an extra high-resolution pyramid level specialized for small object detection. Specifically, we design a novel module, named feature texture transfer (FTT), which is used to super-resolve features and extract credible regional details simultaneously. Moreover, we introduce a cross resolution distillation mechanism to transfer the ability of perceiving details across the scales of the network, where a foreground-background-balanced loss function is designed to alleviate area imbalance of foreground and background. In our experiments, the proposed EFPN is efficient on both computation and memory, and yields state-of-the-art results on small traffic-sign dataset Tsinghua-Tencent 100 K and small category of general object detection dataset MS COCO.

Depth completion aims to recover a dense depth map from a sparse depth map with the corresponding color image as input. Recent approaches mainly formulate the depth completion as a one-stage end-to-end learning task, which outputs dense depth maps directly. However, the feature extraction and supervision in one-stage frameworks are insufficient, limiting the performance of these approaches. To address this problem, we propose a novel end-to-end residual learning framework, which formulates the depth completion as a two-stage learning task, i.e., a sparse-to-coarse stage and a coarse-to-fine stage. First, a coarse dense depth map is obtained by a simple CNN framework. Then, a refined depth map is further obtained using a residual learning strategy in the coarse-to-fine stage with coarse depth map and color image as input. Specially, in the coarse-to-fine stage, a channel shuffle extraction operation is utilized to extract more representative features from color image and coarse depth map, and an energy based fusion operation is exploited to effectively fuse these features obtained by channel shuffle operation, thus leading to more accurate and refined depth maps. We achieve SoTA performance in RMSE on KITTI benchmark. Extensive experiments on other datasets future demonstrate the superiority of our approach over current state-of-the-art depth completion approaches.

Self-supervised learning shows great potential in monocular depth estimation, using image sequences as the only source of supervision. Although people try to use the high-resolution image for depth estimation, the accuracy of prediction has not been significantly improved. In this work, we find the core reason comes from the inaccurate depth estimation in large gradient regions, making the bilinear interpolation error gradually disappear as the resolution increases. To obtain more accurate depth estimation in large gradient regions, it is necessary to obtain high-resolution features with spatial and semantic information. Therefore, we present an improved DepthNet, HR-Depth, with two effective strategies: (1) re-design the skip-connection in DepthNet to get better high-resolution features and (2) propose feature fusion Squeeze-and-Excitation(fSE) module to fuse feature more efficiently. Using Resnet-18 as the encoder, HR-Depth surpasses all previous state-of-the-art(SoTA) methods with the least parameters at both high and low resolution. Moreover, previous SoTA methods are based on fairly complex and deep networks with a mass of parameters which limits their real applications. Thus we also construct a lightweight network which uses MobileNetV3 as encoder. Experiments show that the lightweight network can perform on par with many large models like Monodepth2 at high-resolution with only20%parameters. All codes and models will be available at https://github.com/shawLyu/HR-Depth.

In this paper we tackle the problem of pose guided person image generation, which aims to transfer a person image from the source pose to a novel target pose while maintaining the source appearance. Given the inefficiency of standard CNNs in handling large spatial transformation, we propose a structure-aware flow based method for high-quality person image generation. Specifically, instead of learning the complex overall pose changes of human body, we decompose the human body into different semantic parts (e.g., head, torso, and legs) and apply different networks to predict the flow fields for these parts separately. Moreover, we carefully design the network modules to effectively capture the local and global semantic correlations of features within and among the human parts respectively. Extensive experimental results show that our method can generate high-quality results under large pose discrepancy and outperforms state-of-the-art methods in both qualitative and quantitative comparisons.

The paper proposes a novel generative adversarial network for one-shot face reenactment, which can animate a single face image to a different pose-and-expression (provided by a driving image) while keeping its original appearance. The core of our network is a novel mechanism called appearance adaptive normalization, which can effectively integrate the appearance information from the input image into our face generator by modulating the feature maps of the generator using the learned adaptive parameters. Furthermore, we specially design a local net to reenact the local facial components (i.e., eyes, nose and mouth) first, which is a much easier task for the network to learn and can in turn provide explicit anchors to guide our face generator to learn the global appearance and pose-and-expression. Extensive quantitative and qualitative experiments demonstrate the significant efficacy of our model compared with prior one-shot methods.

Point cloud completion is an interesting and challenging task in 3D vision, aiming to recover complete shapes from sparse and incomplete point clouds. Existing learning-based methods often require vast computation cost to achieve excellent performance, which limits their practical applications. In this paper, we propose a novel Recurrent Forward Network (RFNet), which is composed of three modules: Recurrent Feature Extraction (RFE), Forward Dense Completion (FDC) and Raw Shape Protection (RSP). The RFE extracts multiple global features from the incomplete point clouds for different recurrent levels, and the FDC generates point clouds in a coarse-to-fine pipeline. The RSP introduces details from the original incomplete models to refine the completion results. Besides, we propose a Sampling Chamfer Distance to better capture the shapes of models and a new Balanced Expansion Constraint to restrict the expansion distances from coarse to fine. According to the experiments on ShapeNet and KITTI, our network can achieve the state-of-the-art with lower memory cost and faster convergence.

Remarkable results have been achieved by DCNN based self-supervised depth estimation approaches. However, most of these approaches can only handle either day-time or night-time images, while their performance degrades for all-day images due to large domain shift and the variation of illumination between day and night images. To relieve these limitations, we propose a domain-separated network for self-supervised depth estimation of all-day images. Specifically, to relieve the negative influence of disturbing terms (illumination, etc.), we partition the information of day and night image pairs into two complementary sub-spaces: private and invariant domains, where the former contains the unique information (illumination, etc.) of day and night images and the latter contains essential shared information (texture, etc.). Meanwhile, to guarantee that the day and night images contain the same information, the domain-separated network takes the day-time images and corresponding night-time images (generated by GAN) as input, and the private and invariant feature extractors are learned by orthogonality and similarity loss, where the domain gap can be alleviated, thus better depth maps can be expected. Meanwhile, the reconstruction and photometric losses are utilized to estimate complementary information and depth maps effectively. Experimental results demonstrate that our approach achieves state-of-the-art depth estimation results for all-day images on the challenging Oxford RobotCar dataset, proving the superiority of our proposed approach. Code and data split are available at https://github.com/LINA-lln/ADDS-DepthNet.

Recent works have shown how realistic talking face images can be obtained under the supervision of geometry guidance, e.g., facial landmark or boundary. To alleviate the demand for manual annotations, in this paper, we propose a novel self-supervised hybrid model (DAE-GAN) that learns how to reenact face naturally given large amounts of unlabeled videos. Our approach combines two deforming autoencoders with the latest advances in the conditional generation. On the one hand, we adopt the deforming autoencoder to disentangle identity and pose representations. A strong prior in talking face videos is that each frame can be encoded as two parts: one for video-specific identity and the other for various poses. Inspired by that, we utilize a multi-frame deforming autoencoder to learn a pose-invariant embedded face for each video. Meanwhile, a multi-scale deforming autoencoder is proposed to extract pose-related information for each frame. On the other hand, the conditional generator allows for enhancing fine details and overall reality. It leverages the disentangled features to generate photo-realistic and pose-alike face images. We evaluate our model on VoxCeleb1 and RaFD dataset. Experiment results demonstrate the superior quality of reenacted images and the flexibility of transferring facial movements between identities.

Head pose estimation from RGB images without depth information is a challenging task due to the loss of spatial information as well as large head pose variations in the wild. The performance of existing landmark-free methods remains unsatisfactory as the quality of estimated pose is inferior. In this paper, we propose a novel three-branch network architecture, termed as Feature Decoupling Network (FDN), a more powerful architecture for landmark-free head pose estimation from a single RGB image. In FDN, we first propose a feature decoupling (FD) module to explicitly learn the discriminative features for each pose angle by adaptively recalibrating its channel-wise responses. Besides, we introduce a cross-category center (CCC) loss to constrain the distribution of the latent variable subspaces and thus we can obtain more compact and distinct subspaces. Extensive experiments on both in-the-wild and controlled environment datasets demonstrate that the proposed method outperforms other state-of-the-art methods based on a single RGB image and behaves on par with approaches based on multimodal input resources.

This paper presents a novel multi-identity face reenactment framework, named FReeNet, to transfer facial expressions from an arbitrary source face to a target face with a shared model. The proposed FReeNet consists of two parts: Unified Landmark Converter (ULC) and Geometry-aware Generator (GAG). The ULC adopts an encode-decoder architecture to efficiently convert expression in a latent landmark space, which significantly narrows the gap of the face contour between source and target identities. The GAG leverages the converted landmark to reenact the photorealistic image with a reference image of the target person. Moreover, a new triplet perceptual loss is proposed to force the GAG module to learn appearance and geometry information simultaneously, which also enriches facial details of the reenacted images. Further experiments demonstrate the superiority of our approach for generating photorealistic and expression-alike faces, as well as the flexibility for transferring facial expressions between identities.

Due to the difficulty in generating the effective descriptors which are robust to occlusion and viewpoint changes, place recognition for 3D point cloud remains an open issue. Unlike most of the existing methods that focus on extracting local, global, and statistical features of raw point clouds, our method aims at the semantic level that can be superior in terms of robustness to environmental changes. Inspired by the perspective of humans, who recognize scenes through identifying semantic objects and capturing their relations, this paper presents a novel semantic graph based approach for place recognition. First, we propose a novel semantic graph representation for the point cloud scenes by reserving the semantic and topological information of the raw point cloud. Thus, place recognition is modeled as a graph matching problem. Then we design a fast and effective graph similarity network to compute the similarity. Exhaustive evaluations on the KITTI dataset show that our approach is robust to the occlusion as well as viewpoint changes and outperforms the state-of-the-art methods with a large margin. Our code is available at: https://github.com/kxhit/SG_PR.

Spatiotemporal and motion features are two complementary and crucial information for video action recognition. Recent state-of-the-art methods adopt a 3D CNN stream to learn spatiotemporal features and another flow stream to learn motion features. In this work, we aim to efficiently encode these two features in a unified 2D framework. To this end, we first propose a STM block, which contains a Channel-wise SpatioTemporal Module (CSTM) to present the spatiotemporal features and a Channel-wise Motion Module (CMM) to efficiently encode motion features. We then replace original residual blocks in the ResNet architecture with STM blcoks to form a simple yet effective STM network by introducing very limited extra computation cost. Extensive experiments demonstrate that the proposed STM network outperforms the state-of-the-art methods on both temporal-related datasets (i.e., Something-Something v1 & v2 and Jester) and scene-related datasets (i.e., Kinetics-400, UCF-101, and HMDB-51) with the help of encoding spatiotemporal and motion features together.

Structured output support vector machine (SVM) based tracking algorithms have shown favorable performance recently. Nonetheless, the time-consuming candidate sampling and complex optimization limit their real-time applications. In this paper, we propose a novel large margin object tracking method which absorbs the strong discriminative ability from structured output SVM and speeds up by the correlation filter algorithm significantly. Secondly, a multimodal target detection technique is proposed to improve the target localization precision and prevent model drift introduced by similar objects or background noise. Thirdly, we exploit the feedback from high-confidence tracking results to avoid the model corruption problem. We implement two versions of the proposed tracker with the representations from both conventional hand-crafted and deep convolution neural networks (CNNs) based features to validate the strong compatibility of the algorithm. The experimental results demonstrate that the proposed tracker performs superiorly against several state-of-the-art algorithms on the challenging benchmark sequences while runs at speed in excess of 80 frames per second.

Acquiring the accurate 3-D position of a target person around a robot provides fundamental and valuable information that is applicable to a wide range of robotic tasks, including home service, navigation and entertainment. This paper presents a real-time robotic 3-D human tracking system which combines a monocular camera with an ultrasonic sensor by the extended Kalman filter (EKF). The proposed system consists of three sub-modules: monocular camera sensor tracking model, ultrasonic sensor tracking model and multi-sensor fusion. An improved visual tracking algorithm is presented to provide partial location estimation (2-D). The algorithm is designed to overcome severe occlusions, scale variation, target missing and achieve robust re-detection. The scale accuracy is further enhanced by the estimated 3-D information. An ultrasonic sensor array is employed to provide the range information from the target person to the robot and Gaussian Process Regression is used for partial location estimation (2-D). EKF is adopted to sequentially process multiple, heterogeneous measurements arriving in an asynchronous order from the vision sensor and the ultrasonic sensor separately. In the experiments, the proposed tracking system is tested in both simulation platform and actual mobile robot for various indoor and outdoor scenes. The experimental results show the superior performance of the 3-D tracking system in terms of both the accuracy and robustness.

Autonomous robots enjoy a wide popularity nowadays and have been applied in many applications, such as home security, entertainment, delivery, navigation and guidance. It is vital for robots to track objects accurately in real time in these applications, so it is necessary to focus on tracking algorithms to improve the robustness, speed and accuracy. In this paper, we propose a real-time robust object tracking algorithm based on a hierarchical ensemble framework which incorporates information including individual pixel features, local patches and holistic target models. The framework combines multiple ensemble models simultaneously instead of using a single ensemble model individually. A discriminative model which accounts for the matching degree of local patches is adopted via a bottom ensemble layer, and a generative model which exploits holistic templates is used to search for the object based on the middle ensemble layer as well as an adaptive Kalman filter. We test the proposed tracker on challenging benchmark image sequences. The experimental results demonstrate that the proposed tracker performs superiorly against several state-of-the-art algorithms, especially when the appearance changes dramatically and the occlusions occur.