Cheng-Han Chiang, Xiaofei Wang, Linjie Li, Chung-Ching Lin, Kevin Lin, Shujie Liu, Zhendong Wang, Zhengyuan Yang, Hung-yi Lee, Lijuan Wang. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

In this paper, we introduce zero-shot audio-video editing, a novel task that requires transforming original audio-visual content to align with a specified textual prompt without additional model training. To evaluate this task, we curate a benchmark dataset, AVED-Bench, designed explicitly for zero-shot audio-video editing. AVED-Bench includes 110 videos, each with a 10-second duration, spanning 11 categories from VGGSound. It offers diverse prompts and scenarios that require precise alignment between auditory and visual elements, enabling robust evaluation. We identify limitations in existing zero-shot audio and video editing methods, particularly in synchronization and coherence between modalities, which often result in inconsistent outcomes. To address these challenges, we propose AVED, a zero-shot cross-modal delta denoising framework that leverages audio-video interactions to achieve synchronized and coherent edits. AVED demonstrates superior results on both AVED-Bench and the recent OAVE dataset to validate its generalization capabilities. Our codes, data, and results are available at https://genjib.github.io/project_page/AVED/index.html

Human beings are endowed with a complementary learning system, which bridges the slow learning of general world dynamics with fast storage of episodic memory from a new experience. Previous video generation models, however, primarily focus on slow learning by pre-training on vast amounts of data, overlooking the fast learning phase crucial for episodic memory storage. This oversight leads to inconsistencies across temporally distant frames when generating longer videos, as these frames fall beyond the model's context window. To this end, we introduce SlowFast-VGen, a novel dual-speed learning system for action-driven long video generation. Our approach incorporates a masked conditional video diffusion model for the slow learning of world dynamics, alongside an inference-time fast learning strategy based on a temporal LoRA module. Specifically, the fast learning process updates its temporal LoRA parameters based on local inputs and outputs, thereby efficiently storing episodic memory in its parameters. We further propose a slow-fast learning loop algorithm that seamlessly integrates the inner fast learning loop into the outer slow learning loop, enabling the recall of prior multi-episode experiences for context-aware skill learning. To facilitate the slow learning of an approximate world model, we collect a large-scale dataset of 200k videos with language action annotations, covering a wide range of scenarios. Extensive experiments show that SlowFast-VGen outperforms baselines across various metrics for action-driven video generation, achieving an FVD score of 514 compared to 782, and maintaining consistency in longer videos, with an average of 0.37 scene cuts versus 0.89. The slow-fast learning loop algorithm significantly enhances performances on long-horizon planning tasks as well.

Recent developments in 2D visual generation have been remarkably successful. However, 3D and 4D generation remain challenging in real-world applications due to the lack of large-scale 4D data and effective model design. In this paper, we propose to jointly investigate general 3D and 4D generation by leveraging camera and object movements commonly observed in daily life. Due to the lack of real-world 4D data in the community, we first propose a data curation pipeline to obtain camera poses and object motion strength from videos. Based on this pipeline, we introduce a large-scale real-world 4D scene dataset: CamVid-30K. By leveraging all the 3D and 4D data, we develop our framework, GenXD, which allows us to produce any 3D or 4D scene. We propose multiview-temporal modules, which disentangle camera and object movements, to seamlessly learn from both 3D and 4D data. Additionally, GenXD employs masked latent conditions to support a variety of conditioning views. GenXD can generate videos that follow the camera trajectory as well as consistent 3D views that can be lifted into 3D representations. We perform extensive evaluations across various real-world and synthetic datasets, demonstrating GenXD's effectiveness and versatility compared to previous methods in 3D and 4D generation.

Recent advances in large language models have significantly improved textual reasoning through the effective use of Chain-of-Thought (CoT) and reinforcement learning. However, extending these successes to vision-language tasks remains challenging due to inherent limitations in text-only CoT, such as visual hallucinations and insufficient multimodal integration. In this paper, we introduce Point-RFT, a multimodal reasoning framework explicitly designed to leverage visually grounded CoT reasoning for visual document understanding. Our approach consists of two stages: First, we conduct format finetuning using a curated dataset of 71K diverse visual reasoning problems, each annotated with detailed, step-by-step rationales explicitly grounded to corresponding visual elements. Second, we employ reinforcement finetuning targeting visual document understanding. On ChartQA, our approach improves accuracy from 70.88% (format-finetuned baseline) to 90.04%, surpassing the 83.92% accuracy achieved by reinforcement finetuning relying solely on text-based CoT. The result shows that our grounded CoT is more effective for multimodal reasoning compared with the text-only CoT. Moreover, Point-RFT exhibits superior generalization capability across several out-of-domain visual document reasoning benchmarks, including CharXiv, PlotQA, IconQA, TabMWP, etc., and highlights its potential in complex real-world scenarios.

We introduce ThinkLite-VL, a family of visual reasoning models that achieve state-of-the-art (SoTA) performance using an order of magnitude fewer training samples, relying purely on reinforcement fine-tuning (RFT) self-improvement without any knowledge distillation. Our central insight is that sample difficulty critically influences RFT effectiveness: appropriately challenging examples can drive substantial reasoning improvements, even in low-data regimes. However, quantifying sample difficulty in a reliable and scalable manner remains non-trivial. To address this, we repurpose Monte Carlo Tree Search (MCTS) to measure sample difficulty via the number of reasoning iterations a vision-language model (VLM) requires to solve each instance. This MCTS-based selection procedure identifies samples that induce deeper reasoning while remaining solvable, allowing us to filter a high-quality subset from 70k open-source examples spanning math, natural image understanding, and chart comprehension. Using this approach, we select just 11k challenging samples for RFT on Qwen2.5-VL-7B-Instruct and 7.5k samples for Qwen2.5-VL-72B-Instruct. The resulting models, ThinkLite-VL-7B and ThinkLite-VL-72B, significantly outperform their respective base models across eight visual reasoning benchmarks. In particular, ThinkLite-VL-7B improves the average performance of Qwen2.5-VL-7B-Instruct by 7\% and surpasses all existing 7B-level models, as well as much larger models such as GPT-4o, O1 and Qwen2.5-VL-72B, achieving a new SoTA score of 75.1 on MathVista. ThinkLite-VL-72B further advances the SoTA frontier, achieving an accuracy of 79.7 on MathVista and an average benchmark improvement of 4.42 over the open-source SOTA. These results demonstrate that MCTS-guided difficulty filtering provides a scalable and effective path toward data-efficient self-improvement in multimodal reasoning.

Reinforcement learning (RL) has shown great effectiveness for fine-tuning large language models (LLMs) using tasks that are challenging yet easily verifiable, such as math reasoning or code generation. However, extending this success to visual perception in vision–language models (VLMs) has been impeded by the scarcity of vision-centric tasks that are simultaneously challenging and unambiguously verifiable. To this end, we introduce \textbf{ViCrit} (\textit{Visual Caption Hallucination Critic}), an RL proxy task that trains VLMs to localize a subtle, synthetic visual hallucination injected into paragraphs of human-written image captions. Starting from a 200-word captions, we inject a single, subtle visual description error—altering a few words on objects, attributes, counts, or spatial relations—and task the model to pinpoint the corrupted span given the image and the modified caption. This formulation preserves the full perceptual difficulty while providing a binary, exact-match reward that is easy to compute and unambiguous. Models trained with the \textbf{ViCrit Task} exhibit substantial gains across a variety of VL benchmarks. Crucially, the improvements transfer beyond natural-image training data to abstract image reasoning and visual math, showing promises of learning to perceive rather than barely memorizing seen objects. To facilitate evaluation, we further introduce \textbf{ViCrit-Bench}, a category-balanced diagnostic benchmark that systematically probes perception errors across diverse image domains and error types. Together, our results demonstrate that fine-grained hallucination criticism is an effective and generalizable objective for enhancing visual perception in VLMs.

Generative AI has made significant strides in computer vision, particularly in text-driven image/video synthesis (T2I/T2V). Despite the notable advancements, it remains challenging in human-centric content synthesis such as realistic dance generation. Current methodologies, primarily tailored for human motion transfer, encounter difficulties when confronted with real-world dance scenarios (e.g., social media dance), which require to generalize across a wide spectrum of poses and intricate human details. In this paper, we depart from the traditional paradigm of human motion transfer and emphasize two additional critical attributes for the synthesis of human dance content in social media contexts: (i) Generalizability: the model should be able to generalize beyond generic human viewpoints as well as unseen human subjects, backgrounds, and poses; (ii) Compositionality: it should allow for the seamless composition of seen/unseen subjects, backgrounds, and poses from different sources. To address these challenges, we introduce Disco, which includes a novel model architecture with disentangled control to improve the compositionality of dance synthesis, and an effective human attribute pre-training for better generalizability to unseen humans. Extensive qualitative and quantitative results demonstrate that DISCO can generate high-quality human dance images and videos with diverse appearances and flexible motions. Code is available at https://disco-dance.github.io/.

We present MM-Narrator, a novel system leveraging GPT-4 with multimodal in-context learning for the gener-ation of audio descriptions (AD). Unlike previous methods that primarily focused on downstream fine-tuning with short video clips, MM-Narrator excels in generating precise audio descriptions for videos of extensive lengths, even be-yond hours, in an autoregressive manner. This capability is made possible by the proposed memory-augmented generation process, which effectively utilizes both the short-term textual context and long-term visual memory through an efficient register-and-recall mechanism. These contextual memories compile pertinent past information, including storylines and character identities, ensuring an accurate tracking and depicting of story-coherent and character-centric audio descriptions. Maintaining the training-free design of MM-Narrator, we further propose a complexity-based demonstration selection strategy to largely enhance its multi-step reasoning capability via few-shot multimodal in-context learning (MM-ICL). Experimental results on MAD-eval dataset demonstrate that MM-Narrator consistently outperforms both the existing fine-tuning-based approaches and LLM-based approaches in most scenarios, as measured by standard evaluation metrics. Additionally, we introduce the first segment-based evaluator for recurrent text generation. Empowered by GPT-4, this evaluator comprehensively reasons and marks AD generation performance in various extendable dimensions.

This paper presents a novel approach to object completion, with the primary goal of reconstructing a complete object from its partially visible components. Our method, named MaskComp, delineates the completion process through iterative stages of generation and segmentation. In each iteration, the object mask is provided as an additional condition to boost image generation, and, in return, the generated images can lead to a more accurate mask by fusing the segmentation of images. We demonstrate that the combination of one generation and one segmentation stage effectively functions as a mask denoiser. Through alternation between the generation and segmentation stages, the partial object mask is progressively refined, providing precise shape guidance and yielding superior object completion results. Our experiments demonstrate the superiority of MaskComp over existing approaches, e.g., ControlNet and Stable Diffusion, establishing it as an effective solution for object completion.

Image diffusion distillation achieves high-fidelity generation with very few sampling steps. However, directly applying these techniques to video models results in unsatisfied frame quality. This issue arises from the limited frame appearance quality in public video datasets, affecting the performance of both teacher and student video diffusion models. Our study aims to improve video diffusion distillation and meanwhile enabling the student model to improve frame appearance using the abundant high-quality image data. To this end, we propose motion consistency models (MCM), a single-stage video diffusion distillation method that disentangles motion and appearance learning. Specifically, MCM involves a video consistency model that distills motion from the video teacher model, and an image discriminator that boosts frame appearance to match high-quality image data. However, directly combining these components leads to two significant challenges: a conflict in frame learning objectives, where video distillation learns from low-quality video frames while the image discriminator targets high-quality images, and training-inference discrepancies due to the differing quality of video samples used during training and inference. To address these challenges, we introduce disentangled motion distillation and mixed trajectory distillation. The former applies the distillation objective solely to the motion representation, while the latter mitigates training-inference discrepancies by mixing distillation trajectories from both the low- and high-quality video domains. Extensive experiments show that our MCM achieves state-of-the-art video diffusion distillation performance. Additionally, our method can enhance frame quality in video diffusion models, producing frames with high aesthetic value or specific styles.

Traditional methods of reconstructing 3D human pose and mesh from single images rely on paired image-mesh datasets, which can be difficult and expensive to obtain. Due to this limitation, model scalability is constrained as well as reconstruction performance. Towards addressing the challenge, we introduce Mesh Pre-Training (MPT), an effective pre-training strategy that leverages large amounts of MoCap data to effectively perform pre-training at scale. We introduce the use of MoCap-generated heatmaps as input representations to the mesh regression transformer and propose a Masked Heatmap Modeling approach for improving pre-training performance. This study demonstrates that pre-training using the proposed MPT allows our models to perform effective inference without requiring fine-tuning. We further show that fine-tuning the pre-trained MPT model considerably improves the accuracy of human mesh reconstruction from single images. Experimental results show that MPT outperforms previous state-of-the-art methods on Human3.6M and 3DPW datasets. As a further application, we benchmark and study MPT on the task of 3D hand reconstruction, showing that our generic pre-training scheme generalizes well to hand pose estimation and achieves promising reconstruction performance.

We present a unified framework for camera-space 3D hand pose estimation from a single RGB image based on 3D implicit representation. As opposed to recent works, most of which first adopt holistic or pixel-level dense regression to obtain relative 3D hand pose and then follow with complex second-stage operations for 3D global root or scale recovery, we propose a novel unified 3D dense regression scheme to estimate camera-space 3D hand pose via dense 3D point-wise voting in camera frustum. Through direct dense modeling in 3D domain inspired by Pixel-aligned Implicit Functions for 3D detailed reconstruction, our proposed Neural Voting Field (NVF) fully models 3D dense local evidence and hand global geometry, helping to alleviate common 2D-to-3D ambiguities. Specifically, for a 3D query point in camera frustum and its pixel-aligned image feature, NVF, represented by a Multi-Layer Perceptron, regresses: (i) its signed distance to the hand surface; (ii) a set of 4D offset vectors (1D voting weight and 3D directional vector to each hand joint). Following a vote-casting scheme, 4D offset vectors from near-surface points are selected to calculate the 3D hand joint coordinates by a weighted average. Experiments demonstrate that NVF outperforms existing state-of-the-art algorithms on FreiHAND dataset for camera-space 3D hand pose estimation. We also adapt NVF to the classic task of root-relative 3D hand pose estimation, for which NVF also obtains state-of-the-art results on HO3D dataset.

Unified vision-language frameworks have greatly advanced in recent years, most of which adopt an encoder-decoder architecture to unify image-text tasks as sequence-to-sequence generation. However, existing video-language (VidL) models still require task-specific designs in model architecture and training objectives for each task. In this work, we explore a unified VidL framework LAVENDER, where Masked Language Modeling [13] (MLM) is used as the common interface for all pre-training and downstream tasks. Such unification leads to a simplified model architecture, where only a lightweight MLM head, instead of a decoder with much more parameters, is needed on top of the multimodal encoder. Surprisingly, experimental results show that this unified framework achieves competitive performance on 14 VidL benchmarks, covering video question answering, text-to-video retrieval and video captioning. Extensive analyses further demonstrate Lavender can (i) seamlessly support all downstream tasks with just a single set of parameter values when multi-task fine-tuned; (ii) generalize to various downstream tasks with limited training samples; and (iii) enable zero-shot evaluation on video question answering tasks. Code is available at https://github.com/microsoft/LAVENDER.

The most recent efforts in video matting have focused on eliminating trimap dependency since trimap annotations are expensive and trimap-based methods are less adaptable for real-time applications. Despite the latest tripmapfree methods showing promising results, their performance often degrades when dealing with highly diverse and unstructured videos. We address this limitation by introducing Adaptive Matting for Dynamic Videos, termed AdaM, which is a framework designed for simultaneously differentiating foregrounds from backgrounds and capturing alpha matte details of human subjects in the foreground. Two interconnected network designs are employed to achieve this goal: (1) an encoder-decoder network that produces alpha mattes and intermediate masks which are used to guide the transformer in adaptively decoding foregrounds and backgrounds, and (2) a transformer network in which long- and short-term attention combine to retain spatial and temporal contexts, facilitating the decoding of foreground details. We benchmark and study our methods on recently introduced datasets, showing that our model notably improves matting realism and temporal coherence in complex real-world videos and achieves new best-in-class generalizability. Further details and examples are available at https://github.com/microsoft/AdaM.

This study explores the concept of equivariance in vision-language foundation models (VLMs), focusing specifically on the multimodal similarity function that is not only the major training objective but also the core delivery to support downstream tasks. Unlike the existing image-text similarity objective which only categorizes matched pairs as similar and unmatched pairs as dissimilar, equivariance also requires similarity to vary faithfully according to the semantic changes. This allows VLMs to generalize better to nuanced and unseen multimodal compositions. However, modeling equivariance is challenging as the ground truth of semantic change is difficult to collect. For example, given an image-text pair about a dog, it is unclear to what extent the similarity changes when the pixel is changed from dog to cat? To this end, we propose EqSim, a regularization loss that can be efficiently calculated from any two matched training pairs and easily pluggable into existing image-text retrieval fine-tuning. Meanwhile, to further diagnose the equivariance of VLMs, we present a new challenging benchmark EqBen. Compared to the existing evaluation sets, EqBen is the first to focus on "visual-minimal change". Extensive experiments show the lack of equivariance in current VLMs1and validate the effectiveness of EqSim2.

Non-Foster (NF) circuits are becoming increasingly popular in filters and matching networks as they allow the cancellation of reactances and thus enable wideband matching. With many kinds of NF circuits available in the literature, this paper aims to categorize and provide an in-depth discussion of each variation of these circuits upon their implementation, advantages and challenges.

The paper introduces PaintSeg, a new unsupervised method for segmenting objects without any training. We propose an adversarial masked contrastive painting (AMCP) process, which creates a contrast between the original image and a painted image in which a masked area is painted using off-the-shelf generative models. During the painting process, inpainting and outpainting are alternated, with the former masking the foreground and filling in the background, and the latter masking the background while recovering the missing part of the foreground object. Inpainting and outpainting, also referred to as I-step and O-step, allow our method to gradually advance the target segmentation mask toward the ground truth without supervision or training. PaintSeg can be configured to work with a variety of prompts, e.g. coarse masks, boxes, scribbles, and points. Our experimental results demonstrate that PaintSeg outperforms existing approaches in coarse mask-prompt, box-prompt, and point-prompt segmentation tasks, providing a training-free solution suitable for unsupervised segmentation. Code: https://github.com/lxa9867/PaintSeg.

The canonical approach to video captioning dictates a caption generation model to learn from offline-extracted dense video features. These feature extractors usually operate on video frames sampled at a fixed frame rate and are often trained on image/video understanding tasks, without adaption to video captioning data. In this work, we present SwinBERT, an end-to-end transformer-based model for video captioning, which takes video frame patches directly as inputs, and outputs a natural language description. Instead of leveraging multiple 2D/3D feature extractors, our method adopts a video transformer to encode spatial-temporal representations that can adapt to variable lengths of video input without dedicated design for different frame rates. Based on this model architecture, we show that video captioning can benefit significantly from more densely sampled video frames as opposed to previous successes with sparsely sampled video frames for video-and-language understanding tasks (e.g., video question answering). Moreover, to avoid the inherent redundancy in consecutive video frames, we propose adaptively learning a sparse attention mask and optimizing it for task-specific performance improvement through better long-range video sequence modeling. Through extensive experiments on 5 video captioning datasets, we show that Swinbert achieves across-the-board performance improvements over previous methods, often by a large margin. The learned sparse attention masks in addition push the limit to new state of the arts, and can be transferred between different video lengths and between different datasets. Code is available at https://github.com/microsoft/SwinBERT.

We present a cross-modal Transformer-based frame-work, which jointly encodes video data and text labels for zero-shot action recognition (ZSAR). Our model employs a conceptually new pipeline by which visual representations are learned in conjunction with visual-semantic associations in an end-to-end manner. The model design provides a natural mechanism for visual and semantic representations to be learned in a shared knowledge space, whereby it encourages the learned visual embedding to be discriminative and more semantically consistent. In zero-shot inference, we devise a simple semantic transfer scheme that embeds semantic relatedness information between seen and unseen classes to composite unseen visual prototypes. Accordingly, the discriminative features in the visual structure could be preserved and exploited to alleviate the typical zero-shot issues of information loss, semantic gap, and the hubness problem. Under a rigorous zero-shot setting of not pre-training on additional datasets, the experiment results show our model considerably improves upon the state of the arts in ZSAR, reaching encouraging top-1 accuracy on UCF 101, HMDB51, and ActivityNet benchmark datasets. Code will be made available.11https://github.com/microsoft/ResT

The best beam steering directions are estimated through beam training, which is one of the most important and challenging tasks in millimeter-wave and sub-terahertz communications. Novel array architectures and signal processing techniques are required to avoid prohibitive beam training overhead associated with large antenna arrays and narrow beams. In this work, we leverage recent developments in true-time-delay (TTD) arrays with large delay-bandwidth products to accelerate beam training using frequency-dependent probing beams. We propose and study two TTD architecture candidates, including analog and hybrid analog-digital arrays, that can facilitate beam training with only one wideband pilot. We also propose a suitable algorithm that requires a single pilot to achieve high-accuracy estimation of angle of arrival. The proposed array architectures are compared in terms of beam training requirements and performance, robustness to practical hardware impairments, and power consumption. The findings suggest that the analog and hybrid TTD arrays achieve a sub-degree beam alignment precision with 66% and 25% lower power consumption than a fully digital array, respectively. Our results yield important design trade-offs among the basic system parameters, power consumption, and accuracy of angle of arrival estimation in fast TTD beam training.

This paper presents an integrated ultra-low-power (ULP) wireless sensor system-on-chip (SoC) that can be used for voltage sensing in both Internet of Things applications and bio-potential monitoring. In order to increase the energy efficiency of the analog front-end (AFE), we propose a noise and power efficient push-pull low noise instrumentation amplifier (LNIA) with a built-in ripple reduction loop based on capacitor reuse. A low-power ISM-band harmonic injection locked on-off-keying transmitter (OOK-TX) is also implemented for energy efficient wireless connectivity. Circuit implementations, design considerations, and detailed analysis are presented to improve the overall energy efficiencies of the SoC including the AFE, TX and, the power management unit. The proposed ULP-SoC is fabricated in 130 nm CMOS technology with a total area of 1.92 mm2. The total power consumption of the proposed system-on-chip is 197.1 μW which is one of the lowest among state-of-the-art wireless sensor SoC.

We propose a modified variational autoencoder (VAE) architecture built on top of Mask R-CNN for instance-level video segmentation and tracking. The method builds a shared encoder and three parallel decoders, yielding three disjoint branches for predictions of future frames, object detection boxes, and instance segmentation masks. To effectively solve multiple learning tasks, we introduce a Gaussian Process model to enhance the statistical representation of VAE by relaxing the prior strong independent and identically distributed (iid) assumption of conventional VAEs and allowing potential correlations among extracted latent variables. The network learns embedded spatial interdependence and motion continuity in video data and creates a representation that is effective to produce high-quality segmentation masks and track multiple instances in diverse and unstructured videos. Evaluation on a variety of recently introduced datasets shows that our model outperforms previous methods and achieves the new best in class performance.

The development of an autonomous tracking system which could accurately locate and transfer higher quantities of power wirelessly to deep-tissue catheter implants in a cost-effective manner would prove very useful in medicine. This work presents a backscatter-based energy transfer and tracking system for future deep-tissue implants. To transfer power, an open-top ultrasonic transducer would be placed near the skin, the ultrasonic wave can power circuits on the tip of the catheter inside the patient. For the tracking system, transducers use ultrasonic backscattering and time-of-arrival to precisely locate the tip of the catheter. To show this concept will work in future applications, the two major components of the system were tested and validated using commercial off-the-shelf devices with air as medium. We were able to show a one-way 0.34mm localization resolution and track an object in two dimensions with 6.32mm accuracy on the x-axis and 4.27mm on the y-axis. Ultrasonic Wireless Power Transfer (WPT) was modeled using a linear rectangular array approximation via an open-source MATLAB Toolbox known as k-Wave. The developed array model was experimentally validated at distances from 11mm to 310mm using a linear rectangular array approximation of commercially available transducers for transfer efficiencies through air of up to 81.9 percent.

On-chip integrated oscillators are highly needed for the next-generation low-cost sensors used in IoT/biomedical applications to realize sub-millimeter-sized packages at ultra-low-power and fast startup. This paper presents a 25.6μW phase-locked relaxation oscillator (PLL-ROSC) with 8.972ps period jitter using an energy-efficient PLL. A self-locking mechanism is proposed that tracks the temperature variation from -20 to 100°C A current-reuse differential ROSC with boosted output swing is employed to improve noise performance and reduce the power consumption. The simulated power consumption is 15x better than state-of-art with -237.2dBc/Hz FoM and <; 1μs startup time using 180nm CMOS technology.

This paper presents a prior-less method for tracking and clustering an unknown number of human faces and maintaining their individual identities in unconstrained videos. The key challenge is to accurately track faces with partial occlusion and drastic appearance changes in multiple shots resulting from significant variations of makeup, facial expression, head pose and illumination. To address this challenge, we propose a new multi-face tracking and re-identification algorithm, which provides high accuracy in face association in the entire video with automatic cluster number generation, and is robust to outliers. We develop a co-occurrence model of multiple body parts to seamlessly create face tracklets, and recursively link tracklets to construct a graph for extracting clusters. A Gaussian Process model is introduced to compensate the deep feature insufficiency, and is further used to refine the linking results. The advantages of the proposed algorithm are demonstrated using a variety of challenging music videos and newly introduced body-worn camera videos. The proposed method obtains significant improvements over the state of the art [51], while relying less on handling video-specific prior information to achieve high performance.

In this work, we examine the resiliency of a state-of-the-art end-to-end video summarization (VS) application that serves as a representative emerging workload in the domain of real time edge computing. The VS application constitutes key video and image analytic elements that are processed by embedded systems aboard unmanned aerial vehicles (UAVs). Real-time performance and energy constraints motivate the consideration of approximations to the VS algorithm. However, mission-critical UAV applications also demand stringent levels of resilience to soft errors that are exacerbated with higher altitude. In this work, we study the effects of three different types of software approximations on the application level resiliency (to soft errors) of the VS algorithm. We show that our approximations yield significant energy savings (up to 68%), with commensurate improvement in performance, without a degradation in the application resilience. Further, by proposing a novel quality metric (appropriate for the UAV vision analytics domain) for the summarized video output, we show that even though the rate of Silent Data Corruptions (SDCs) increases slightly (<2%), the impact of these SDCs on output quality is limited. Thus, we conclude that software approximation can be utilized to achieve significant gains in performance and energy without affecting application resiliency.

The goal of image stitching is to create natural-looking mosaics free of artifacts that may occur due to relative camera motion, illumination changes, and optical aberrations. In this paper, we propose a novel stitching method, that uses a smooth stitching field over the entire target image, while accounting for all the local transformation variations. Computing the warp is fully automated and uses a combination of local homography and global similarity transformations, both of which are estimated with respect to the target. We mitigate the perspective distortion in the non-overlapping regions by linearizing the homography and slowly changing it to the global similarity. The proposed method is easily generalized to multiple images, and allows one to automatically obtain the best perspective in the panorama. It is also more robust to parameter selection, and hence more automated compared with state-of-the-art methods. The benefits of the proposed approach are demonstrated using a variety of challenging cases.

In this paper, we propose a hierarchical computational system architecture to support the target domain of realtime mobile computing in the context of unmanned aerial vehicles (UAVs). The overall architectural vision includes support for system resilience in the presence of uncertainties in the operational environment of surveillance UAVs. We report measurement-based results that are obtained from a UAV proxy demonstration apparatus. The apparatus consists of a Raspberry Pi (RPi) board that serves as an on-board UAV computer, working with support from a laptop that serves as the on-ground computing infrastructure where an operator "consumes" video information received from the UAV. We quantify the gap between the on-board UAV camera frame rate (input) and the on-ground operator-observed frame rate (output) for a specialized class of computer vision applications germane to the UAV-based aerial surveillance domain. The goal is to keep the frame rate observed by the ground operator as close (or ideally equal) to the on-board UAV camera frame rate (i.e. to preserve the real-time aspect) despite the unstable bandwidth availability in the channel connecting both ends. The proposed hierarchical approach significantly outperforms two considered baselines: one in which computation takes place entirely on the UAV computer and another in which computation takes place entirely on the ground. This improved performance is due to a more balanced resource sharing between the on-board UAV computer and UAV-to-ground communication channel. Later, we show how the observed frame rate improves when the RPi board is replaced with an NVIDIA Jetson TK1 board. Based on the observations gleaned from these "proxy" experiments, we sketch the fundamentals of our ongoing work in model-based predictive analysis of resilient "UAV swarm" computational architectures of the future.

The importance of the internet-of-things (IOT) is now an established reality. With that backdrop, the phenomenal emergence of cameras/sensors mounted on unmanned aerial, ground and marine vehicles (UAVs, UGVs, UMVs) and body worn cameras is a notable new development. The swarms of cameras and real-time computing thereof are at the heart of new technologies like connected cars, drone-based city-wide surveillance and precision agriculture, etc. Smart computer vision algorithms (with or without dynamic learning) that enable object recognition and tracking, supported by baseline video content summarization or 2D/3D image reconstruction of the scanned environment are at the heart of such new applications. In this article, we summarize our recent innovations in this space. We focus primarily on algorithms and architectural design considerations for video summarization systems.

The advances of mobile computing and sensor technology have turned the mobile devices into powerful instruments. The integration of thermal and visual cameras extends the capability of computer vision, due to the fact that both images reveal different characteristics in images, however, image alignment is a challenge. This paper proposes an effective approach to align image pairs for event detection on mobile through image recognition. We leverage thermal and visual cameras as multi-modality sources for image recognition. By analyzing the heat pattern, the proposed APP can identify the heating sources and help users inspect their house heating system, on the other hand, with applying image recognition, the proposed APP furthermore can help field workers identify the asset condition and provide the guidance to solve their issues.

It is estimated that less than five percent of videos are currently analyzed to any degree. In addition to petabyte-sized multimedia archives, continuing innovations in optics, imaging sensors, camera arrays, (aerial) platforms, and storage technologies indicates that for the foreseeable future existing and new applications will continue to generate enormous volumes of video imagery. Contextual video summarizations and activity maps offers one innovative direction to tackling this Big Data problem in computer vision. The goal of this work is to develop semi-automatic exploitation algorithms and tools to increase utility, dissemination and usage potential by providing quick dynamic overview geospatial mosaics and motion maps. We present a framework to summarize (multiple) video streams from unmanned aerial vehicles (UAV) or drones which have very different characteristics compared to structured commercial and consumer videos that have been analyzed in the past. Using both metadata geospatial characteristics of the video combined with fast low-level image-based algorithms, the proposed method first generates mini-mosaics that can then be combined into geo-referenced meta-mosaics imagery. These geospatial maps enable rapid assessment of hours long videos with arbitrary spatial coverage from multiple sensors by generating quick look imagery, composed of multiple mini-mosaics, summarizing spatiotemporal dynamics such as coverage, dwell time, activity, etc. The overall summarization pipeline was tested on several DARPA Video and Image Retrieval and Analysis Tool (VIRAT) datasets. We evaluate the effectiveness of the proposed video summarization framework using metrics such as compression and hours of viewing time.

This paper proposes a new ego-motion estimation and background/foreground classification method to effectively segment moving objects from videos captured by a moving camera on a moving platform. Existing methods for moving-camera detecting impose serious constraints. In our approach, ellipsoid scene shape is applied in the motion model and a complicated ego-motion estimation formula is derived. Genetic algorithm is introduced to accurately solve ego-motion parameters. After motion recovery, noisy result is refined by motion vector correlation and foreground is classified by pixel level probability model. Experiment results show that the method demonstrates significant detecting performance without further restrictions and performs effectively in complex detecting environment.