Human pose and shape estimation (HPS) has attracted increasing attention in recent years. While most existing studies focus on HPS from 2D images or videos with inherent depth ambiguity, there are surging need to investigate HPS from 3D point clouds as depth sensors have been frequently employed in commercial devices. However, real-world sensory 3D points are usually noisy and incomplete, and also human bodies could have different poses of high diversity. To tackle these challenges, we propose a principled framework, PointHPS, for accurate 3D HPS from point clouds captured in real-world settings, which iteratively refines point features through a cascaded architecture. Specifically, each stage of PointHPS performs a series of downsampling and upsampling operations to extract and collate both local and global cues, which are further enhanced by two novel modules: 1) Cross-stage Feature Fusion (CFF) for multi-scale feature propagation that allows information to flow effectively through the stages, and 2) Intermediate Feature Enhancement (IFE) for body-aware feature aggregation that improves feature quality after each stage. To facilitate a comprehensive study under various scenarios, we conduct our experiments on two large-scale benchmarks, comprising i) a dataset that features diverse subjects and actions captured by real commercial sensors in a laboratory environment, and ii) controlled synthetic data generated with realistic considerations such as clothed humans in crowded outdoor scenes. Extensive experiments demonstrate that PointHPS, with its powerful point feature extraction and processing scheme, outperforms State-of-the-Art methods by significant margins across the board. Homepage: https://caizhongang.github.io/projects/PointHPS/.

Video-to-audio (V2A) generation extends image-to-audio generation (I2A) by introducing consecutive frames that provide essential temporal cues for audio synthesis. However, existing conditional diffusion-based V2A methods typically enhance visual conditioning with additional audio-visual supervision, acoustic structure prediction, or reasoning from large multimodal models, requiring extra networks or strong inductive biases. Inspired by recent advances in visual representation learning, we introduce TD-V2A, which leverages temporal differences (TD) as the key representation that distinguishes V2A from I2A, enriching visual conditioning with minimal architectural modification. We first investigate TD at both the frame and feature levels to identify the most effective representation level at which TD complements visual representations. Based on these findings, we develop a hierarchically continual learning strategy and an annealed temporal differences guidance method to progressively learn and exploit TD information during diffusion training and sampling process, respectively. Extensive experiments on benchmark datasets demonstrate that effectively exploiting TD through our proposed framework significantly improves end-to-end V2A generation quality, even outperforming dedicated V2A representations such as contrastive audio-visual pretraining.

Manipulating dynamic objects remains an open challenge for Vision-Language-Action (VLA) models, which, despite strong generalization in static manipulation, struggle in dynamic scenarios requiring rapid perception, temporal anticipation, and continuous control. We present DynamicVLA, a framework for dynamic object manipulation that integrates temporal reasoning and closed-loop adaptation through three key designs: 1) a compact 0.4B VLA using a convolutional vision encoder for spatially efficient, structurally faithful encoding, enabling fast multimodal inference; 2) Continuous Inference, enabling overlapping reasoning and execution for lower latency and timely adaptation to object motion; and 3) Latent-aware Action Streaming, which bridges the perception-execution gap by enforcing temporally aligned action execution. To fill the missing foundation of dynamic manipulation data, we introduce the Dynamic Object Manipulation (DOM) benchmark, built from scratch with an auto data collection pipeline that efficiently gathers 200K synthetic episodes across 2.8K scenes and 206 objects, and enables fast collection of 2K real-world episodes without teleoperation. Extensive evaluations demonstrate remarkable improvements in response speed, perception, and generalization, positioning DynamicVLA as a unified framework for general dynamic object manipulation across embodiments.

Embedded camera is widely adopted in multimedia and optical camera communication (OCC) applications thanks to its images capturing ability and ease of deployment. However, it faces a serious conflict between integrated image-based object recognition and optical signal demodulation. On one hand, the captured frames suffer from low light and stripe patterns incurred by OCC, resulting in degraded visual contrast and confusion of object features. On the other hand, objects in communication link interfere with the received optical signal, leading to poor decoding performance. In this paper, we propose TransferCam as a solution to address the conflict. Essentially, TransferCam learns the features of OCC frames in both the pixel and frequency domains, and reconstructs the normal clean objects via style transfer for improved recognition. Meanwhile, TransferCam utilizes the reconstructed frames to identify the regions affected by object and background, and employs an object-aware sampling method to avoid the interference for OCC decoding. TransferCam provides integrated object recognition and communication, and extensive evaluations fully demonstrate its feasibility in real-scene application.

Modeling dynamic 3D environments from LiDAR sequences is central to building reliable 4D worlds for autonomous driving and embodied AI. Existing generative frameworks, however, often treat all spatial regions uniformly, overlooking the varying uncertainty across real-world scenes. This uniform generation leads to artifacts in complex or ambiguous regions, limiting realism and temporal stability. In this work, we present U4D, an uncertainty-aware framework for 4D LiDAR world modeling. Our approach first estimates spatial uncertainty maps from a pretrained segmentation model to localize semantically challenging regions. It then performs generation in a "hard-to-easy" manner through two sequential stages: (1) uncertainty-region modeling, which reconstructs high-entropy regions with fine geometric fidelity, and (2) uncertainty-conditioned completion, which synthesizes the remaining areas under learned structural priors. To further ensure temporal coherence, U4D incorporates a mixture of spatio-temporal (MoST) block that adaptively fuses spatial and temporal representations during diffusion. Extensive experiments show that U4D produces geometrically faithful and temporally consistent LiDAR sequences, advancing the reliability of 4D world modeling for autonomous perception and simulation.

Hypothesis. Artificial general intelligence is, at its core, a compression problem. Effective compression demands resonance: deep learning scales best when its architecture aligns with the fundamental structure of the data. These are the fundamental principles. Yet, modern vision architectures have strayed from these truths: visual signals are highly redundant, while discriminative information, the surprise, is sparse. Current models process dense pixel grids uniformly, wasting vast compute on static background rather than focusing on the predictive residuals that define motion and meaning. We argue that to solve visual understanding, we must align our architectures with the information-theoretic principles of video, i.e., Codecs. Method. OneVision-Encoder encodes video by compressing predictive visual structure into semantic meaning. By adopting Codec Patchification, OV-Encoder abandons uniform computation to focus exclusively on the 3.1 Evidence. The results validate our core hypothesis: efficiency and accuracy are not a trade-off; they are positively correlated. When integrated into LLM, it consistently outperforms strong vision backbones such as Qwen3-ViT and SigLIP2 across 16 image, video, and document understanding benchmarks, despite using substantially fewer visual tokens and pretraining data. Notably, on video understanding tasks, OV-Encoder achieves an average improvement of 4.1

Humans acquire knowledge through three cognitive stages: perceiving information, comprehending knowledge, and adapting knowledge to solve novel problems. Videos serve as an effective medium for knowledge acquisition, facilitating a natural progression through these learning stages. However, existing video benchmarks fail to evaluate the knowledge acquisition capabilities of Large Multimodal Models (LMMs). To address this gap, we introduce Video-MMMU, a multi-modal, multi-discipline, multi-track benchmark that evaluates LMMs’ ability to acquire knowledge from college-level, educational videos. Video-MMMU features a collection of 300 videos and 900 human-annotated questions across six disciplines, evaluating knowledge acquisition through stage-aligned question-answer pairs: Perception, Comprehension, and Adaptation. Beyond measuring final accuracy, Video-MMMU proposes the performance gain metric that quantifies an LMM’s learning gain from video, shifting the focus of evaluation from absolute performance to learning efficiency. Our evaluation reveals a substantial gap between human learners and current LMMs, highlighting the need to improve models’ ability to learn and adapt knowledge from video content.

AIGC has rapidly expanded from text-to-image generation toward high-quality multimodal synthesis across video and audio. Within this context, joint audio-video generation (JAVG) has emerged as a fundamental task that produces synchronized and semantically aligned sound and vision from textual descriptions. However, compared with advanced commercial models such as Veo3, existing open-source methods still suffer from limitations in generation quality, temporal synchrony, and alignment with human preferences. To bridge the gap, this paper presents JavisDiT++, a concise yet powerful framework for unified modeling and optimization of JAVG. First, we introduce a modality-specific mixture-of-experts (MS-MoE) design that enables cross-modal interaction efficacy while enhancing single-modal generation quality. Then, we propose a temporal-aligned RoPE (TA-RoPE) strategy to achieve explicit, frame-level synchronization between audio and video tokens. Besides, we develop an audio-video direct preference optimization (AV-DPO) method to align model outputs with human preference across quality, consistency, and synchrony dimensions. Built upon Wan2.1-1.3B-T2V, our model achieves state-of-the-art performance merely with around 1M public training entries, significantly outperforming prior approaches in both qualitative and quantitative evaluations. Comprehensive ablation studies have been conducted to validate the effectiveness of our proposed modules. All the code, model, and dataset are released at https://JavisVerse.github.io/JavisDiT2-page.

3D human motion generation plays a crucial role in many applications. However, in previous studies, motion generation algorithms often focused on a limited number of motion categories or required complex methods to control the desired outputs. The goal of this chapter is to address the question of how to use natural language as the sole control signal to generate diverse and realistic motion sequences while enabling users to perform fine-grained control. This form of control greatly lowers the threshold for using this technology, allowing even layman users to benefit from the technology. Specifically, this chapter first uses MotionDiffuse as an example to introduce a classic text-driven motion generation pipeline design and studies how to fully integrate the text modality with the motion modality. Then, the chapter discusses how to leverage the concept of retrieval-augmented generation to further enhance the model's generation performance. Finally, this chapter introduces two effective solutions to solve the problem of fine-grained motion generation and demonstrates how such generation capability can benefit real-world applications.

Autonomous systems are increasingly deployed in open and dynamic environments – from city streets to aerial and indoor spaces – where perception models must remain reliable under sensor noise, environmental variation, and platform shifts. However, even state-of-the-art methods often degrade under unseen conditions, highlighting the need for robust and generalizable robot sensing. The RoboSense 2025 Challenge is designed to advance robustness and adaptability in robot perception across diverse sensing scenarios. It unifies five complementary research tracks spanning language-grounded decision making, socially compliant navigation, sensor configuration generalization, cross-view and cross-modal correspondence, and cross-platform 3D perception. Together, these tasks form a comprehensive benchmark for evaluating real-world sensing reliability under domain shifts, sensor failures, and platform discrepancies. RoboSense 2025 provides standardized datasets, baseline models, and unified evaluation protocols, enabling large-scale and reproducible comparison of robust perception methods. The challenge attracted 143 teams from 85 institutions across 16 countries, reflecting broad community engagement. By consolidating insights from 23 winning solutions, this report highlights emerging methodological trends, shared design principles, and open challenges across all tracks, marking a step toward building robots that can sense reliably, act robustly, and adapt across platforms in real-world environments.

A central goal of biomedicine is to understand, predict and ultimately control the dynamic mechanisms by which biological systems respond to perturbations, disease progression and therapeutic intervention. Although foundation models and large language models have accelerated biomedical data interpretation, most current systems remain focused on static pattern recognition rather than prospective simulation of biological futures. Here we propose biomedical world models as a paradigm for AI-driven discovery. These models learn latent representations of molecular, cellular, tissue and clinical states, together with intervention-conditioned dynamics that allow future trajectories to be simulated before actions are taken. We discuss how biomedical world models could function as data engines, environment simulators and scientific planning substrates across applications including virtual cells, organoids, virtual patients and surgical simulation. We outline the data infrastructure, evaluation benchmarks, safety constraints and governance frameworks required. Biomedical world models may provide a foundation for simulation-guided, closed-loop and experimentally actionable biomedical discovery.

Simulating physically realistic garment deformations is an essential task for virtual immersive experience, which is often achieved by physics simulation methods. However, these methods are typically time-consuming, computationally demanding, and require costly hardware, which is not suitable for real-time applications. Recent learning-based methods tried to resolve this problem by training graph neural networks to learn the garment deformation on vertices, which, however, fail to capture the intricate deformation of complex garment meshes with complex topologies. In this paper, we introduce a novel neural deformation field-based method, named UNIC, to animate the garments of an avatar in real time, given the motion sequences. Our key idea is to learn the instance-specific neural deformation field to animate the garment meshes. Such an instance-specific learning scheme does not require UNIC to generalize to new garments but only to new motion sequences, which greatly reduces the difficulty in training and improves the deformation quality. Moreover, neural deformation fields map the 3D points to their deformation offsets, which not only avoids handling topologies of the complex garments but also injects a natural smoothness constraint in the deformation learning. Extensive experiments have been conducted on various kinds of garment meshes to demonstrate the effectiveness and efficiency of UNIC over baseline methods, making it potentially practical and useful in real-world interactive applications like video games.

Creating 3D avatars has become essential in the digital landscape, yet the process remains extremely time-consuming and requires significant expertise. To make this technology accessible to a broader audience, this chapter discusses a novel framework called AvatarCLIP, which enables generation and animation of 3D avatars using text in a zero-shot manner. Unlike traditional software, which requires advanced technical knowledge, AvatarCLIP enables users with no expertise to design 3D avatars with specific shapes and textures and animate them using only descriptions in natural language. The core innovation lies in using the CLIP vision-language model to guide the generation of 3D human models, including their geometry, textures, and animations. Specifically, we begin by generating the 3D human shape through a VAE-based network, which is conditioned on textual prompts. Once the 3D shapes are produced, we apply a volume rendering approach to refine the geometry and textures. For animation, we introduce a novel method that combines motion priors learned in a motion VAE with CLIP-based guidance, enabling reference-driven motion synthesis. Comprehensive experiments demonstrate the robustness and versatility of AvatarCLIP, which can produce previously unseen avatars and animations with impressive zero-shot performance. Our code is publicly available at haps: higithub.com/hongfz 16/AvatarCLIP.

While exocentric video synthesis has achieved great progress, egocentric video generation remains largely underexplored, which requires modeling first-person view content along with camera motion patterns induced by the wearer's body movements. To bridge this gap, we introduce a novel task of joint egocentric video and human motion generation, characterized by two key challenges: 1) Viewpoint Alignment: the camera trajectory in the generated video must accurately align with the head trajectory derived from human motion; 2) Causal Interplay: the synthesized human motion must causally align with the observed visual dynamics across adjacent video frames. To address these challenges, we propose EgoTwin, a joint video-motion generation framework built on the diffusion transformer architecture. Specifically, EgoTwin introduces a head-centric motion representation that anchors the human motion to the head joint and incorporates a cybernetics-inspired interaction mechanism that explicitly captures the causal interplay between video and motion within attention operations. For comprehensive evaluation, we curate a large-scale real-world dataset of synchronized text-video-motion triplets and design novel metrics to assess video-motion consistency. Extensive experiments demonstrate the effectiveness of the EgoTwin framework. Qualitative results are available on our project page: https://egotwin.pages.dev/.

We introduce Ego-R1, a novel framework for reasoning over ultra-long (i.e., in days and weeks) egocentric videos, which leverages a structured Chain-of-Tool-Thought (CoTT) process, orchestrated by an Ego-R1 Agent trained via reinforcement learning (RL). Inspired by human problem-solving strategies, CoTT decomposes complex reasoning into modular steps, with the RL agent invoking specific tools, one per step, to iteratively and collaboratively answer sub-questions tackling such tasks as temporal retrieval and multi-modal understanding. We design a two-stage training paradigm involving supervised finetuning (SFT) of a pretrained language model using CoTT data and RL to enable our agent to dynamically propose step-by-step tools for long-range reasoning. To facilitate training, we construct a dataset called Ego-R1 Data, which consists of Ego-CoTT-25K for SFT and Ego-QA-4.4K for RL. Furthermore, our Ego-R1 agent is evaluated on a newly curated week-long video QA benchmark, Ego-R1 Bench, which contains human-verified QA pairs from hybrid sources. Extensive results demonstrate that the dynamic, tool-augmented chain-of-thought reasoning by our Ego-R1 Agent can effectively tackle the unique challenges of understanding ultra-long egocentric videos, significantly extending the time coverage from few hours to a week.

Video diffusion models have achieved remarkable progress in generating high-quality videos. However, these models struggle to represent the temporal succession of multiple events in real-world videos and lack explicit mechanisms to control when semantic concepts appear, how long they persist, and the order in which multiple events occur. Such control is especially important for movie-grade video synthesis, where coherent storytelling depends on precise timing, duration, and transitions between events. When using a single paragraph-style prompt to describe a sequence of complex events, models often exhibit semantic entanglement, where concepts intended for different moments in the video bleed into one another, resulting in poor text-video alignment. To address these limitations, we propose Prompt Relay, an inference-time, plug-and-play method to enable fine-grained temporal control in multi-event video generation, requiring no architectural modifications and no additional computational overhead. Prompt Relay introduces a penalty into the cross-attention mechanism, so that each temporal segment attends only to its assigned prompt, allowing the model to represent one semantic concept at a time and thereby improving temporal prompt alignment, reducing semantic interference, and enhancing visual quality.

We introduce PerceptionComp, a manually annotated benchmark for complex, long-horizon, perception-centric video reasoning. PerceptionComp is designed so that no single moment is sufficient: answering each question requires multiple temporally separated pieces of visual evidence and compositional constraints under conjunctive and sequential logic, spanning perceptual subtasks such as objects, attributes, relations, locations, actions, and events, and requiring skills including semantic recognition, visual correspondence, temporal reasoning, and spatial reasoning. The benchmark contains 1,114 highly complex questions on 279 videos from diverse domains including city walk tours, indoor villa tours, video games, and extreme outdoor sports, with 100

When an omnimodal large language model accepts a question whose textual premise contradicts what it actually sees or hears, does the failure lie in perception or in action? Recent omnimodal models are positioned as perception-grounded agents that jointly process video, audio, and text, yet a basic form of grounding remains untested: catching a textual claim that conflicts with the model's own sensory input. We introduce IMAVB, a curated 500-clip benchmark of long-form movies with a 2x2 design crossing target modality (vision, audio) and premise condition (standard, misleading), which lets us measure conflict detection separately from ordinary multimodal comprehension. Across eight open-source omnimodal LLMs and Gemini 3.1 Pro, we document a Representation-Action Gap: hidden states reliably encode premise-perception mismatches even when the same models almost never reject the false claim in their outputs. Behaviorally, models fall into two failure modes: under-rejection, in which they answer misleading questions as if the false premise were true; and over-rejection, in which they reject more often but also reject standard questions, sacrificing ordinary comprehension accuracy. The gap is modality-asymmetric (audio grounding underperforms vision) and prompt-resistant across seven variants. As an initial diagnostic intervention, a probe-guided logit adjustment (PGLA) re-injects the encoded mismatch signal into decoding and consistently improves rejection behavior. Together, these results suggest the bottleneck for omnimodal grounding lies in translation, not perception.

Diffusion models arise as a powerful generative tool recently. Despite the great progress, existing diffusion models mainly focus on unimodal control, i.e., the diffusion process is driven by only one modality of condition. To further unleash the users' creativity, it is desirable for the model to be controllable by multiple modalities simultaneously, e.g., generating and editing faces by describing the age (text driven) while drawing the face shape (mask driven). In this chapter, we discuss Collaborative Diffusion, where pre-trained unimodal diffusion models collaborate to achieve multimodal face generation and editing without retraining. The key insight is that diffusion models driven by different modalities are inherently complementary regarding the latent denoising steps, where bilateral connections can be established upon. Specifically, we discuss dynamic diffuser, a meta-network that adaptively hallucinates multimodal denoising steps by predicting the spatial temporal influence functions for each pre-trained unimodal model. Collaborative diffusion not only collaborates generation capabilities from unimodal diffusion models but also integrates multiple unimodal manipulations to perform multimodal editing. Extensive qualitative and quantitative experiments demonstrate the superiority of this framework in both image quality and condition consistency.

The size of vision models has grown exponentially in recent years, particularly with the rise of Vision Transformers. This rapid growth has driven the development of parameter-efficient tuning methods, such as learning adapter layers or low-rank adaptation layers, which enable fine-tuning of a small subset of model parameters while keeping the vast majority of pretrained parameters frozen. However, designing an effective tuning method is not straightforward: it often involves exploring numerous design choices, and each downstream dataset may require custom-tailored solutions. In this chapter, we introduce Neural prOmpt seArcH (NOAH), a novel approach that leverages a neural architecture search algorithm to automatically learn the optimal design of prompt modules for large vision models, tailored specifically for each downstream dataset.