Wenhang Ge

Chengzhi Lin, Ancong Wu, Junwei Liang et al. · cmu, tencent-ai

Cross-modal retrieval between videos and texts has gained increasing research interest due to the rapid emergence of videos on the web. Generally, a video contains rich instance and event information and the query text only describes a part of the information. Thus, a video can correspond to multiple different text descriptions and queries. We call this phenomenon the ``Video-Text Correspondence Ambiguity'' problem. Current techniques mostly concentrate on mining local or multi-level alignment between contents of a video and text (\textit{e.g.}, object to entity and action to verb). It is difficult for these methods to alleviate the video-text correspondence ambiguity by describing a video using only one single feature, which is required to be matched with multiple different text features at the same time. To address this problem, we propose a Text-Adaptive Multiple Visual Prototype Matching model, which automatically captures multiple prototypes to describe a video by adaptive aggregation of video token features. Given a query text, the similarity is determined by the most similar prototype to find correspondence in the video, which is termed text-adaptive matching. To learn diverse prototypes for representing the rich information in videos, we propose a variance loss to encourage different prototypes to attend to different contents of the video. Our method outperforms state-of-the-art methods on four public video retrieval datasets.

Aditi, Niket Agarwal, Arslan Ali et al.

We introduce Cosmos 3, a family of omnimodal world models designed to jointly process and generate language, image, video, audio, and action sequences within a unified mixture-of-transformers architecture. By supporting highly flexible input-output configurations, Cosmos 3 seamlessly unifies critical modalities for Physical AI -- effectively subsuming vision-language models, video generators, world simulators, and world-action models into a single framework. Our evaluation demonstrates that Cosmos 3 establishes a new state-of-the-art across a diverse suite of understanding and generation tasks, demonstrating omnimodal world models as scalable, general-purpose backbones for embodied agents. Our post-trained Cosmos 3 models were ranked as the best open-source Text-to-Image and Image-to-Video models by Artificial Analysis, and the best policy model by RoboArena at the time the technical report was written. To accelerate open research and deployment in Physical AI, we make our code, model checkpoints, curated synthetic datasets, and evaluation benchmark available under the Linux Foundation's OpenMDW-1.1 https://openmdw.ai/license/1-1/ License at https://github.com/nvidia/cosmos}{github.com/nvidia/cosmos and https://huggingface.co/collections/nvidia/cosmos3 . The project website is available at https://research.nvidia.com/labs/cosmos-lab/cosmos3 .

Wenhang Ge, Tao Hu, Haoyu Zhao et al.

Neural implicit surface learning has shown significant progress in multi-view 3D reconstruction, where an object is represented by multilayer perceptrons that provide continuous implicit surface representation and view-dependent radiance. However, current methods often fail to accurately reconstruct reflective surfaces, leading to severe ambiguity. To overcome this issue, we propose Ref-NeuS, which aims to reduce ambiguity by attenuating the effect of reflective surfaces. Specifically, we utilize an anomaly detector to estimate an explicit reflection score with the guidance of multi-view context to localize reflective surfaces. Afterward, we design a reflection-aware photometric loss that adaptively reduces ambiguity by modeling rendered color as a Gaussian distribution, with the reflection score representing the variance. We show that together with a reflection direction-dependent radiance, our model achieves high-quality surface reconstruction on reflective surfaces and outperforms the state-of-the-arts by a large margin. Besides, our model is also comparable on general surfaces.

Luozhou Wang, Guibao Shen, Wenhang Ge et al.

Text-to-image diffusion models have advanced towards more controllable generation via supporting various additional conditions (e.g.,depth map, bounding box) beyond text. However, these models are learned based on the premise of perfect alignment between the text and extra conditions. If this alignment is not satisfied, the final output could be either dominated by one condition, or ambiguity may arise, failing to meet user expectations. To address this issue, we present a training free approach called Text-Anchored Score Composition (TASC) to further improve the controllability of existing models when provided with partially aligned conditions. The TASC firstly separates conditions based on pair relationships, computing the result individually for each pair. This ensures that each pair no longer has conflicting conditions. Then we propose an attention realignment operation to realign these independently calculated results via a cross-attention mechanism to avoid new conflicts when combining them back. Both qualitative and quantitative results demonstrate the effectiveness of our approach in handling unaligned conditions, which performs favorably against recent methods and more importantly adds flexibility to the controllable image generation process. Our code will be available at: https://github.com/EnVision-Research/Decompose-and-Realign.

Chao Wu, Wenhang Ge, Ancong Wu et al.

To learn camera-view invariant features for person Re-IDentification (Re-ID), the cross-camera image pairs of each person play an important role. However, such cross-view training samples could be unavailable under the ISolated Camera Supervised (ISCS) setting, e.g., a surveillance system deployed across distant scenes. To handle this challenging problem, a new pipeline is introduced by synthesizing the cross-camera samples in the feature space for model training. Specifically, the feature encoder and generator are end-to-end optimized under a novel method, Camera-Conditioned Stable Feature Generation (CCSFG). Its joint learning procedure raises concern on the stability of generative model training. Therefore, a new feature generator, $σ$-Regularized Conditional Variational Autoencoder ($σ$-Reg.~CVAE), is proposed with theoretical and experimental analysis on its robustness. Extensive experiments on two ISCS person Re-ID datasets demonstrate the superiority of our CCSFG to the competitors.

Luozhou Wang, Zhifei Chen, Yihua Du et al.

Large-scale video generation models have demonstrated emergent physical coherence, positioning them as potential world models. However, a gap remains between contemporary "stateless" video architectures and classic state-centric world model theories. This work bridges this gap by proposing a novel taxonomy centered on two pillars: State Construction and Dynamics Modeling. We categorize state construction into implicit paradigms (context management) and explicit paradigms (latent compression), while dynamics modeling is analyzed through knowledge integration and architectural reformulation. Furthermore, we advocate for a transition in evaluation from visual fidelity to functional benchmarks, testing physical persistence and causal reasoning. We conclude by identifying two critical frontiers: enhancing persistence via data-driven memory and compressed fidelity, and advancing causality through latent factor decoupling and reasoning-prior integration. By addressing these challenges, the field can evolve from generating visually plausible videos to building robust, general-purpose world simulators.

Zhifei Chen, Tianshuo Xu, Wenhang Ge et al.

Rendering and inverse rendering are pivotal tasks in both computer vision and graphics. The rendering equation is the core of the two tasks, as an ideal conditional distribution transfer function from intrinsic properties to RGB images. Despite achieving promising results of existing rendering methods, they merely approximate the ideal estimation for a specific scene and come with a high computational cost. Additionally, the inverse conditional distribution transfer is intractable due to the inherent ambiguity. To address these challenges, we propose a data-driven method that jointly models rendering and inverse rendering as two conditional generation tasks within a single diffusion framework. Inspired by UniDiffuser, we utilize two distinct time schedules to model both tasks, and with a tailored dual streaming module, we achieve cross-conditioning of two pre-trained diffusion models. This unified approach, named Uni-Renderer, allows the two processes to facilitate each other through a cycle-consistent constrain, mitigating ambiguity by enforcing consistency between intrinsic properties and rendered images. Combined with a meticulously prepared dataset, our method effectively decomposition of intrinsic properties and demonstrates a strong capability to recognize changes during rendering. We will open-source our training and inference code to the public, fostering further research and development in this area.

Guibao Shen, Yihua Du, Wenhang Ge et al.

The rapid growth of stereoscopic displays, including VR headsets and 3D cinemas, has led to increasing demand for high-quality stereo video content. However, producing 3D videos remains costly and complex, while automatic Monocular-to-Stereo conversion is hindered by the limitations of the multi-stage ``Depth-Warp-Inpaint'' (DWI) pipeline. This paradigm suffers from error propagation, depth ambiguity, and format inconsistency between parallel and converged stereo configurations. To address these challenges, we introduce UniStereo, the first large-scale unified dataset for stereo video conversion, covering both stereo formats to enable fair benchmarking and robust model training. Building upon this dataset, we propose StereoPilot, an efficient feed-forward model that directly synthesizes the target view without relying on explicit depth maps or iterative diffusion sampling. Equipped with a learnable domain switcher and a cycle consistency loss, StereoPilot adapts seamlessly to different stereo formats and achieves improved consistency. Extensive experiments demonstrate that StereoPilot significantly outperforms state-of-the-art methods in both visual fidelity and computational efficiency. Project page: https://hit-perfect.github.io/StereoPilot/.

Wenhang Ge, Guibao Shen, Jiawei Feng et al.

Recent advances in camera-controlled video diffusion models have significantly improved video-camera alignment. However, the camera controllability still remains limited. In this work, we build upon Reward Feedback Learning and aim to further improve camera controllability. However, directly borrowing existing ReFL approaches faces several challenges. First, current reward models lack the capacity to assess video-camera alignment. Second, decoding latent into RGB videos for reward computation introduces substantial computational overhead. Third, 3D geometric information is typically neglected during video decoding. To address these limitations, we introduce an efficient camera-aware 3D decoder that decodes video latent into 3D representations for reward quantization. Specifically, video latent along with the camera pose are decoded into 3D Gaussians. In this process, the camera pose not only acts as input, but also serves as a projection parameter. Misalignment between the video latent and camera pose will cause geometric distortions in the 3D structure, resulting in blurry renderings. Based on this property, we explicitly optimize pixel-level consistency between the rendered novel views and ground-truth ones as reward. To accommodate the stochastic nature, we further introduce a visibility term that selectively supervises only deterministic regions derived via geometric warping. Extensive experiments conducted on RealEstate10K and WorldScore benchmarks demonstrate the effectiveness of our proposed method. Project page: \href{https://a-bigbao.github.io/CamPilot/}{CamPilot Page}.

Harold Haodong Chen, Sirui Chen, Yingjie Xu et al.

The scalability of robotic manipulation is fundamentally bottlenecked by the scarcity of task-aligned physical interaction data. While vision-language models (VLMs) and video generation models (VGMs) hold promise for autonomous data synthesis, they suffer from semantic-spatial misalignment and physical hallucinations, respectively. To bridge this gap, we introduce RoboEvolve, a novel framework that couples a VLM planner and a VGM simulator into a mutually reinforcing co-evolutionary loop. Operating purely on unlabeled seed images, RoboEvolve leverages a cognitive-inspired dual-phase mechanism: (i) daytime exploration fosters physically grounded behavioral discovery through a semantic-controlled multi-granular reward, and (ii) nighttime consolidation mines "near-miss" failures to stabilize policy optimization. Guided by an autonomous progressive curriculum, the system naturally scales from simple atomic actions to complex tasks. Extensive experiments demonstrate that RoboEvolve (I) achieves superior effectiveness, elevating base planners by 30 absolute points and amplifying simulator success by 48% on average; (II) exhibits extreme data efficiency, surpassing fully supervised baselines with merely 500 unlabeled seeds--a 50x reduction; and (III) demonstrates robust continual learning without catastrophic forgetting.

Guibao Shen, Luozhou Wang, Jiantao Lin et al.

Recent advancements in text-to-image generation have been propelled by the development of diffusion models and multi-modality learning. However, since text is typically represented sequentially in these models, it often falls short in providing accurate contextualization and structural control. So the generated images do not consistently align with human expectations, especially in complex scenarios involving multiple objects and relationships. In this paper, we introduce the Scene Graph Adapter(SG-Adapter), leveraging the structured representation of scene graphs to rectify inaccuracies in the original text embeddings. The SG-Adapter's explicit and non-fully connected graph representation greatly improves the fully connected, transformer-based text representations. This enhancement is particularly notable in maintaining precise correspondence in scenarios involving multiple relationships. To address the challenges posed by low-quality annotated datasets like Visual Genome, we have manually curated a highly clean, multi-relational scene graph-image paired dataset MultiRels. Furthermore, we design three metrics derived from GPT-4V to effectively and thoroughly measure the correspondence between images and scene graphs. Both qualitative and quantitative results validate the efficacy of our approach in controlling the correspondence in multiple relationships.

Zhongqi Yang, Wenhang Ge, Yuqi Li et al.

Explorable 3D world generation from a single image or text prompt forms a cornerstone of spatial intelligence. Recent works utilize video model to achieve wide-scope and generalizable 3D world generation. However, existing approaches often suffer from a limited scope in the generated scenes. In this work, we propose Matrix-3D, a framework that utilize panoramic representation for wide-coverage omnidirectional explorable 3D world generation that combines conditional video generation and panoramic 3D reconstruction. We first train a trajectory-guided panoramic video diffusion model that employs scene mesh renders as condition, to enable high-quality and geometrically consistent scene video generation. To lift the panorama scene video to 3D world, we propose two separate methods: (1) a feed-forward large panorama reconstruction model for rapid 3D scene reconstruction and (2) an optimization-based pipeline for accurate and detailed 3D scene reconstruction. To facilitate effective training, we also introduce the Matrix-Pano dataset, the first large-scale synthetic collection comprising 116K high-quality static panoramic video sequences with depth and trajectory annotations. Extensive experiments demonstrate that our proposed framework achieves state-of-the-art performance in panoramic video generation and 3D world generation. See more in https://matrix-3d.github.io.

Tao Hu, Wenhang Ge, Yuyang Zhao et al.

We introduce X-Ray, a novel 3D sequential representation inspired by the penetrability of x-ray scans. X-Ray transforms a 3D object into a series of surface frames at different layers, making it suitable for generating 3D models from images. Our method utilizes ray casting from the camera center to capture geometric and textured details, including depth, normal, and color, across all intersected surfaces. This process efficiently condenses the whole 3D object into a multi-frame video format, motivating the utilize of a network architecture similar to those in video diffusion models. This design ensures an efficient 3D representation by focusing solely on surface information. Also, we propose a two-stage pipeline to generate 3D objects from X-Ray Diffusion Model and Upsampler. We demonstrate the practicality and adaptability of our X-Ray representation by synthesizing the complete visible and hidden surfaces of a 3D object from a single input image. Experimental results reveal the state-of-the-art superiority of our representation in enhancing the accuracy of 3D generation, paving the way for new 3D representation research and practical applications.

Xinli Xu, Wenhang Ge, Jiantao Lin et al.

In this work, we introduce FlexGen, a flexible framework designed to generate controllable and consistent multi-view images, conditioned on a single-view image, or a text prompt, or both. FlexGen tackles the challenges of controllable multi-view synthesis through additional conditioning on 3D-aware text annotations. We utilize the strong reasoning capabilities of GPT-4V to generate 3D-aware text annotations. By analyzing four orthogonal views of an object arranged as tiled multi-view images, GPT-4V can produce text annotations that include 3D-aware information with spatial relationship. By integrating the control signal with proposed adaptive dual-control module, our model can generate multi-view images that correspond to the specified text. FlexGen supports multiple controllable capabilities, allowing users to modify text prompts to generate reasonable and corresponding unseen parts. Additionally, users can influence attributes such as appearance and material properties, including metallic and roughness. Extensive experiments demonstrate that our approach offers enhanced multiple controllability, marking a significant advancement over existing multi-view diffusion models. This work has substantial implications for fields requiring rapid and flexible 3D content creation, including game development, animation, and virtual reality. Project page: https://xxu068.github.io/flexgen.github.io/.

Lutao Jiang, Jiantao Lin, Kanghao Chen et al.

We propose DiMeR, a novel geometry-texture disentangled feed-forward model with 3D supervision for sparse-view mesh reconstruction. Existing methods confront two persistent obstacles: (i) textures can conceal geometric errors, i.e., visually plausible images can be rendered even with wrong geometry, producing multiple ambiguous optimization objectives in geometry-texture mixed solution space for similar objects; and (ii) prevailing mesh extraction methods are redundant, unstable, and lack 3D supervision. To solve these challenges, we rethink the inductive bias for mesh reconstruction. First, we disentangle the unified geometry-texture solution space, where a single input admits multiple feasible solutions, into geometry and texture spaces individually. Specifically, given that normal maps are strictly consistent with geometry and accurately capture surface variations, the normal maps serve as the sole input for geometry prediction in DiMeR, while the texture is estimated from RGB images. Second, we streamline the algorithm of mesh extraction by eliminating modules with low performance/cost ratios and redesigning regularization losses with 3D supervision. Notably, DiMeR still accepts raw RGB images as input by leveraging foundation models for normal prediction. Extensive experiments demonstrate that DiMeR generalises across sparse-view-, single-image-, and text-to-3D tasks, consistently outperforming baselines. On the GSO and OmniObject3D datasets, DiMeR significantly reduces Chamfer Distance by more than 30%.

Xinli Xu, Wenhang Ge, Dicong Qiu et al.

Estimating physical properties for visual data is a crucial task in computer vision, graphics, and robotics, underpinning applications such as augmented reality, physical simulation, and robotic grasping. However, this area remains under-explored due to the inherent ambiguities in physical property estimation. To address these challenges, we introduce GaussianProperty, a training-free framework that assigns physical properties of materials to 3D Gaussians. Specifically, we integrate the segmentation capability of SAM with the recognition capability of GPT-4V(ision) to formulate a global-local physical property reasoning module for 2D images. Then we project the physical properties from multi-view 2D images to 3D Gaussians using a voting strategy. We demonstrate that 3D Gaussians with physical property annotations enable applications in physics-based dynamic simulation and robotic grasping. For physics-based dynamic simulation, we leverage the Material Point Method (MPM) for realistic dynamic simulation. For robot grasping, we develop a grasping force prediction strategy that estimates a safe force range required for object grasping based on the estimated physical properties. Extensive experiments on material segmentation, physics-based dynamic simulation, and robotic grasping validate the effectiveness of our proposed method, highlighting its crucial role in understanding physical properties from visual data. Online demo, code, more cases and annotated datasets are available on \href{https://Gaussian-Property.github.io}{this https URL}.

Wenhang Ge, Jiantao Lin, Guibao Shen et al.

We propose PRM, a novel photometric stereo based large reconstruction model to reconstruct high-quality meshes with fine-grained local details. Unlike previous large reconstruction models that prepare images under fixed and simple lighting as both input and supervision, PRM renders photometric stereo images by varying materials and lighting for the purposes, which not only improves the precise local details by providing rich photometric cues but also increases the model robustness to variations in the appearance of input images. To offer enhanced flexibility of images rendering, we incorporate a real-time physically-based rendering (PBR) method and mesh rasterization for online images rendering. Moreover, in employing an explicit mesh as our 3D representation, PRM ensures the application of differentiable PBR, which supports the utilization of multiple photometric supervisions and better models the specular color for high-quality geometry optimization. Our PRM leverages photometric stereo images to achieve high-quality reconstructions with fine-grained local details, even amidst sophisticated image appearances. Extensive experiments demonstrate that PRM significantly outperforms other models.

Hao Lu, Zhuang Ma, Guangfeng Jiang et al.

Current generative models struggle to synthesize dynamic 4D driving scenes that simultaneously support temporal extrapolation and spatial novel view synthesis (NVS) without per-scene optimization. Bridging generation and novel view synthesis remains a major challenge. We present PhiGenesis, a unified framework for 4D scene generation that extends video generation techniques with geometric and temporal consistency. Given multi-view image sequences and camera parameters, PhiGenesis produces temporally continuous 4D Gaussian splatting representations along target 3D trajectories. In its first stage, PhiGenesis leverages a pre-trained video VAE with a novel range-view adapter to enable feed-forward 4D reconstruction from multi-view images. This architecture supports single-frame or video inputs and outputs complete 4D scenes including geometry, semantics, and motion. In the second stage, PhiGenesis introduces a geometric-guided video diffusion model, using rendered historical 4D scenes as priors to generate future views conditioned on trajectories. To address geometric exposure bias in novel views, we propose Stereo Forcing, a novel conditioning strategy that integrates geometric uncertainty during denoising. This method enhances temporal coherence by dynamically adjusting generative influence based on uncertainty-aware perturbations. Our experimental results demonstrate that our method achieves state-of-the-art performance in both appearance and geometric reconstruction, temporal generation and novel view synthesis (NVS) tasks, while simultaneously delivering competitive performance in downstream evaluations. Homepage is at \href{https://jiangxb98.github.io/PhiGensis}{PhiGensis}.

Wenhang Ge, Chunyan Pan, Ancong Wu et al.

Person re-identification (Re-ID) aims to match person images across non-overlapping camera views. The majority of Re-ID methods focus on small-scale surveillance systems in which each pedestrian is captured in different camera views of adjacent scenes. However, in large-scale surveillance systems that cover larger areas, it is required to track a pedestrian of interest across distant scenes (e.g., a criminal suspect escapes from one city to another). Since most pedestrians appear in limited local areas, it is difficult to collect training data with cross-camera pairs of the same person. In this work, we study intra-camera supervised person re-identification across distant scenes (ICS-DS Re-ID), which uses cross-camera unpaired data with intra-camera identity labels for training. It is challenging as cross-camera paired data plays a crucial role for learning camera-invariant features in most existing Re-ID methods. To learn camera-invariant representation from cross-camera unpaired training data, we propose a cross-camera feature prediction method to mine cross-camera self supervision information from camera-specific feature distribution by transforming fake cross-camera positive feature pairs and minimize the distances of the fake pairs. Furthermore, we automatically localize and extract local-level feature by a transformer. Joint learning of global-level and local-level features forms a global-local cross-camera feature prediction scheme for mining fine-grained cross-camera self supervision information. Finally, cross-camera self supervision and intra-camera supervision are aggregated in a framework. The experiments are conducted in the ICS-DS setting on Market-SCT, Duke-SCT and MSMT17-SCT datasets. The evaluation results demonstrate the superiority of our method, which gains significant improvements of 15.4 Rank-1 and 22.3 mAP on Market-SCT as compared to the second best method.