Garment refitting, the task of adapting a garment from a source to a target avatar, must preserve the original design features and fine-scale wrinkles, a challenge exacerbated by significant shape variations and varying poses without registration to a shared canonical pose. Existing methods struggle to balance robustness, efficiency, and fidelity of detail: physics-based simulation is costly, data-driven approaches lack generalizability, and geometry optimization in the full vertex space is often ill-conditioned and prone to local minima with unsatisfactory quality. We identify that a fundamental limitation lies in the representation: deforming garments directly in global coordinates couples vertices non-locally, creating a complex and poorly-structured optimization landscape. Therefore, we introduce LoBoFit, a robust refitting method built upon a novel Local Bone Mapping Blending (LoBoMap Blending) representation. Instead of manipulating global vertex positions, LoBoMap Blending expresses garment geometry as a linear blend of its mappings into local bone coordinate frames. This representation is highly expressive and flexible: local bone mappings yield a pose-robust initialization and a well-conditioned parameterization, while blending weights smooth the optimization landscape and broaden the space of plausible solutions for stable convergence with fine-scale detail preservation. The subsequent refinement efficiently resolves collisions and preserves details by optimizing localized residuals, effectively decomposing the complex global deformation into manageable subproblems. Our experiments demonstrate that LoBoFit reliably refits high-resolution, single- and multi-layer garments across avatars with large shape and topological differences, while faithfully preserving intricate wrinkles and the intended fit style, outperforming state-of-the-art methods in robustness and output quality.

Existing depth estimation methods are fundamentally limited to predicting depth on discrete image grids. Such representations restrict their scalability to arbitrary output resolutions and hinder the geometric detail recovery. This paper introduces InfiniDepth, which represents depth as neural implicit fields. Through a simple yet effective local implicit decoder, we can query depth at continuous 2D coordinates, enabling arbitrary-resolution and fine-grained depth estimation. To better assess our method's capabilities, we curate a high-quality 4K synthetic benchmark from five different games, spanning diverse scenes with rich geometric and appearance details. Extensive experiments demonstrate that InfiniDepth achieves state-of-the-art performance on both synthetic and real-world benchmarks across relative and metric depth estimation tasks, particularly excelling in fine-detail regions. It also benefits the task of novel view synthesis under large viewpoint shifts, producing high-quality results with fewer holes and artifacts. Code and data will be made publicly available.

As a critical,rapidly evolving domain within artificial intelligence(AI),embodied AI represents the conver-gence of computer vision,natural language processing,and robotics,and aims to create intelligent agents capable of per-ceiving,reasoning,and acting within the physical world.However,despite the transformative success of large language models(LLMs)in the digital realm,embodied AI faces unprecedented,multifaceted challenges that hinder the direct rep-lication of the"large-scale pre-training plus scaling law"paradigm.These challenges include extreme data heterogeneity across different robot morphologies,strong physical constraints that demand safety and precision,and the prohibitively expensive interaction costs associated with collecting real-world robotic data.Consequently,simply scaling up model parameters without addressing these domain-specific hurdles has proven insufficient for achieving general-purpose robotic intelligence.This paper comprehensively reviews the frontier technical evolution of embodied AI and offers a systematic analysis across four critical dimensions,namely,data,models,systems,and evaluation,to chart a path toward more robust and generalized embodied agents.In terms of data,this paper proposes a"Data Pyramid"structure designed to maximize data efficiency and transferability.This hierarchical framework advocates for the foundational use of massive,low-cost simulation data and Internet-scale video datasets at the bottom layer to build broad physical commonsense and visual representations;the utilization of human interaction data(such as ego-centric videos and teleoperation logs)in the middle layer to facilitate behavioral mapping and intent understanding;and the strategic application of a small,high-quality amount of real-world robot data at the top layer for fine-tuning and final skill deployment,an approach bridging the reality gap.Regarding models,the paper critically discusses the current state of mainstream vision-language-action models and highlights that while they excel at semantic understanding,they encounter significant sealing bottlenecks in continuous control and fine-grained manipulation.To overcome this,this paper identifies"World Models"as a pivotal new direction for embodied pretraining.By learning to simulate environmental dynamics,predict future states,and understand causal relationships without explicit supervision,world models promise to endow agents with deeper physical intuition and supe-rior generalization capabilities in unseen environments.In terms of systems,this paper observes a paradigm shift,where the architecture is evolving from monolithic,single end-to-end models toward an operating system-like"Hierarchical Archi-tecture."This evolution achieves the necessary decoupling of high-level semantic planning——powered by the reasoning capabilities of LLMs——and low-level motion control,which ensures precise execution and hardware compliance.This modular approach not only improves system robustness but also facilitates easier debugging and component upgrades.Finally,the paper examines the critical issues within current evaluation systems and specifically focuses on the challenges of authenticity in simulation benchmarks and the lack of reproducibility in real-world experiments.The field suffers from fragmented metrics that fail to capture the complexity of open-world interaction.In conclusion,this paper provides a forward-looking perspective on the inevitable integration of locomotion and manipulation——moving beyond stationary arms to mobile manipulators——and anticipates the arrival of the"ImageNet moment"for embodied AI,where standardized datasets and benchmarks will catalyze a Cambrian explosion of robotic capabilities,a development ultimately bridging the gap between digital intelligence and physical reality.

Depth estimation is a fundamental task in 3D computer vision, crucial for applications such as 3D reconstruction, free-viewpoint rendering, robotics, autonomous driving, and AR/VR technologies. Traditional methods relying on hardware sensors like LiDAR are often limited by their high costs, low resolution, and sensitivity to the environment, limiting their applicability to real-world scenarios. Recent advances in vision-based methods offer a promising alternative, yet they face challenges in generalization and stability due to either the low capacity of model architectures or reliance on domain-specific and small-scale datasets. The emergence of scaling laws and foundation models in other domains has inspired the development of “depth foundation models”: deep neural networks trained on large datasets with strong zeroshot generalization capabilities. This paper surveys the evolution of deep learning architectures and paradigms for depth estimation across monocular, stereo, multiview, and monocular video settings. We explore the potential of these models to address existing challenges and we also provide a comprehensive overview of large-scale datasets that can facilitate their development. By identifying key architectures and training strategies, we aim to highlight the path towards robust depth foundation models, offering insights for future research and applications.

We present Skeletal-Anchored Dual Harmonics (SADH), a novel 3D shape representation that tightly couples local surface geometry with internal meso-skeletal organization. SADH represents a shape as a collection of compact surface patches rooted on internal anchors optimized directly inside the object volume. Each patch is parameterized using a dual-channel spherical harmonic (SH) formulation, where one channel models local radial geometry while the other defines adaptive patch support through a generalized viewing cone. Unlike isotropic primitives such as medial spheres or Gaussian kernels, SH patches directly encode anisotropic local surface geometry together with adaptive spatial support, enabling compact representation of detailed and directionally varying surface regions. Starting from unorganized point clouds, SADH jointly optimizes surface geometry, anchor locations, patch orientations, and structural connectivity through a staged optimization process that progressively forms a coherent meso-skeletal structure. A geodesic anchor graph further preserves structural relationships between neighboring patches. Experiments on complex 3D shapes demonstrate that SADH achieves accurate surface reconstruction together with compact and coherent skeletal organization across a wide range of geometries.

Accurate digital twin reconstruction of articulated objects is essential for graphics-driven simulation, visualization, and embodied applications. However, existing 3D Gaussian Splatting (3DGS) approaches primarily target static scenes and fail to capture the structural dynamics of multi-joint objects, leading to geometric inconsistencies and limited realism. We present a structure-aware reconstruction framework that attaches articulated Gaussian primitives to the kinematic hierarchy, enabling dynamic and editable digital twin reconstruction from multi-view observations via differentiable rendering. By explicitly modeling link-level geometry and motion coupling, the proposed method produces temporally consistent reconstructions that generalize to novel viewpoints and unseen joint configurations. Extensive experiments demonstrate improved visual fidelity, geometric accuracy, and cross-view consistency on articulated robotic systems. Furthermore, the reconstructed digital twins facilitate scalable simulation and enhance sim-to-real policy transfer, highlighting the importance of dynamic Gaussian-based modeling for high-fidelity digital twin applications.

Large-scale floorplan generation is critical for virtual space planning and architectural simulation. Although existing methods have shown success in generating small-scale floorplans with simple room shapes, they struggle to handle the complex room connections and irregular room shapes that arise in large-scale floorplans. In this paper, we propose CG-Floor, a centroid-guided hierarchical framework that explicitly decouples topology and geometry to address these issues. We first introduce the size-aware semantic centroid heatmap, derived from predicted room centroids, which provides a structured representation to precisely guide the effective generation of a coarse-to-fine floorplan generator while ensuring semantic alignment. Additionally, we train a vector quantized codebook of floorplans with complex room shapes to capture the diversity of room shapes and employ a latent diffusion transformer to generate large-scale floorplans featuring non-Manhattan room shapes. CG-Floor achieves state-of-the-art performance on the large-scale MSD dataset, and supports 3D floorplan conversion and editing, demonstrating the practicality of our approach.The code will be publicly available for research purposes.

The household rearrangement task involves spotting misplaced objects in a scene and accommodate them with proper places. It depends both on common-sense knowledge on the objective side and human user preference on the subjective side. In achieving such a task, we propose to mine object functionality with user preference alignment directly from the scene itself, without relying on human intervention. To do so, we work with scene graph representation and propose LLM-enhanced scene graph learning which transforms the input scene graph into an Affordance Enhanced Graph (AEG) with information-enriched nodes and newly discovered edges (relations). In AEG, the nodes corresponding to the receptacle objects are augmented with context-induced affordance which encodes what kind of carriable objects can be placed on it. New edges are discovered with newly discovered non-local relations. With AEG, we perform task planning for scene rearrangement by detecting misplaced carriables and determining a proper placement for each of them. We implement an end-to-end robot system for autonomous household rearrangement in unseen environments and test our method by implementing a tiding robot in both simulated environments and real-world scenarios, and perform evaluation on a new benchmark we build. Extensive evaluations demonstrate that our method achieves state-of-the-art performance in misplacement detection and rearrangement planning.

Achieving real-time physics-based animation that generalizes across diverse 3D shapes and discretizations remains a fundamental challenge. We introduce PhysSkin, a physics-informed framework that addresses this challenge. In the spirit of Linear Blend Skinning, we learn continuous skinning fields as basis functions lifting motion subspace coordinates to full-space deformation, with subspace defined by handle transformations. To generate mesh-free, discretization-agnostic, and physically consistent skinning fields that generalize well across diverse 3D shapes, PhysSkin employs a new neural skinning fields autoencoder which consists of a transformer-based encoder and a cross-attention decoder. Furthermore, we also develop a novel physics-informed self-supervised learning strategy that incorporates on-the-fly skinning-field normalization and conflict-aware gradient correction, enabling effective balancing of energy minimization, spatial smoothness, and orthogonality constraints. PhysSkin shows outstanding performance on generalizable neural skinning and enables real-time physics-based animation.

Single-image feed-forward 3D Gaussian Splatting (3DGS) aims to directly generate a renderable 3D scene representation from one input image, avoiding the cost of multi-view capture and per-scene optimization. However, existing methods are often constrained by a pixel-aligned representation, where Gaussians are predicted from fixed image-grid locations. Such pixel-aligned primitives can produce promising nearby-view renderings, but they remain weakly coupled to underlying scene surfaces and struggle to preserve coherent structures under large viewpoint shifts. We present InfiniSplat, a feed-forward single-image 3DGS framework that moves from a pixel-aligned representation toward a surface-aligned representation. InfiniSplat constructs this representation by first using geometry-guided sampling to place 2D supports according to depth-induced local surface structure, and then applying a query-conditioned implicit decoder to predict Gaussian attributes from the image features queried at these supports. By grounding support locations in geometry while decoupling Gaussian prediction from fixed pixel centers, InfiniSplat produces Gaussian layouts that better follow scene surfaces and reduce scattered primitives caused by grid discretization. Across multiple cross-dataset NVS evaluations, InfiniSplat achieves state-of-the-art performance compared with single-image feed-forward baselines, and demonstrates zero-shot generalization from Hypersim indoor synthetic training to complex open-world scenes. Project page: https://zju3dv.github.io/InfiniSplat.

Real-time navigation in cluttered and dynamic environments requires collision-free and dynamically feasible motion under limited perception. However, feasible navigation behaviors are inherently multimodal because multiple paths may exist around obstacles. In this paper, we formulate navigation as learning a transferable goal-conditioned stochastic path prior that models a reusable distribution over goal-aligned geometry consistent local paths conditioned on local observations. This formulation enables structured sampling of navigation candidates, allowing multiple feasible paths to be explored through sampling without relying on robot-specific motion constraints. To this end, we introduce a goal-aligned canonical state representation that removes in-plane rotational ambiguity and normalizes local geometry with respect to the goal, enabling rotation-invariant path distribution learning. We further develop a structured prior learning framework that parameterizes local paths using a geometry-aware polar action manifold and incorporates risk sensitive utility shaping with multi-goal distributional rollouts for stable and safety-aware planning. Extensive experiments in dense static environments and dynamic pedestrian scenarios demonstrate that the proposed method achieves consistently high success rates with competitive efficiency while enabling cross platform transfer of a single path prior learned on differential drive robots to quadruped platforms without retraining.

The Flexible Job Shop Scheduling Problem (FJSP) originates from real production lines, while some practical constraints are often ignored or idealized in current FJSP studies, among which the limited buffer problem has a particular impact on production efficiency. To this end, we study an extended problem that is closer to practical scenarios—the Flexible Job Shop Scheduling Problem with Limited Buffers and Material Kitting. In recent years, deep reinforcement learning (DRL) has demonstrated considerable potential in scheduling tasks. However, its capacity for state modeling remains limited when handling complex dependencies and long-term constraints. To address this, we leverage a heterogeneous graph network within the DRL framework to model the global state. By constructing efficient message passing among machines, operations, and buffers, the network focuses on avoiding decisions that may cause frequent pallet changes during long-sequence scheduling, thereby helping improve buffer utilization and overall decision quality. Experimental results on both synthetic and real production line datasets show that the proposed method outperforms traditional heuristics and advanced DRL methods in terms of makespan and pallet changes, and also achieves a good balance between solution quality and computational cost. Furthermore, a supplementary video is provided to showcase a simulation system that effectively visualizes the progression of the production line.

In modern industrial production, multiple robots often collaborate to complete complex manufacturing tasks. Large language models (LLMs), with their strong reasoning capabilities, have shown potential in coordinating robots for simple household and manipulation tasks. However, in industrial scenarios, stricter sequential constraints and more complex dependencies within tasks present new challenges for LLMs. To address this, we propose IMR-LLM, a novel LLM-driven Industrial Multi-Robot task planning and program generation framework. Specifically, we utilize LLMs to assist in constructing disjunctive graphs and employ deterministic solving methods to obtain a feasible and efficient high-level task plan. Based on this, we use a process tree to guide LLMs to generate executable low-level programs. Additionally, we create IMR-Bench, a challenging benchmark that encompasses multi-robot industrial tasks across three levels of complexity. Experimental results indicate that our method significantly surpasses existing methods across all evaluation metrics.

World foundation models (WFMs) are powerful simulators, yet they predominantly operate in a single-view setting and lack the multi-view 3D consistency required for robotic manipulation. While robotic systems rely on multiple cameras (egocentric, eye-to-hand, and wrist-mounted) for policy learning, current multi-view world models simply concatenate view tokens without explicit geometric reasoning. This causes cross-view object drift, depth inconsistency, and texture misalignment. We trace these failures to two deficiencies: the absence of an explicit inter-view communication mechanism and the lack of a 3D geometric prior. We argue that resolving both simultaneously is necessary and sufficient. To address this, we present PAIWorld, a framework that augments diffusion-transformer world models via three core components: (1) Geometry-Aware Cross-View Attention blocks that establish an explicit pathway across views, (2) Geometric Rotary Position Embedding that encodes camera ray directions and extrinsic poses into the attention mechanism, and (3) Latent 3D-REPA, which distills 3D-aware features from frozen 3D foundation models to ensure 3D consistency. Built upon a DiT-based world foundation model, PAIWorld achieves state-of-the-art multi-view 3D consistency on robotic manipulation benchmarks, ranking 1st on the WorldArena leaderboard and 2nd on the AgiBot-Challenge2026 leaderboard, while enabling downstream applications such as model-based planning, world action models, and multi-view policy post-training.

Training embodied AI agents depends critically on the visual fidelity of simulation environments and the ability to model dynamic humans. Current simulators rely on mesh-based rasterization with limited visual realism, and their support for dynamic human avatars, where available, is constrained to mesh representations, hindering agent generalization to human-populated real-world scenarios. We present Habitat-GS, a navigation-centric embodied AI simulator extended from Habitat-Sim that integrates 3D Gaussian Splatting scene rendering and drivable gaussian avatars while maintaining full compatibility with the Habitat ecosystem. Our system implements a 3DGS renderer for real-time photorealistic rendering and supports scalable 3DGS asset import from diverse sources. For dynamic human modeling, we introduce a gaussian avatar module that enables each avatar to simultaneously serve as a photorealistic visual entity and an effective navigation obstacle, allowing agents to learn human-aware behaviors in realistic settings. Experiments on point-goal navigation demonstrate that agents trained on 3DGS scenes achieve stronger cross-domain generalization, with mixed-domain training being the most effective strategy. Evaluations on avatar-aware navigation further confirm that gaussian avatars enable effective human-aware navigation. Finally, performance benchmarks validate the system's scalability across varying scene complexity and avatar counts.

We present a novel method for recovering world-grounded human motion from monocular video. The main challenge lies in the ambiguity of defining the world coordinate system, which varies between sequences. Previous approaches attempt to alleviate this issue by predicting relative motion in an auto-regressive manner, but are prone to accumulating errors. Instead, we propose estimating human poses in a novel Gravity-View (GV) coordinate system, which is defined by the world gravity and the camera view direction. The proposed GV system is naturally gravity-aligned and uniquely defined for each video frame, largely reducing the ambiguity of learning image-pose mapping. The estimated poses can be transformed back to the world coordinate system using camera rotations, forming a global motion sequence. Additionally, the per-frame estimation avoids error accumulation in the auto-regressive methods. Experiments on in-the-wild benchmarks demonstrate that our method recovers more realistic motion in both the camera space and world-grounded settings, outperforming state-of-the-art methods in both accuracy and speed. We further introduce a Stationary Label Predictor (SLP) to estimate the motion stationary states of hands and feet. By producing sharper decision boundaries, the module mitigates unnatural artifacts such as foot-sliding and shows improved performance under fast-motion inputs. The code is available at https://zju3dv.github.io/gvhmr.

Modeling wind-driven object dynamics from video observations is highly challenging due to the invisibility and spatio–temporal variability of wind, as well as the complex deformations of objects. We present DiffWind, a physics-informed differentiable framework that unifies wind–object interaction modeling, video-based reconstruction, and forward simulation. Specifically, we represent wind as a grid-based physical field and objects as particle systems derived from 3D Gaussian Splatting, with their interaction modeled by the Material Point Method (MPM). To recover wind-driven object dynamics, we introduce a reconstruction framework that jointly optimizes the spatio–temporal wind force field and object motion through differentiable rendering and simulation. To ensure physical validity, we incorporate the Lattice Boltzmann Method (LBM) as a physics-informed constraint, enforcing compliance with fluid dynamics laws. Beyond reconstruction, our method naturally supports forward simulation under novel wind conditions and enable new applications such as wind retargeting. We further introduce WD-Objects, a dataset of synthetic and real-world wind-driven scenes. Extensive experiments demonstrate that our method significantly outperforms prior dynamic scene modeling approaches in both reconstruction accuracy and simulation fidelity, opening a new avenue for video-based wind–object interaction modeling.

World Foundation Models (WFMs) offer remarkable visual dynamics simulation capabilities, yet their application to precise robotic control remains limited by the gap between generative realism and control-oriented precision. While existing approaches use WFMs as synthetic data generators, they suffer from high computational costs and underutilization of pre-trained VLA policies. We introduce AdaPower (Adapt and Empower), a lightweight adaptation framework that transforms general-purpose WFMs into specialist world models through two novel components: Temporal-Spatial Test-Time Training (TS-TTT) for inference-time adaptation and Memory Persistence (MP) for long-horizon consistency. Integrated within a Model Predictive Control framework, our adapted world model empowers pre-trained VLAs, achieving over 41% improvement in task success rates on LIBERO benchmarks without policy retraining, while preserving computational efficiency and generalist capabilities.

Task-oriented dexterous grasping holds broad application prospects in robotic manipulation and human-object interaction. However, most existing methods still struggle to generalize across diverse objects and task instructions, as they heavily rely on costly labeled data to ensure task-specific semantic alignment. In this study, we propose ZeroDexGrasp, a zero-shot task-oriented dexterous grasp synthesis framework integrating Multimodal Large Language Models with grasp refinement to generate human-like grasp poses that are well aligned with specific task objectives and object affordances. Specifically, ZeroDexGrasp employs prompt-based multi-stage semantic reasoning to infer initial grasp configurations and object contact information from task and object semantics, then exploits contact-guided grasp optimization to refine these poses for physical feasibility and task alignment. Experimental results demonstrate that ZeroDexGrasp enables high-quality zero-shot dexterous grasping on diverse unseen object categories and complex task requirements, advancing toward more generalizable and intelligent robotic grasping.

Data-driven models of robot motion constructed using principles from Geometric Mechanics have been shown to produce useful predictions of robot motion for a variety of robots. For robots with a useful number of DoF, these geometric mechanics models can only be constructed in the neighborhood of a gait. Here we show how Gaussian Mixture Models (GMM) can be used as a form of manifold learning that learns the structure of the Geometric Mechanics "motility map" and demonstrate: [i] a sizable improvement in prediction quality when compared to the previously published methods; [ii] a method that can be applied to any motion dataset and not only periodic gait data; [iii] a way to pre-process the data-set to facilitate extrapolation in places where the motility map is known to be linear. Our results can be applied anywhere a data-driven geometric motion model might be useful.