HandVQA: Diagnosing and Improving Fine-Grained Spatial Reasoning about Hands in Vision-Language Models

• URL: http://arxiv.org/abs/2603.26362v1
• Date: Fri, 27 Mar 2026 12:42:26 GMT
• Title: HandVQA: Diagnosing and Improving Fine-Grained Spatial Reasoning about Hands in Vision-Language Models
• Abstract summary: Understanding the fine-grained articulation of human hands is critical in high-stakes settings such as robot-assisted surgery.<n>Current vision-language models (VLMs) struggle with fine-grained spatial reasoning, especially in interpreting complex and articulated hand poses.<n>We introduce HandVQA, a large-scale diagnostic benchmark designed to evaluate VLMs' understanding of detailed hand anatomy through visual question answering.
• Score: 14.981963881449486
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Understanding the fine-grained articulation of human hands is critical in high-stakes settings such as robot-assisted surgery, chip manufacturing, and AR/VR-based human-AI interaction. Despite achieving near-human performance on general vision-language benchmarks, current vision-language models (VLMs) struggle with fine-grained spatial reasoning, especially in interpreting complex and articulated hand poses. We introduce HandVQA, a large-scale diagnostic benchmark designed to evaluate VLMs' understanding of detailed hand anatomy through visual question answering. Built upon high-quality 3D hand datasets (FreiHAND, InterHand2.6M, FPHA), our benchmark includes over 1.6M controlled multiple-choice questions that probe spatial relationships between hand joints, such as angles, distances, and relative positions. We evaluate several state-of-the-art VLMs (LLaVA, DeepSeek and Qwen-VL) in both base and fine-tuned settings, using lightweight fine-tuning via LoRA. Our findings reveal systematic limitations in current models, including hallucinated finger parts, incorrect geometric interpretations, and poor generalization. HandVQA not only exposes these critical reasoning gaps but provides a validated path to improvement. We demonstrate that the 3D-grounded spatial knowledge learned from our benchmark transfers in a zero-shot setting, significantly improving accuracy of model on novel downstream tasks like hand gesture recognition (+10.33%) and hand-object interaction (+2.63%).

Related papers

• Benchmarking Dexterity of Multifingered Robot Hands: A Review and Perspective [2.6272316560877362]
  We survey the state of the art in benchmarking the dexterity of multifingered robot hands.<n>We present the perspective of the U.S. National Science Foundation HAND Engineering Research Center.<n>We introduce a framework consisting of three benchmark levels that correspond to increasing system complexity.
  arXiv  Detail & Related papers  (2026-09-04T16:02:04Z)
• Embodied3DBench: Benchmarking Low-Level Embodied Spatial Intelligence of Vision Language Models [50.6415287154632]
  Embodied3DBench is a robot-centric benchmark targeting low-level spatial intelligence in 3D environments.<n>The benchmark spans 12 subcategories and contains over 21k high-quality question-answer pairs.
  arXiv  Detail & Related papers  (2026-05-27T20:28:56Z)
• HandMCM: Multi-modal Point Cloud-based Correspondence State Space Model for 3D Hand Pose Estimation [60.2305990057581]
  3D hand pose estimation is crucial for many human-computer interaction applications such as augmented reality.<n>HandMCM is a novel method based on the powerful state space model (Mamba)
  arXiv  Detail & Related papers  (2026-02-02T03:25:43Z)
• Monocular 3D Hand Pose Estimation with Implicit Camera Alignment [9.199465050084296]
  Estimating the 3D hand articulation from a single color image is an important problem with applications in Augmented Reality (AR), Virtual Reality (VR), Human-Computer Interaction (HCI)<n>We propose an optimization pipeline for estimating the 3D hand articulation from 2D keypoint input, which includes a keypoint alignment step and a fingertip loss.<n>We evaluate our approach on the EgoDexter and Dexter+Object benchmarks to showcase that it performs competitively with the state-of-the-art.
  arXiv  Detail & Related papers  (2025-06-10T18:45:22Z)
• ViewSpatial-Bench: Evaluating Multi-perspective Spatial Localization in Vision-Language Models [68.46716645478661]
  Vision-language models (VLMs) have demonstrated remarkable capabilities in understanding and reasoning about visual content.<n>Current VLMs excel primarily at egocentric spatial reasoning (from the camera's perspective) but fail to generalize to allocentric viewpoints.<n>We introduce ViewSpatial-Bench, the first comprehensive benchmark designed specifically for multi-viewpoint spatial localization recognition evaluation.
  arXiv  Detail & Related papers  (2025-05-27T17:59:26Z)
• In My Perspective, In My Hands: Accurate Egocentric 2D Hand Pose and Action Recognition [1.4732811715354455]
  Action recognition is essential for egocentric video understanding, allowing automatic and continuous monitoring of Activities of Daily Living (ADLs) without user effort. Existing literature focuses on 3D hand pose input, which requires computationally intensive depth estimation networks or wearing an uncomfortable depth sensor. We introduce two novel approaches for 2D hand pose estimation, namely EffHandNet for single-hand estimation and EffHandEgoNet, tailored for an egocentric perspective.
  arXiv  Detail & Related papers  (2024-04-14T17:33:33Z)
• Benchmarks and Challenges in Pose Estimation for Egocentric Hand Interactions with Objects [89.95728475983263]
  holistic 3Dunderstanding of such interactions from egocentric views is important for tasks in robotics, AR/VR, action recognition and motion generation. We design the HANDS23 challenge based on the AssemblyHands and ARCTIC datasets with carefully designed training and testing splits. Based on the results of the top submitted methods and more recent baselines on the leaderboards, we perform a thorough analysis on 3D hand(-object) reconstruction tasks.
  arXiv  Detail & Related papers  (2024-03-25T05:12:21Z)
• HMP: Hand Motion Priors for Pose and Shape Estimation from Video [52.39020275278984]
  We develop a generative motion prior specific for hands, trained on the AMASS dataset which features diverse and high-quality hand motions. Our integration of a robust motion prior significantly enhances performance, especially in occluded scenarios. We demonstrate our method's efficacy via qualitative and quantitative evaluations on the HO3D and DexYCB datasets.
  arXiv  Detail & Related papers  (2023-12-27T22:35:33Z)
• LG-Hand: Advancing 3D Hand Pose Estimation with Locally and Globally Kinematic Knowledge [0.693939291118954]
  We propose LG-Hand, a powerful method for 3D hand pose estimation. We argue that kinematic information plays an important role, contributing to the performance of 3D hand pose estimation. Our method achieves promising results on the First-Person Hand Action Benchmark dataset.
  arXiv  Detail & Related papers  (2022-11-06T15:26:32Z)
• Measuring Generalisation to Unseen Viewpoints, Articulations, Shapes and Objects for 3D Hand Pose Estimation under Hand-Object Interaction [137.28465645405655]
  HANDS'19 is a challenge to evaluate the abilities of current 3D hand pose estimators (HPEs) to interpolate and extrapolate the poses of a training set. We show that the accuracy of state-of-the-art methods can drop, and that they fail mostly on poses absent from the training set.
  arXiv  Detail & Related papers  (2020-03-30T19:28:13Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.