Vision-Language Models for Occupational Physical Exposure Assessment: Estimating External Hand Forces in Manual Material Handling Tasks from RGB Video
Authors: Mohammad Sadra Rajabi, Aanuoluwapo Ojelade, Sunwook Kim, Maury A. Nussbaum
Organizations: Department of Industrial and Systems Engineering, Virginia Tech, Blacksburg VA 24061, USA · St. Jude Children's Research Hospital, Memphis, TN 38105, USA
External hand forces are important inputs to biomechanical analyses of occupational physical exposure and injury risk, yet continuous force measurements during manual material handling (MMH) typically requires instrumented objects or specialized sensing. We evaluated a vision-language model (VLM)-based pipeline that combines task-specific textual cues, visual representations, and known box mass to estimate dynamic, triaxial, bilateral external hand forces from RGB video. Thirty-five healthy young adults performed five MMH tasks involving lifting, carrying, pushing, and pulling with box masses of 6, 9, and 12 kg. The pipeline used text-guided localization of participant and handled-object regions of interest (ROIs), pretrained vision-transformer feature extraction, and transformer-based temporal regression. Performance was evaluated using leave-one-subject-out validation across seven camera-view conditions (three single-view and four multi-view conditions) and four ROI strategies. Overall, root mean square error was ~4.7-5.6 N for the horizontal and mediolateral force components and ~10.6-11.0 N for the vertical component. Including the handled object as a second ROI generally improved force estimation, with some of the largest benefits under single-camera conditions, whereas pixel-level segmentation provided little additional improvement. Multi-camera capture provided the clearest benefit for peak-force estimation, particularly for the vertical component, whereas differences in overall frame-level error among camera configurations were comparatively modest. These findings demonstrate the feasibility of estimating continuous, bilateral, directional hand-force estimates from RGB video and known load mass without requiring sensors on the worker or handled objects as model inputs, supporting the development of more scalable occupational physical exposure and risk assessments.
Manual lifting tasks are a major contributor to work-related musculoskeletal disorders, and effective ergonomic risk assessment is essential for quantifying physical exposure and informing ergonomic interventions. The Revised NIOSH Lifting Equation (RNLE) is a widely used ergonomic risk assessment tool for lifting tasks that relies on six task variables, including horizontal (H) and vertical (V) hand distances; such distances are typically obtained through manual measurement or specialized sensing systems and are difficult to use in real-world environments. We evaluated the feasibility of using innovative vision-language models (VLMs) to non-invasively estimate H and V from RGB video streams. Two multi-stage VLM-based pipelines were developed: a text-guided detection-only pipeline and a detection-plus-segmentation pipeline. Both pipelines used text-guided localization of task-relevant regions of interest, visual feature extraction from those regions, and transformer-based temporal regression to estimate H and V at the start and end of a lift. For a range of lifting tasks, estimation performance was evaluated using leave-one-subject-out validation across the two pipelines and seven camera view conditions. Results varied significantly across pipelines and camera view conditions, with the segmentation-based, multi-view pipeline consistently yielding the smallest errors, achieving mean absolute errors of approximately 6-8 cm when estimating H and 5-8 cm when estimating V. Across pipelines and camera view configurations, pixel-level segmentation reduced estimation error by approximately 20-30% for H and 35-40% for V relative to the detection-only pipeline. These findings support the feasibility of VLM-based pipelines for video-based estimation of RNLE distance parameters.
Mohammad Sadra Rajabi, Aanuoluwapo Ojelade, Sunwook Kim +1
Department of Industrial and Systems Engineering, Virginia Tech, Blacksburg VA 24061, USA · St. Jude Children's Research Hospital, Memphis, TN 38105, USA
Understanding hand-object interaction from egocentric vision is essential for modeling how people physically engage with the surrounding world. Yet reasoning about physically grounded interaction requires estimating the forces acting on hands and objects, beyond localizing contact. We present EgoPHI, the first method that jointly estimates dense contact maps and 3D force distributions on hand and object meshes from a single monocular RGB image and object geometry. To address the lack of scalable ground-truth force annotations, we introduce a physics-based simulation pipeline that augments existing hand-object datasets with dense per-vertex force supervision. EgoPHI then learns dense 3D contact and force on interacting hand and articulated object meshes, extending vision-based force estimation beyond image-space or planar settings. Our evaluation on in-distribution and out-of-distribution benchmarks shows that EgoPHI improves force estimation over existing approaches while generalizing to unseen datasets. To evaluate sim-to-real transfer, we constructed two physical objects that capture dense object contact and force magnitude and used them to record a dataset of interactions from eight participants across diverse touch and grasp types. Our results demonstrate that EgoPHI recovers meaningful 3D contact and force distributions in simulated, out-of-distribution, and real-world settings, advancing egocentric hand-object understanding from contact localization toward physically grounded interaction reasoning.
Andela Ilic, Rachel Schuchert, Yijing Jiang +1
Department of Computer Science, ETH Zürich, Zürich, Switzerland
Vision-Language-Action (VLA) models have emerged as general-purpose robotic manipulation policies. However, in dexterous hand manipulation, contact forces are typically provided as separate states or force-specific representations, making it difficult to explicitly represent the spatial correspondence between force and their corresponding visual locations. In this work, we propose VisForce, which visually grounds the current and desired forces at their corresponding fingertip locations. VisForce renders current and desired visual force cues on the current wrist image and a task-specific goal image, and combines the two representations through goal-conditioned cross-attention to generate force-aware actions. We evaluate VisForce using a real UR10 robot equipped with an RH56F1 dexterous hand through force-conditioned grasping and three multi-stage manipulation tasks. In force-conditioned grasping experiments, VisForce exhibited a consistent grip-force response as the desired force increased, and achieved grasp-and-lift success rates of 70% and 80% for an egg and a toothpaste tube, respectively. It further achieved final success rates of 70%, 55%, and 40% on cup insertion/bottle pouring, tong-assisted bread transfer, and slip-modulated peg-in-hole, respectively. These results show that fingertip-aligned visual force representations can be effectively used for force-aware conditioning in VLA-based dexterous hand manipulation.