New method improves realistic hand grasp generation for virtual reality
DEAL-Grasp: Decoupled Alignment Representation for Geometry-Aware Dexterous Grasp Generation
RoboticsGraphics
Summary
Making computer-generated hands grasp objects realistically is hard because hand movements combine big motions like moving your whole hand and small finger movements. The authors created DEAL-Grasp, a method that separates these parts to better model hand poses. This method learns to generate grasps by focusing on key hand-object points and finger motions, producing stable and realistic grips without extra trial-and-error during use. It works well on test datasets, making hand-object interactions look natural and physically plausible in virtual environments.
What this means in practice
- •For virtual reality developers: Create more natural and physically accurate hand-object interactions in VR applications without slow optimization steps or unrealistic grips.
- •For robotic system integrators: Generate stable and diverse robotic hand grasps that respect physical contact constraints using a model that separates hand pose components.
Authors
Fuqiang Zhao, Qian Liu
Abstract
Synthesizing realistic articulated hand-object interactions is a fundamental problem in virtual reality, embodied intelligence, and digital human applications. Existing methods for dexterous grasp synthesis typically regress or denoise poses in a joint space that couples global rigid motion with local articulation, which often yields unstable samples and physically implausible contacts. We introduce DEAL-Grasp, built upon the Decoupled Alignment (DEAL) representation, which reformulates grasp synthesis as alignment-space generation: the interaction state comprises task-space geometric anchors and articulation parameters, from which the rigid transform is recovered via closed-form Procrustes alignment while preserving local articulation. On this mixed state, we model grasp generation using heterogeneous-state flow matching with component-wise vector fields, incorporating time-adaptive physical regularization during training. At inference, grasps are synthesized solely by integrating the learned vector field, without test-time optimization or auxiliary physical guidance. Across MultiDex and zero-shot RealDex benchmarks, DEAL-Grasp attains high force-perturbation success rates alongside minimal penetration and high diversity of generated grasps, while substantially reducing native inference latency compared to optimization-heavy baselines. The project page is available at https://wmtlab.github.io/DEAL-Grasp/.