Yanli Li ¹ Congyi Wang ¹ Huan Wang ^2*

• ¹ Fuzhou Medical College of Nanchang University, Fuzhou, China
• ² Industrial Technology Research Center,Guangdong Institute of Scientific & Technical Information, Guangzhou, China

With an enormous number of hand images generated over time, leveraging unlabeled images for pose estimation is an emerging yet challenging topic. While some semi-supervised and self-supervised methods have emerged, they are constrained by their reliance on high-quality keypoint detection models or complicated network architectures. We propose a novel selfsupervised pre-training strategy for 3D hand mesh regression. Our approach integrates a multigranularity strategy with pseudo-keypoint alignment in a teacher-student framework, employing self-distillation and masked image modeling for comprehensive representation learning. We pair this with a robust pose estimation baseline, combining a standard Vision Transformer backbone with a Pyramidal Mesh Alignment Feedback head. Extensive experiments demonstrate HandMIM's competitive performance across diverse datasets, notably achieving an 8.00mmProcrustes Alignment Vertex-Point-Error on the challenging HO3Dv2 test set, which features severe hand occlusions, surpassing many specially optimized architectures.