The field of hand pose estimation and reconstruction is witnessing significant advancements, driven by innovative approaches that address long-standing challenges such as occlusion, intra-class variation, and synthetic-to-real domain gaps. Recent developments have focused on improving the accuracy and robustness of hand pose estimation, particularly in scenarios involving occlusions, interactions with objects, or complex hand postures. Researchers have proposed novel frameworks that integrate foundation models, diffusion-based methods, and transformer architectures to achieve state-of-the-art performance. Noteworthy papers include:

• Aligning Foundation Model Priors and Diffusion-Based Hand Interactions for Occlusion-Resistant Two-Hand Reconstruction, which proposes a novel framework for occlusion-resistant two-hand reconstruction.
• Diff-Palm: Realistic Palmprint Generation with Polynomial Creases and Intra-Class Variation Controllable Diffusion Models, which introduces a polynomial-based palm crease representation and achieves superior recognition performance.
• Analyzing the Synthetic-to-Real Domain Gap in 3D Hand Pose Estimation, which presents a systematic study of the synthetic-to-real gap and proposes a data synthesis pipeline to synthesize high-quality data.
• HORT: Monocular Hand-held Objects Reconstruction with Transformers, which proposes a transformer-based model for efficient reconstruction of dense 3D point clouds of hand-held objects.
• OccRobNet : Occlusion Robust Network for Accurate 3D Interacting Hand-Object Pose Estimation, which proposes an occlusion robust method for estimating 3D hand-object pose from RGB images.