Non Rigid Shape

Non-rigid shape matching aims to establish correspondences between two shapes undergoing non-isometric deformations, a crucial task in computer vision and graphics. Current research heavily utilizes functional maps, often integrated with optimal transport methods or deep learning architectures like encoder-decoder networks and diffusion processes, to achieve robust and accurate matching even in the presence of noise and topological inconsistencies. These advancements improve the accuracy and efficiency of shape analysis across diverse applications, including object tracking, medical image registration, and animation. The field is actively exploring methods to enhance robustness to various types of noise and to handle increasingly complex non-isometric deformations.