Skeleton Extraction

Skeleton extraction aims to create simplified, skeletal representations of objects from various data sources, such as images, point clouds, and even 3D models, primarily to capture essential structural information while reducing complexity. Current research emphasizes improving accuracy and robustness, particularly when dealing with low-quality data, occlusions, and complex structures, often employing deep learning models like UNets and employing techniques such as knowledge distillation and attention mechanisms. These advancements have significant implications across diverse fields, including robotics (e.g., for object manipulation), medical image analysis (e.g., vessel segmentation), and computer vision (e.g., action recognition and image restoration).