Skeleton Representation
Skeleton representation in computer vision and related fields focuses on efficiently encoding the structural and kinematic information of skeletal data, primarily for tasks like action recognition, pose estimation, and shape analysis. Current research emphasizes developing robust and efficient models, often employing graph convolutional networks (GCNs), transformers, and diffusion models to capture spatial and temporal relationships within skeletal data, along with techniques like contrastive learning and knowledge distillation to improve model performance. These advancements have significant implications for various applications, including human-computer interaction, healthcare (e.g., gait analysis), and animation, by enabling more accurate and efficient analysis of human movement and shape.
Papers
USDRL: Unified Skeleton-Based Dense Representation Learning with Multi-Grained Feature Decorrelation
Wanjiang Weng, Hongsong Wang, Junbo He, Lei He, Guosen Xie
Motif Guided Graph Transformer with Combinatorial Skeleton Prototype Learning for Skeleton-Based Person Re-Identification
Haocong Rao, Chunyan Miao