Preprint
Aug 2026
MASQ: Mask-Aware Spatiotemporal Quantization for Unsupervised Skeleton Action Segmentation
A novel Mask-aware Action Spatiotemporal Quantization framework that decouples the conflicting tasks of spatial feature inference and temporal smoothing, and establishes a comprehensive and substantial leading advantage in the Mean over Frames accuracy.
Xingchen Qin, Lin-Xiang Peng, You-Bao Ye et al.
· 0 citations