Taiji Bafa-Wubu Six-view Synchronized Video Dataset: Rule-driven Key Event Modeling and Unified Segment-Frame Encoding Framework for Action Quality Assessment

This dataset contains synchronized six-view RGB videos of Taiji Bafa-Wubu, collected for Action Quality Assessment (AQA). The data includes 41 participants performing a complete Taiji routine, with annotations for action segments and scoring keyframes. This is the first part of the dataset, covering samples 1–15.