Dance movement style recognition and automatic classification technology based on artificial intelligence

This paper focuses on the research of dance action style recognition and automatic classification system based on artificial intelligence (AI). Through extensive collection of multi-style dance action data from professional dancers' action capture and public dance video library, the hierarchical architecture is used to design the system, covering data collection, feature extraction and classification model construction. In feature extraction, both spatio-temporal and semantic features are taken into account, the classification model adopts the combination of CNN and RNN. Five typical dance style data were selected in the experiment, and the model was trained and evaluated after being divided into training set and testing set. The results show that the overall performance of the model is good, with an accuracy rate of 0.85, and the recognition accuracy of different dance styles is different, which is better than the common classification models. The research shows that the system can effectively recognize and automatically classify dance movements, which provides strong support for the intelligent development of dance field.