Music feature analysis for synchronization with dance animation

This paper presents an approach for music feature analysis with an application in dance animation, in which a dancing avatar dances according to the rhythm and energy of the music. Extracted music features, such as drum beats and meters, are used for synchronization with motion sequences from a database of dance animation files. In addition, music segmentation with music intensity labels are produced for the selection of motion sequences of proper intensity levels. Experiments on club dancing music, including hip-hop and R&B, were conducted, and the results validate the performance of the proposed techniques.