An Intelligent Music–Dance Modeling Approach for Assisted Dance Teaching Based on Cycle Consistency

Read full text Summarise with VIRAT
Authors: Lun Lu
Cited by: 0
References: 12

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Abstract Dance teaching relies heavily on the close correspondence between music and body movements to help learners develop rhythmic perception, timing control, and expressive skills. However, existing computational approaches for music-driven dance generation often treat music as a simple control input, which limits their ability to capture musical semantics and reduces their effectiveness in educational applications. To address this limitation, this study proposes an intelligent dance teaching framework based on cycle-consistent music–motion learning. The proposed framework integrates a forward Music-to-Dance generation model with a reverse Dance-to-Music reconstruction model, forming a cycle-consistent learning process that encourages semantic preservation between music and dance movements. Through this closed-loop mechanism, the generated dance motions are constrained to retain essential musical information, enabling a more meaningful alignment between musical structure and body movement. This property is particularly beneficial in dance teaching scenarios, where understanding the relationship between music and motion is critical for learning. Experimental evaluations conducted on two public dance datasets, AIST++ and PopDanceSet, demonstrate that the proposed approach significantly improves motion quality, rhythmic consistency, and music–dance correspondence compared with existing methods. The results suggest that cycle-consistent music–motion learning provides an effective computational foundation for intelligent dance teaching systems, supporting both instructional demonstration and learner engagement.

Fields of Study & Subjects

Human Motion and Animation Music Technology and Sound Studies Generative Adversarial Networks and Image Synthesis

Publication details

Journal / source
Research Square
Publication date
27 March 2026
Language
EN
Document type
Preprint