Ethnic dance movement instruction guided by artificial intelligence and 3D convolutional neural networks

Area of Science:

• Computer Science
• Artificial Intelligence
• Dance Studies

Background:

• Ethnic dance instruction traditionally relies on in-person guidance.
• Automated movement recognition can enhance accessibility and standardization.
• Integrating AI offers novel pedagogical tools for cultural arts.

Purpose of the Study:

• To develop and evaluate an AI model for ethnic dance action recognition.
• To apply three-dimensional convolutional neural networks (3D-CNNs) combined with residual networks (ResNet) for dance classification.
• To assess the model's performance across diverse ethnic dance forms.

Main Methods:

• Construction of a dataset comprising six ethnic dances: Miao, Dai, Tibetan, Uygur, Mongolian, and Yi.
• Implementation of a 3D-ResNet model for classifying dance movements.
• Performance evaluation using a self-built dataset and the NTU-RGBD60 database.

Main Results:

• The proposed 3D-ResNet model achieved an accuracy exceeding 95%.
• The model demonstrated strong performance across various ethnic dance types.
• High versatility and adaptability were observed for different cultural dance contexts.

Conclusions:

• The 3D-ResNet model provides effective technical support for ethnic dance instruction.
• The study validates the model's universality and adaptability in cross-cultural dance education.
• This research offers a foundation for AI-driven dance pedagogy and preservation.