Virtual Simulation of Dance by Integrating VR Technology and Motion Capture Technology

Abstract

With the rapid development of science and technology, the integration of virtual simulation technology has injected unprecedented vitality and innovation into the field of dance. However, there are still some key problems such as interactivity, multi-person collaboration and real-time performance. In order to further solve these problems, a virtual dance simulation method combining virtual reality technology and motion capture technology was proposed. Firstly, advanced motion capture technology was used to accurately capture dancers' movements and transform them into 3D movements in virtual environment. It then provides users with more engaging and effective learning and performance tools through real-time feedback and interactivity. The results showed that the study's approach achieved the highest accuracy of 81% on the self-built dataset, and achieved 55% and 80% accuracy on the other two datasets. Overall, on large public data sets, graph convolution models perform better than most traditional deep learning models, but not as well as two-flow recursive convolution models. The performance of the graph convolution model is better in the data set specially designed for virtual reality technology interaction. When testing four body movements and their operation instructions in virtual reality technology, the accuracy rate is more than 70%, confirming the effectiveness of body movements for VR interaction. The study shows that this research is expected to promote the further development of virtual dance experience, providing more opportunities and innovation space for dance lovers and professional dancers.