Integrating electronics and data analysis for movement recognition in Chinese traditional dance performances
Shan Liang · Australian Journal of Electrical & Electronics Engineering · 2025
With the collection of a wide range of ethnic dance data with different styles, how to reasonably organise and support exploratory browsing of ethnic dance data and effectively analyse the relationship between dance movements and performances has become a hot topic in Chinese traditional dance performance research. This research combines data mining, pattern recognition, action interaction and other technologies and proposes extraction and recognition methods for dance actions and performance videos in traditional Chinese dance performances. First, the dance video is divided into frames. Then, optical flow is calculated on each image with good frames to obtain the movement characteristics of the dancers in the image. Then, the image entropy is calculated using the optical flow image to get the entropy sequence. Finally, the sequence is used to extract key frames. Continuous iterations are needed to select the best threshold, obtain high-quality keyframes, and analyse the iterative relationship between dance action and performance. The experimental results prove that the similarity-matching model constructed in this study considers the factors of time order, granularity, interference, and other factors. By comparing the similarities, it can fully analyse the correlation between the dance action and the performance effect, performance time, stage scenery and other elements.