Research on System Architecture Design and Implementation Path for the Integration of Dance Theater Performing Arts and Technology

Abstract

In this paper, we fit the performance information by constructing an LSTM-based fusion model and a fully connected layer to recognize the dance performance movements. A system architecture for fusion of dance performance and technology is designed to manipulate the stage position through system hardware and software. The system functional modules can be divided into four key modules: data processing, machine learning and prediction, visualization and display, and system management and maintenance. In the case of 60/MB-150/MB of action information, the action recognition latency of the proposed system is always around 3.5/ms, and in the case of 150/MB of action information, the action recognition latency of the proposed system for the dance drama performance recognition method is 1.3/ms. The recognition latency curve changes more gently, and does not have a large floating fluctuation. It proves that the action recognition method proposed in this paper has a greater advantage in recognition efficiency and can better meet the needs of the fusion of dance drama performance art and technology, and explores the implementation path from the content side, theoretical side and market side.