Research on Motion Capture Technologies and AI Video Synthesis for Creating Digital Bulgarian Folk Choreographies

Abstract

The article is focused on the following three scientific and applied activities: research into motion capture technologies, analysis of applications for AI video synthesis, and developing a methodology for creating digital choreographies of Bulgarian folk dances. The possibilities for automation through AI and MoCap were assessed, including extraction, adaptation, and synchronisation of choreographic movements with virtual avatars using retargeting technologies. When analysing the leading AI-based applications for markerless MoCap, it was found that they differ in the detail of motion capture, complexity of use, hardware and software requirements. Methods for 3D reconstruction of people from images, significantly facilitates the process of 3D modelling but require specific additional animation settings that improve their work, especially when used to create realistic movements in a digital environment. Automated rigging and animation tools are extremely useful for rapid prototyping, although in some cases additional processing is required to achieve realistic movements. Effective software solutions were identified, a comparative analysis of different AI tools was made and recommendations for the selection of technologies were proposed, based on their purpose, degree of complexity, need for additional software/hardware, price and compatibility with Motion Capture. Based on this research, a step-by-step methodology for creating digital choreographies of Bulgarian folk dances was developed. It integrates MoCap, AI-based tools and automated animation techniques.