Real-time Classification, Localization and Tracking System (Based on Rhythmic Gymnastics)

Rhythmic gymnastics, which includes both fast movements and turns of various body parts, as well as a large set of complex non-standard poses and their combinations, was chosen to research the problems of developing an effective real- time classification, localization and tracking system. The choice of rhythmic gymnastics as a model task is dictated by the fact that researchers can create a controlled environment for conducting experiments without restriction of movement and the possibility of rapid repetition of poses, which significantly increases the effectiveness of research and accelerates its implementation.