Identification of Improper Posture in Female Bharatanatyam Dancers - A Computational Approach

India is a diverse country filled with a rich culture and heritage. Bharatanatyam is one of the traditional classical dance forms of India. To achieve expertise in the art form proper maintenance of different postures for a long time is very important, which imposes immense physical demands on the body during training and performance. About 90% of the dancers experience recurrent pain prevalently in the lower extremities. In the proposed work, a survey was conducted among 30 dancers to get insights about dance-related pain and injuries. A computer vision model was built and used to detect and analyze the most significant posture in Bharatanatyam namely the Aramandi. Posture analysis on the pose was performed by sketching the nodal points. The x, y, and z coordinate values of the nodal points were extracted, and the angle formed between the joints of the knee and ankle was also determined. This extracted data was used to identify deviation from the proper form, thereby determining the proper posture. Also, an attempt has been devised to overcome the problem of key point estimation by using the Media Pipe pose algorithm. The results of the work confirmed that the extracted features were useful in providing accurate classification between the proper and improper Aramandi posture.