AI-Powered Bharatanatyam Mudra Identification and Description System: Preserving Heritage Through Technology

Bharatanatyam is a traditional Indian dance form. It involves hand gestures, known as mudras, and facial expressions to represent emotions and convey stories. In this paper describes a new system for classifying and recognizing Bharatanatyam mudras of single-hand and double-hand styles. Advanced computer vision techniques have been used, including the You Only Look Once (YOLO) framework for object detection, MediaPipe for accurate hand keypoint tracking, and MobileNet for efficient mudra classification. High accuracy has been achieved by the system in detecting mudras across various poses and conditions. The curated data set on which the system has been trained contains various lighting scenarios and camera angles. Each identified mudra is linked to its cultural and symbolic meaning, with this information stored in an Excel sheet to enhance the educational value system. The Excel sheet has been intended to serve as a reference for students, researchers, and enthusiasts interested in the symbolic aspects of the dance. The proposed system is expected to contribute significantly to the learning, research, and preservation of Bharatanatyam performance. Its potential application can also be seen in cultural heritage preservation and educational contexts. The educational value of the system has been enriched, and students and researchers have been provided with a detailed understanding of the mudras. The finding shows that the system can successfully identify the mudra and provide a description, helping Bharatanatyam to be preserved and studied.