Major Project · Web Development

Kathakali Mudra Recognition System Using Computer Vision Project

A real-time hand gesture recognition system that identifies and displays Kathakali mudras using Computer Vision, MediaPipe hand tracking, and OpenCV through a Django web application.

Difficulty Level

Advanced

Core Modules

5 Modules

Project Overview

The AI-Based Kathakali Mudra Recognition System is an intelligent web application developed to recognize and classify traditional Kathakali hand gestures (Mudras) in real time. The system utilizes Computer Vision techniques, MediaPipe hand landmark detection, and OpenCV image processing to track hand movements from a live webcam feed. Based on the spatial positions and distances between hand landmarks, the application identifies specific Kathakali mudras and displays the recognized gesture instantly. The project helps preserve and promote the cultural heritage of Kathakali by providing an interactive learning and recognition platform for students, artists, and researchers.

Module Breakdown 5 Modules

User Interface Module

Provides web pages for accessing the application, controlling video streaming, and displaying detected Kathakali mudras in real time.

Video Capture Module

Captures live video from the webcam and continuously streams frames to the Django application for processing.

Hand Landmark Detection Module

Uses MediaPipe Hands to detect and track 21 hand landmarks, providing accurate finger and palm position information.

Feature Extraction Module

Calculates distances, finger positions, and hand orientations using detected landmarks to generate gesture features.

Mudra Recognition Module

Identifies Kathakali mudras such as Pathaka, Mushti, Kapithakam, Hamsapaksham, Kartharee Mukham, Mukuram, Mrigashershanam, and others using rule-based gesture classification.