Automated Cricket Shot Classification using LSTM and Deep Learning Techniques for Enhanced Accuracy and Real-Time Application

A robust framework for real-time image recognition was developed, leveraging key body landmarks identified using Mediapipe and OpenCV libraries. This research focused on creating an advanced system for detecting cricket shots with cutting-edge computer vision and machine learning technologies. Long Short-Term Memory (LSTM) neural networks, a specialized deep learning model, were utilized to accurately classify various cricket strokes. The model was trained on a comprehensive dataset encompassing a wide range of cricket shots and demonstrated a high level of recognition accuracy. The system was rigorously tested with both live camera feeds and pre-recorded video files to ensure real-time functionality. Results indicate that the system performs with notable accuracy and efficiency in real-time scenarios, marking a significant advancement in the application of computer vision for sports analytics. This innovative approach holds substantial potential for practical use in cricket coaching and performance analysis, offering a valuable tool for enhancing training methodologies and analytical precision.