DAASH-IUTCricShot

This repository contains the code and resources for a Machine Learning model designed to automatically detect cricket shot types from video footage.

About The Project

Welcome to our project for detecting cricket shot types from video footage using machine learning! This project aims to classify cricket shots in real time, leveraging advanced computer vision and deep learning techniques.

Key Features

• Advanced Object Detection with YOLO v9: Utilizes a custom-trained YOLO v9 model to locate and crop the batsman from the video frames.

• LRCN for Shot Type Classification: Employs a Long-term Recurrent Convolutional Network (LRCN) to analyze the cropped video sequence and classify the shot type. Initially, we are detecting 4 types of shots.

  • PULL

    

  • DEFENSE

    

  • DRIVE

    

  • FLICK

    

Methodology

• Object Detection:
  • YOLO v9 is used to detect the batsman in each frame.
  • The detected region around the batsman is cropped for further analysis.
• Shot Classification:
  • The cropped video sequence is processed by an LRCN model.
  • The LRCN model classifies the type of cricket shot based on the batsman’s movements.

License

Distributed under the MIT License. See LICENSE for more information.