Cricket shot classification using Deep Learning

Cricket is a popular sport that involves various batting techniques and shot types. Accurate and real-time detection of cricket shots can provide priceless visual images for coaches, players, and analysts. This study proposes an artificial intelligence-based approach for cricket shot detection aimed at automatically recognizing and classifying different shots played by batsmen. Images are extracted and pre-processed to enhance their quality. The pre-processed images extract key features like bat trajectory, player movement, and batsman position. A training dataset is used to enable shot classification, consisting of labeled samples of different cricket shots. Neural network models, such as a convolutional neural network, VGG16, VGG19, are trained using an image set to learn the patterns and characteristics of each shot type. The model is then employed to classify shots in real time for predicting the corresponding shot type based on the selected features. The proposed approach is evaluated on a diverse dataset of cricket match images, encompassing a broad extent of shots played by professional batsmen. Execution metrics such as accuracy, precision, recall, and F1-score are used to assess the effect of the shot detection system. A VGG16 model in cricket shot prediction of the dataset yielded an impressive accuracy of 95.81%. The solution establishes the efficiency of the planned approach in accurately detecting and classifying cricket shots in real-time. The developed system has the potential to enhance cricket coaching, player performance analysis, and broadcast coverage by providing valuable insights into the shot selection and execution of batters.