Cricket Legends: Exploring VGG-16 for Sports Figure Identification

Abstract

Cricket is among the world's most popular sports, appreciated for its competitive spirit and for the legendary players who have shaped its history. This study presents a deep learning framework that automatically classifies thirty renowned cricket legends. A custom dataset with over 22,817 images was assembled and used to fine-tune a pre-trained VGG-16 convolutional neural network via transfer learning. To ensure accuracy, the model was tested using 5-fold stratified cross-validation, achieving an average accuracy of 94.37% (±0.39%) and consistent results across validation sets. These findings demonstrate the effectiveness of transfer learning for sports image classification and point to valuable applications in digital sports archiving, media analysis, and fan engagement platforms.