Sports Recognition in Videos Using Deep Learning

This paper presents a sports recognition technique that applies a deep learning approach for a video data, addressing the major problems posed by traditional models which find it difficult catching the spatial as well as temporal features. We integrate a Convolutional Neural Network (CNN) with Long Short-Term Memory (LSTM) networks to address these challenges, aiming for improved recognition performance. In first phase, our methodology contains data collection, frame extraction, data cleaning, feature extraction, model selection, and model training. The second phase utilizes the trained model for video retrieval using key frame from given video as input. Validation and performance evaluation assess the model's accuracy in recognizing sports activities. The approach may be useful in applications including sports analytics, video content analysis, and human-computer interaction. The work presented here highlights the transformative potential of deep learning in sports recognition.