Tennis Shot Classification | YOLO + LSTM | Forehand, Backhand and Serve
🎾 Tennis Shot Classification App | Real-time Pose-Based AI Classifier
Welcome to the demo of my AI-powered Tennis Shot Classification App that can recognize tennis shots in real-time using pose estimation and LSTM-based sequence modeling.
🔍 What this app does:
✅ Detects tennis shots — Forehand, Backhand, and Serve
✅ Works in real-time using webcam or pre-recorded videos
✅ Displays predictions with visual overlays (FPS, shot type)
✅ Built with modern ML tools: YOLOv8, LSTM, Streamlit
🧠 Technologies Used:
YOLOv8 Pose Estimation (Ultralytics)
LSTM (Long Short-Term Memory) neural network
PyTorch for model training & inference
OpenCV for frame processing
Streamlit for the user interface
🗂️ Project Highlights:
Real-time frame processing using a rolling 30-frame buffer
Custom-trained LSTM model for action recognition
Easily extendable for other sports or gesture detection
📫 Contact: [email protected]