GoalKeeper Action Prediction (Proof of Concept)

An AI system that suggests optimal goalkeeper dive directions during penalty kicks.

What's Implemented

• Object Detection: YOLOv8 tracks ball and players in real-time video
• Decision Engine: Q-learning model processes positions to suggest dive direction (Left/Center/Right)
• Web Interface: Flask app displays predictions on processed video

How It Works

1. Processes input video frame-by-frame
2. YOLOv8 detects:
   • Ball position (bounding box)
   • Kicker position
   • Goalkeeper position
3. Reinforcement learning model analyzes positions and suggests dive direction

Requirements

• Python 3.8+
• OpenCV
• PyTorch
• Ultralytics (YOLOv8)
• Flask (for web interface)