Project Description

Metamorph Arena is an innovative computer vision application that transforms live sports broadcasts by allowing users to modify players’ appearances in real-time. The system would use YOLO object detection and tracking to identify specific players, then applies real-time modifications to their appearance while maintaining visual consistency throughout the game. Users will be able to choose from various modification styles, from making a player bald or changing them to alien while preserving natural movement and game flow.

The project is meant to be exciting by turning watching sports into an interactive experience where viewers can create their own visual narratives. The challenges in include player detection, consistent tracking, and real-time image manipulation while maintaining performance.

Success looks like a smooth, responsive system that can:

• Accurately track specific players across multiple plays
• Apply consistent visual modifications without any large bugs or glitches
• Handle both live streams and uploaded video content with minimal latency

Project Goals

1. Implement robust player detection and tracking using YOLO CNNs that maintains accuracy
2. Develop a real-time modification system that can handle style transformations while preserving natural movement
3. Create an intuitive user interface that supports both live streaming and video file input