player-detection

🎮 🕹️ 🔫 Intelligent opponents (enemy AI) in Unity using a Finite State Machine (FSM) approach.

An implementation of seminal CVPR 2016 paper: "A Hierarchical Deep Temporal Model for Group Activity Recognition."

Real-time football player and ball detection and tracking using YOLO26 Nano and ByteTrack, achieving approximately 28 FPS on video inference.

This project uses a pre-trained YOLOv8 model to detect football players from live camera feeds or video files. It includes extended features such as action recognition, team classification, goal event highlighting, and player heatmap generation — all presented through a user-friendly

Manually annotated football (soccer) object-detection dataset in YOLO format — 612 Full HD images, classes: ball, player, referee.