Military Object Detection with YOLOv8

Real-time multi-class object detection system for military imagery using YOLOv8s.

This project implements a real-time multi-class object detection system for military imagery using YOLOv8s.

Key aspects:

  • 12-class military object detection dataset (personnel, vehicles, equipment, aerial/naval)
  • Training, evaluation, and inference notebooks for reproducible experiments
  • GPU-ready pipeline with ~5ms inference time per image (on supported hardware)

See metrics, dataset details, and training configuration in the project repository.

View the repository