Minecraft Mob Detection
Real-time object detection of 15 Minecraft mobs from live screen capture using YOLOv8s fine-tuned on a custom merged dataset.
About this project
Built to satisfy the Neural Network course's CNN requirement, this fine-tunes YOLOv8s (CSPDarknet backbone) on a 15-class Minecraft mob dataset — 8 classes from a Roboflow base plus 7 extras (bee, goat, polar_bear, skeleton, slime, spider, villager) merged via a custom remapping script. The detector runs against a live `mss` screen capture and displays bounding boxes in a tkinter window with FPS counter, deliberately avoiding `cv2.imshow` because OpenCV's GUI module conflicts with the headless wheel many Python environments ship. After 100 epochs at imgsz=640 the model reached mAP50 ≈ 0.47, Precision ≈ 0.82, Recall ≈ 0.75.
- YOLOv8s fine-tuned on 15 Minecraft mob classes (Roboflow base + 4 merged extras)
- Custom class-ID remapping across heterogeneous sources
- Real-time inference loop: mss screen capture → YOLO → tkinter display
- tkinter chosen over cv2.imshow to dodge opencv-python vs opencv-python-headless conflict
Key features
YOLOv8s
CSPDarknet + PANet + decoupled head
15 mob classes
Creeper, enderman, villager, …
Live capture
mss screen grab @ ~30fps
mAP50 0.47
After 100-epoch fine-tune
Screenshots
Detecting polar bear
Detecting slime
Tech stack
Other projects
Three-Class Building Damage Classifier
Drone-image classifier for undamaged / damaged / collapsed buildings, comparing 12 model variants and a Grad-CAM-guided crop pipeline.
No Car No Problem — Ride-sharing platform
Full ride-sharing app with customer/driver/admin/support roles, real-time chat, geospatial ride matching, and driver verification workflow.