Workflowhealtree
HealTree
A commission project built with 3 other devs — a mobile app for real-time tree disease detection using YOLOv8 and TensorFlow via smartphone cameras.
stack
- Flutter
- Firebase
- Python
- TensorFlow
Architecture
- YOLOv8 model — trained on leaf spot, blight, rust, and canker; TensorFlow for model optimization
- Natural-variation class and augmentation — "healthy with natural variation" class plus color jittering and brightness shifts to cut false positives
- Real-time detection — camera feed yields disease identification with confidence scores
- Offline inference — model runs on-device via TensorFlow Lite
- Firebase integration — user accounts, detection history, and monitoring dashboards
Results
- Real-time disease detection via smartphone camera
- On-device inference for offline use in rural areas
- Firebase-backed monitoring history for tracking disease spread
Action · Send message
Hiring for a remote AI engineering role?
Contact me