Label 20 images and let the model suggest labels for the next thousand
Advanced28h2–3
PythonCLIPEmbeddings
16/20Plant Leaf Disease Spotter
Photograph a leaf and see the likely disease plus treatment steps.
Intermediate20h2–4 people16/20 fit
Build Plant Leaf Disease Spotter for home gardeners and small farmers in ~20h with 2–4 people. Stack: Python + YOLO + Next.js.
Why it wins: Real agricultural impact and a classic computer-vision demo.
Build flow
01 · Problem
Farmers lose crops because diseases are spotted late and expert advice is far away.
02 · Solution
Fine-tune or use a pretrained image classifier on a public plant-disease dataset, show confidence and link to treatment guidance with a 'consult an expert' note.
03 · Build
- 1Pick a public leaf-disease dataset
- 2Train or fine-tune a small classifier
- 3Serve predictions from an API
- 4Show results with confidence
04 · Demo
Show three leaf photos and the top-3 predictions with confidence bars.
MVP checklist
- Leaf photo upload
- Top-3 predictions
- Treatment guidance
- Low-confidence warning
Tech stack
| Tech | Role | |
|---|---|---|
| Python | Backend / scripting | Backend |
| YOLO | Object detection | AI |
| Next.js | Frontend + API routes | Web |
Reference repos
Open-source projects you can build on. Links checked 2026-10-31.
Related ideas
More AIIncoming tickets get a category, urgency and suggested reply automatically
Intermediate14h1–3EmbeddingsLLMPostgres16/20Climate
Recycling Sorter Camera
Point a camera at an item and it tells you which bin it goes in
Intermediate20h2–3YOLOPythonNext.js16/20