Classes
3
Healthy, Powdery, Rust
Orientation
S
Quick Project Summary
The migrated stack keeps Streamlit feature parity while adding a cleaner UI layer and API-first backend.
Feature ParityNext.js + FastAPITensorFlow Inference
Core Routes
5
Summary to Performance
Detector Inputs
2
Upload files and URLs
Business Direction
- Automates healthy vs powdery vs rust detection from leaf images.
- Supports multi-image and multi-URL predictions with confidence scores.
- Delivers treatment guidance and downloadable CSV analysis reports.
Dataset Context
Uses the apple leaf dataset split into Train, Validation, and Test subsets with three classes: Healthy, Powdery, and Rust.
Read complete project background in the repository README.
Architecture Snapshot
Frontend handles interaction and presentation. Backend handles prediction, CSV reports, metrics, and montage generation.
Current Parity Status
Summary, Visualizer, Detector, Hypothesis, and Performance routes are now available in Next.js.