+ Computer Vision

PaddyCare AI: Rice Leaf Disease Classifier

EfficientNet-B0 classifier that detects 6 rice leaf conditions (blast, brown spot, dead heart, hispa, tungro, healthy) with a web interface.

View source on GitHub

Web interface with dual input methods: file upload and URL-based image processing for rice leaf images

Overview

An agricultural computer vision system for early detection of rice leaf diseases. A PyTorch EfficientNet-B0 model (ImageNet pre-trained) is fine-tuned on a class-balanced dataset with augmentation, early stopping and learning-rate scheduling, and evaluated with per-class precision, recall, F1 and a confusion matrix. Users upload a photo or URL and get the predicted condition with confidence.

Highlights

6-class classification: blast, brown spot, dead heart, hispa, tungro and healthy
Transfer learning with EfficientNet-B0
Balanced training, augmentation, early stopping and LR scheduling
Per-class precision/recall/F1 and confusion-matrix evaluation

Tech stack

Python
PyTorch
EfficientNet-B0
OpenCV
Scikit-learn
Pandas

Gallery

Example input image of rice leaf showing disease symptoms before classification

Classification result displaying predicted disease class, confidence score, and health status