Plant Disease Detection Using CNN

Overview

Automated plant disease detection using deep convolutional neural networks to help identify plant diseases from leaf images. Built with TensorFlow and Keras.

Dataset

PlantVillage dataset from Kaggle: https://www.kaggle.com/sumanismcse/plant-disease-detection-using-keras

Preprocessing

• Cropped images with 500+ pixel resolution
• Manual labeling with disease classifications
• Removed duplicates and low-quality images

Model Architecture

Deep CNN with:

• Convolutional layers with ReLU activation
• Batch normalization
• Max pooling
• Dropout
• Fully connected layers

Results

Accuracy: 96.77%

Tested on validation set with individual class performance metrics.