Plant leaf diseases are a major threat to global agricultural productivity.
Early and accurate detection is essential for preventing crop damage and ensuring food security.
This project presents a comparative analysis of five Convolutional Neural Network (CNN) architectures for automated plant leaf disease classification using a publicly available dataset.

We implemented and evaluated:

• Basic CNN
• VGG16
• VGG19
• ResNet50
• InceptionResNetV2

The dataset was preprocessed with normalization, augmentation, and class balancing to improve generalization.
Models were trained under identical experimental conditions to ensure fair comparison.

• Source: Plant Disease Recognition Dataset
• Classes: Multiple plant disease categories + healthy leaves
• Data Splits: Train / Validation / Test
• Image Size: 224 × 224 pixels
• Augmentation Techniques:
  • Random Flip
  • Random Rotation
  • Random Zoom
  • Random Contrast

After training all models on the Plant Disease Recognition Dataset, we evaluated them using Accuracy, Precision, Recall, and F1-Score.
The table below summarizes the performance of each architecture:

Key Observations:

• The Basic CNN achieved the highest accuracy (93.99%) while maintaining low computational complexity, making it suitable for real-time and resource-limited applications.
• InceptionResNetV2 performed competitively with 92% accuracy, excelling in complex feature extraction but requiring more computational resources.
• VGG16 and VGG19 achieved moderate accuracy, performing well on high-contrast lesions but struggling with subtle disease symptoms.
• ResNet50 performed poorly on this dataset due to insufficient domain-specific fine-tuning, with an accuracy of only 36.67%.

Visual Evaluation:

• Confusion matrices were plotted for each model to analyze per-class performance.
• Predicted image samples demonstrated that Basic CNN and InceptionResNetV2 made consistent predictions across disease classes.