Plant Disease Detection using Deep Learning

Overview

This project implements a deep learning pipeline for detecting plant diseases using the PlantVillage dataset. It leverages PyTorch for training and testing convolutional neural networks (CNNs) on plant leaf images, providing accurate classification of various plant diseases.

Features

• Supports multiple pre-trained models, including MobileNetV2, SqueezeNet 1.1, EfficientNet-B0, and ShuffleNet V2.
• Implements transfer learning with optional freezing of model layers.
• Early stopping and learning rate scheduling to optimize training.
• Evaluation metrics: Accuracy, Precision, Recall, and F1 Score.
• Visualization of training statistics (loss and accuracy).

Requirements

• Python 3.7+
• PyTorch 1.10+
• torchvision
• NumPy
• Matplotlib
• argparse
• PlantVillage dataset (downloaded automatically)

Setup

1. Clone the repository:

   git clone <repository_url>
   cd <repository_name>

2. Install the required dependencies:

   pip install -r requirements.txt

3. Ensure GPU support by installing the appropriate version of PyTorch for your system from pytorch.org.

Usage

Commands

The script provides two primary commands:

• train: Train a model on the PlantVillage dataset.
• test: Evaluate a trained model on the test dataset.

Training

Run the following command to train a model:

python main.py train --model <model_name> --freeze

Parameters

• --model: Specify the model architecture (e.g., mobilenet_v2, squeezenet1_1, efficientnet_b0, shufflenet_v2).
• --freeze: Optional flag to freeze the pre-trained model's layers.

Example:

python main.py train --model mobilenet_v2 --freeze

Testing

To evaluate a trained model, run:

python main.py test --model <model_name> --freeze

Parameters

• --model: Specify the model architecture used during training (e.g., mobilenet_v2, squeezenet1_1, efficientnet_b0, shufflenet_v2).
• --freeze: Optional flag to freeze the model's layers.

Example:

python main.py test --model mobilenet_v2 --freeze

Outputs

• Trained model weights are saved in the weights/ directory.
• Training statistics (loss and accuracy) are visualized after training.
• Evaluation metrics (Accuracy, Precision, Recall, F1 Score) are printed during testing.

Project Structure

.
├── main.py              # Entry point for training and testing
├── data.py              # Dataset preparation and utilities
├── model.py             # Model architecture definitions
├── train.py             # Training logic
├── eval.py              # Evaluation and prediction utilities
├── vis.py               # Visualization utilities
├── weights/             # Directory for saving model weights
└── requirements.txt     # Required Python packages

Key Hyperparameters

• Batch size: 64
• Image size: 224x224
• Epochs: 50
• Learning rates:
  • Fine-tuning: 0.0001
  • Standard training: 0.001
• Weight decay: 0.0001
• Early stopping:
  • Patience: 5 epochs
  • Delta: 0.001
• Learning rate scheduler:
  • Type: ReduceLROnPlateau
  • Patience: 3 epochs
  • Factor: 0.1

Notes

• Ensure the weights/<model_name>.pth file exists before testing. If not, train the model first.
• The PlantVillage dataset is automatically downloaded and prepared by the script.

References

• PlantVillage Dataset
• PyTorch Documentation
• MobileNetV2: Inverted Residuals and Linear Bottlenecks
• SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size
• EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
• ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design
• Transfer Learning with PyTorch

License

This project is currently not licensed. In the meantime, feel free to explore, modify, and use the code for educational purposes.

Acknowledgments

• The creators of the PlantVillage dataset for their valuable contributions to plant disease research.
• The PyTorch community for providing extensive documentation and tools.