Student-Teacher Model for Image Anomaly Detection

This repository contains the PyTorch implementation of our paper published in the IEEE Access journal. You can read the full paper here: IEEE Xplore.

📌 Overview

This project implements a Student-Teacher model for anomaly detection in images. The method leverages a teacher network to guide a student network in learning normal patterns, allowing anomalies to be detected as deviations.

🚀 Getting Started

1. Prerequisites

Ensure you have Python 3.6 and PyTorch 1.6.0 installed before running the code.

2. Install Dependencies

Run the following command to install all required dependencies:

pip install -r requirements.txt

3. Prepare Datasets

Download and place the required datasets in the Datasets/ folder. Ensure the data structure matches the expected format.

4. Configure Hyperparameters

Modify hyperparameters such as the number of epochs, learning rate, etc., in the configs.yaml file located in the configs/ directory.

🔥 Training the Model

To train the model, use the following command:

python train.py

🎯 Testing the Model

To evaluate the model on test data, run:

python test.py

📂 Project Structure

Student-Teacher-Model-For-Image-Anomaly-Detection/
│── configs/           # Configuration files (configs.yaml)
│── datasets/          # Folder to store datasets
│── models/            # Model architecture definitions
│── utils/             # Helper functions and utilities
│── train.py           # Script for training the model
│── test.py            # Script for testing the model
│── requirements.txt   # List of dependencies
│── README.md          # Project documentation

🛠 Troubleshooting & Tips

• Ensure you are using the correct Python and PyTorch versions.
• If you encounter missing dependencies, manually install them using pip install <package-name>.
• Adjust hyperparameters in configs.yaml for better performance.

📜 Citation

If you find this work useful, please cite our paper:

@article{rakhmonov2023extensive,
  author    = {Rakhmonov, Akhrorjon Akhmadjon Ugli and Subramanian, Barathi and Olimov, Bekhzod and Kim, Jeonghong},
  title     = {Extensive knowledge distillation model: An end-to-end effective anomaly detection model for real-time industrial applications},
  journal   = {IEEE Access},
  year      = {2023},
  doi       = {10.1109/ACCESS.2023.3293108}
}