FahimFBA / SafeLicensing

Safeguard vehicles' licenseplates from unauthorized access (Image and Video both)

SafeLicensing

This project demonstrates a pipeline for detecting license plates in images using YOLOv8 and encrypting the detected regions with a Chaotic Logistic Map encryption algorithm. It provides a user-friendly interface built with Streamlit.

Tip

You can directly test the application on the web using the following link:

Features

• License Plate Detection: Uses the YOLOv8 model to detect license plates in uploaded images.
• Chaotic Encryption: Encrypts the detected license plate regions using a two-layer XOR-based chaotic logistic map algorithm.
• Streamlit Web App: A simple interface to upload images, detect license plates, encrypt them, and download the results.

Installation

1. Clone the repository:

   git clone https://github.com/FahimFBA/SafeLicensing.git
   cd SafeLicensing

2. Install ffmpeg for video processing (Linux):

   sudo apt-get install ffmpeg

   or, for macOS:

   brew install ffmpeg

3. Install the required dependencies:

   pip install -r requirements.txt

4. Download the YOLOv8 weights file (best.pt) and place it in the root directory of the project. You can train your own model or use a pre-trained one. This repository already have our model from SEncrypt located in best.pt file.

Usage

1. Run the Streamlit app:

   streamlit run app.py

2. Open the app in your browser (typically at http://localhost:8501).

3. Follow the steps:

   • Upload an image or provide a URL.
   • Adjust the encryption key seed using the slider.
   • Click the "Detect & Encrypt" button to process the image.

4. Download the encrypted image directly from the app.

Docker Usage

You can run this application using Docker. There are two options: CPU-only and GPU-enabled.

CPU Version

1. Build the Docker image:

   docker build -t safelicensing .

2. Run the Docker container:

   docker run -d -p 8501:8501 --name safelicensing_container safelicensing

GPU Version (for NVIDIA GPU users)

1. Ensure you have the NVIDIA Container Toolkit installed. If not, follow the official installation guide.

2. Build the GPU-enabled Docker image:

   docker build -f Dockerfile.gpu -t safelicensing-gpu .

3. Run the GPU-enabled Docker container:

   docker run -d -p 8501:8501 --gpus all --name safelicensing_gpu_container safelicensing-gpu

Common Steps

4. Check if the container is running:

   docker ps

   You should see a container named safelicensing_container or safelicensing_gpu_container in the list.

5. View the container logs:

   docker logs safelicensing_container  # or safelicensing_gpu_container for GPU version

   This will show you the Streamlit startup logs and any errors if they occur.

6. Open the app in your browser by navigating to http://localhost:8501.

If you encounter any issues:

• Ensure that port 8501 is not being used by another application.
• Check the container logs for any error messages.
• If needed, you can stop and remove the container using:

  docker stop safelicensing_container  # or safelicensing_gpu_container
  docker rm safelicensing_container  # or safelicensing_gpu_container

  Then, try running the container again.

Note: The GPU version requires an NVIDIA GPU and proper drivers. If you don't have a compatible GPU, use the CPU version instead.

Workflow

1. License Plate Detection:

   • The YOLOv8 model is used to detect license plates in the input image. The model has been taken from SEncrypt.
   • Detected regions are highlighted with bounding boxes.

2. Chaotic Logistic Map Encryption:

   • A chaotic logistic map generates two XOR-based encryption keys.
   • Pixels in the license plate regions are shuffled and encrypted in two stages.
   • The encrypted region replaces the original plate in the image.

3. Visualization and Download:

   • The original, detected, and encrypted images are displayed in the app.
   • Encrypted images can be downloaded as PNG files.

Files

• app.py: The main Streamlit app file.
• requirements.txt: Python dependencies for the project.
• best.pt: YOLOv8 weights file (not included, add your own).

Key Parameters

• Encryption Key Seed: A slider in the app adjusts the seed value for the chaotic logistic map, affecting the encryption's randomness.

Example Screenshots

Original Image

Encrypted Image

License

This project is licensed under the MIT License. See the LICENSE file for more details.

Contact

For any queries, feel free to reach out:

• Author: Md. Fahim Bin Amin

• GitHub: FahimFBA

• Other Authors: Rafid Mehda, Israt Jahan Khan