cell-detection-package

#need to use better frontend

Goal

#Goal The cell-detection-package is a Python package designed for detecting and segmenting cells in histology images. This package leverages deep learning models to provide accurate and efficient cell detection, facilitating research and analysis in pathology. It supports images of any size, including whole slide images.

Installation

First, clone the repository and install the necessary dependencies.

git clone https://github.com/idso-fa1-pathology/cell-sight-package.git
cd cell-sight-package
pip install -r requirements.txt

Ensure YOLO Command is Available

After installing the dependencies, you need to ensure the yolo command is available in your PATH. This can be done by adding the installation path to your PATH environment variable.

export PATH=$PATH:~/.local/bin

To make this change permanent, add the above export command to your ~/.bashrc or ~/.profile file:

echo 'export PATH=$PATH:~/.local/bin' >> ~/.bashrc
source ~/.bashrc

Usage

Predict New Images

To predict new images using the pre-trained model, follow these steps:

Download Pre-trained Model:

The pre-trained model weights will be automatically downloaded and saved in the runs/pretrained-model/weights directory when you run the prediction script.

1. Upload your images to the data/predict/images/ directory.

2. Run the prediction script with the following command:

   python scripts/predict.py --images_directory data/predict/images/ --output_directory runs/predict/

3. The results will be saved in the specified output_directory, with annotated images and prediction data.

Train the Model

To train the model on your own dataset, follow these steps:

1. Organize your dataset into three main folders: train, test, and valid. Each of these folders should contain two subfolders: images and labels.

   • images/: Contains the image files.

   • labels/: Contains the label files, where each label file corresponds to an image file with the same name but a .txt extension. The format of each line in the label file should be as follows:

     class_id center_x center_y width height
     

     • class_id: The class ID of the object.
     • center_x: The x-coordinate of the center of the bounding box, normalized to the [0, 1] range.
     • center_y: The y-coordinate of the center of the bounding box, normalized to the [0, 1] range.
     • width: The width of the bounding box, normalized to the [0, 1] range.
     • height: The height of the bounding box, normalized to the [0, 1] range.

     Example content of a label file:

     0 0.942446 0.0212446 0.0742872 0.0422503
     0 0.124674 0.327 0.0790858 0.078888
     0 0.527016 0.342674 0.0800738 0.0796084
     

2. Make sure all input images are resized to 640x640 pixels.

3. Place your dataset in the data/raw/ directory.

4. Run the training script with the following command:

   python scripts/train.py --config_path data/raw/config.yaml

5. The model will be trained using the images and labels in the specified directories, and the results will be saved in the runs/trained/ directory.

Running with Docker

To run the cell-sight-package Docker image, follow these steps:

1. Pull the Docker image:

   docker pull yshokrolllahi/cell-sight-package:latest

2. Run the Docker container:

   docker run -it --rm yshokrolllahi/cell-sight-package:latest

This will start the container and run your application as defined in the Dockerfile.

From GitHub Packages

1. Pull the Docker image:

    docker pull ghcr.io/idso-fa1-pathology/cell-sight-package:latest

2. Run the Docker container:

    docker run -it --rm ghcr.io/idso-fa1-pathology/cell-sight-package:latest
   

docker run -p 5001:5001 -v /Users/yshokrollahi/Documents/GitHub/cell-sight-package/backend:/app/backend -v /Users/yshokrollahi/Documents/GitHub/cell-sight-package/frontend:/app/frontend cell-sight-package:latest

This will start the container and run your application as defined in the Dockerfile.