https://github.com/joeybyc/cell_detection

https://github.com/joeybyc/cell_detection

Last synced: about 1 month ago
JSON representation

• Host: GitHub
• URL: https://github.com/joeybyc/cell_detection
• Owner: joeybyc
• Created: 2024-05-29T11:49:22.000Z (8 months ago)
• Default Branch: main
• Last Pushed: 2024-11-24T20:08:02.000Z (about 2 months ago)
• Last Synced: 2024-11-24T21:19:39.415Z (about 2 months ago)
• Language: Python
• Size: 3.4 MB
• Stars: 2
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• Awesome-Segment-Anything - [code

README

        
# Anterior Chamber Cell Detector (ACCDor)

This repository contains the code for ACCDor, an anterior chamber cell detector.

## Installation

1. Create a new conda environment with Python 3.9:

```

conda create -n accdor python=3.9

```

2. Activate the conda environment:

```

conda activate accdor # sometimes could be `source activate accdor`

```

3. Install the required dependencies:

```

pip install -r requirements.txt

```

## Checkpoints

ACCDor requires a pre-trained ViT model. Download the model from the [ViT-H SAM model](https://dl.fbaipublicfiles.com/segment_anything/sam_vit_h_4b8939.pth) link.

The link is provided in the [SAM](https://github.com/facebookresearch/segment-anything?tab=readme-ov-file#model-checkpoints) repository. Put the pre-trained model into the **models** folder.

You can also use

```

wget https://dl.fbaipublicfiles.com/segment_anything/sam_vit_h_4b8939.pth

```

to download the ViT model.

## Getting Started

To process an image, segment the AC area, and detect cells, run the following command:

```

python -m apps.detect_cell

```

By default, this command will process the image located at `data/example/example1.jpeg`. The output will be saved in the `data/output/{image_name}` directory, where `{image_name}` is the name of the example image (in this case, `example1`).

## Showcase

After running `apps.detect_cell`, intermediate stage images will also be generated. Below are examples of the generated images for the sample input image.

### Original image

![](data/example/example1.jpeg)

### Anterior Chamber Mask

![](data/output/example1/chamber_mask.png)

### Cell Mask (Processed by Adjusted Cutoff)

![](data/output/example1/candidate_cell_mask.png)

### Cell Mask (Discard the False Positive Candidate Cells)

![](data/output/example1/cell_mask.png)

### Cell Dection by ACCDor

![](data/output/example1/image_with_cell_bounding_boxes.png)

## Citation

arXiv: https://arxiv.org/abs/2406.17577

To cite ACCDor in publications, please use:

```bibtex

@article{chen2024advancing,

      title={Advancing Cell Detection in Anterior Segment Optical Coherence Tomography Images}, 

      author={Boyu Chen and Ameenat L. Solebo and Paul Taylor},

      year={2024},

      journal={arXiv preprint arXiv:2406.17577}

}

```

## Acknowledgements

Thanks to the support of AWS Doctoral Scholarship in Digital Innovation, awarded through the UCL Centre for Digital Innovation. We thank them for their generous support.

![](AWS.png)

![](CDI.png)