https://github.com/aleemsidra/SaLIP

This repository contains the pytorch code for our CVPRW 2024 paper "Test-Time Adaptation with SaLIP: A Cascade of SAM and CLIP for Zero-shot Medical Image Segmentation".
https://github.com/aleemsidra/SaLIP

Last synced: about 1 month ago
JSON representation

This repository contains the pytorch code for our CVPRW 2024 paper "Test-Time Adaptation with SaLIP: A Cascade of SAM and CLIP for Zero-shot Medical Image Segmentation".

• Host: GitHub
• URL: https://github.com/aleemsidra/SaLIP
• Owner: aleemsidra
• License: mit
• Created: 2024-04-18T14:09:13.000Z (9 months ago)
• Default Branch: main
• Last Pushed: 2024-04-18T14:16:14.000Z (9 months ago)
• Last Synced: 2024-04-18T16:03:42.757Z (9 months ago)
• Language: Python
• Homepage:
• Size: 376 KB
• Stars: 1
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• Awesome-Segment-Anything - [code

README

        
## Test-Time Adaptation with SaLIP: A Cascade of SAM and CLIP for Zero shot Medical Image Segmentation

This project is an implementation of the paper ["Test-Time Adaptation with SaLIP: A Cascade of SAM and CLIP for Zero shot Medical Image Segmentation"](https://arxiv.org/pdf/2404.06362.pdf),  accepted at [CVPRW 2024].

## SaLIP



## Setup

To set up the project environment using conda, follow these steps:

1. Clone the repository: ```[email protected]:aleemsidra/SaLIP.git```

2. Navigate to the project directory: ```cd SaLIP```

3. Create a conda environment: ```conda create --name sam python=3.8```

5. Download SAM's checkpoint from [here ](https://github.com/facebookresearch/segment-anything?tab=readme-ov-file#model-checkpoints)

## Arguments

The ```main.py``` script is the main entry point of the project. It performs the following steps:

```mode:``` Specifies the baseline e.g. sam_clip, sam_clip_3d

```prompt_mode:``` Specifies the visual prompt mode for CLIP (default: "crops")

```vit_name: ``` Specifies SAM's checkpoint version (default: "vit_h" )

```data:``` Specifies the dataset 

```config:``` Specifies the path to config file

```seed:``` Specifies seed value for reproducibility

## Example command

```

python main.py --config ./config/lung.json --mode "sam_clip"  --prompt_mode "crops" --dataset "cxr" --seed 1234 

```

## Contact

Feel free to raise an issue or contact me at [email protected] for queries and discussions.