CEmb-SAM

This is the official repository for CEmb-SAM: Segment Anything Model with Condition Embedding for Joint Learning from Heterogeneous Datasets.

Getting Started

We provide GUI to test on sample images

Install PyQt5

python gui.py --sam_ckpt <path/to/sam_vit_b/checkpoint> --ckpt <path/to/CEmbSam/checkpoint> --emb_class <the_number_of_embedding_classes>

Load the image to the GUI and specify segmentation targets by drawing bounding boxes.

Training Model (CEmb-sam)

Datasets Preparation

We use two datasets, the public benchmark BUSI dataset and Nerve dataset.

BUSI

Training

The model was trained on RTX3090.

Sample result images

Segmentation results on BUSI (1st and 2nd rows) and peripheral nerve dataset (3rd and 4th rows).

Reference

@inproceedings{shin2023cemb,
  title={CEmb-SAM: Segment Anything Model with Condition Embedding for Joint Learning from Heterogeneous Datasets},
  author={Shin, Dongik and Kim, MD, Beomsuk and Baek, Seungjun},
  booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
  pages={275--284},
  year={2023},
  organization={Springer}
}