This is the official repository for CEmb-SAM: Segment Anything Model with Condition Embedding for Joint Learning from Heterogeneous Datasets.
We provide GUI to test on sample images
- GUI
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.
We use two datasets, the public benchmark BUSI dataset and Nerve dataset.
- BUSI
The model was trained on RTX3090.
Segmentation results on BUSI (1st and 2nd rows) and peripheral nerve dataset (3rd and 4th rows).
@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}
}