Xinyu Xiong, Zihuang Wu, Shuangyi Tan, Wenxue Li, Feilong Tang, Ying Chen, Siying Li, Jie Ma, Guanbin Li
Image segmentation plays an important role in vision understanding. Recently, the emerging vision foundation models continuously achieved superior performance on various tasks. Following such success, in this paper, we prove that the Segment Anything Model 2 (SAM2) can be a strong encoder for U-shaped segmentation models. We propose a simple but effective framework, termed SAM2-UNet, for versatile image segmentation. Specifically, SAM2-UNet adopts the Hiera backbone of SAM2 as the encoder, while the decoder uses the classic U-shaped design. Additionally, adapters are inserted into the encoder to allow parameter-efficient fine-tuning. Preliminary experiments on various downstream tasks, such as camouflaged object detection, salient object detection, marine animal segmentation, mirror detection, and polyp segmentation, demonstrate that our SAM2-UNet can simply beat existing specialized state-of-the-art methods without bells and whistles.
git clone https://github.com/WZH0120/SAM2-UNet.git
cd SAM2-UNet/You can refer to the following repositories and their papers for the detailed configurations of the corresponding datasets.
- Camouflaged Object Detection. Please refer to FEDER.
- Salient Object Detection. Please refer to SALOD.
- Marine Animal Segmentation. Please refer to MASNet.
- Mirror Detection. Please refer to HetNet.
- Polyp Segmentation. Please refer to PraNet.
Our project does not depend on installing SAM2. If you have already configured an environment for SAM2, then directly using this environment should also be fine. You may also create a new conda environment:
conda create -n sam2-unet python=3.10
conda activate sam2-unet
pip install -r requirements.txtIf you want to train your own model, please download the pre-trained segment anything 2 from the official repository. You can also directly download sam2_hiera_large.pt from here. After the above preparations, you can run train.sh to start your training.
Our pre-trained models and prediction maps can be found on Google Drive. Also, you can run test.sh to obtain your own predictions.
After obtaining the prediction maps, you can run eval.sh to get most of the quantitative results. For the evaluation of mirror detection, please refer to eval.py in HetNet to obtain the results.