Multi-scale attention subtraction networks and an ultra-wide field retinal hemorrhage dataset

Renkai Wu, Pengchen Liang, Yiqi Huang, Qing Chang* and Huiping Yao*

1. Ruijin Hospital, Shanghai Jiao Tong University School of Medicine, Shanghai, China
2. Shanghai University, Shanghai, China

News🚀

(2024.10.10) The UWF-RHS dataset is now fully public. If you need it, follow the requirements below to submit an application.🔥

(2024.09.17) Our model code has been uploaded! The UWF-RHS Dataset will provide access links next. Stay tuned!

(2024.09.10) This work has been accepted for early access by IEEE Journal of Biomedical and Health Informatics!🔥

(2024.03.02) Upload the corresponding running code.

(2024.03.02) Manuscript submitted for review. 📃

0. Main Environments.

• python 3.8
• pytorch 1.12.0

1. The proposed datasets (UWF-RHS).
(1) To obtain the UWF-RHS dataset, you need to provide your name, affiliation, reason for applying (one sentence description of your work), and an assurance that you will not share the dataset privately. Specifically, complete the information in the following format and send it to '[email protected]' with the subject name 'UWF-RHS dataset request'. We will usually review your request and provide you with a link within 3 days. If you do not register your information as required, your application may fail. Please understand!

Name:
Affiliation:
Reason for applying (one sentence description of your work):
I (the applicant) guarantee that the data will be used only for academic communication and not for any commercial purposes. I (the applicant) guarantee that I will not privately disseminate the data to any public place without the consent of the author.

video.mp4

2. Train the MASNet.

python train.py

• After trianing, you could obtain the outputs in './results/'

3. Test the MASNet.
First, in the test.py file, you should change the address of the checkpoint in 'resume_model' and fill in the location of the test data in 'data_path'.

python test.py

• After testing, you could obtain the outputs in './results/'

Citation

If you find this repository helpful, please consider citing:

@article{wu2024automatic,
  title={Automatic Segmentation of Hemorrhages in the Ultra-wide Field Retina: Multi-scale Attention Subtraction Networks and An Ultra-wide Field Retinal Hemorrhage Dataset},
  author={Wu, Renkai and Liang, Pengchen and Huang, Yiqi and Chang, Qing and Yao, Huiping},
  journal={IEEE Journal of Biomedical and Health Informatics},
  year={2024},
  publisher={IEEE}
}