Request for Detailed Information on Replicating Segmentation Performance with U-Net Model

[CVKim]

Hi Author,

I would like to replicate the segmentation performance evaluation as described in your paper.

Could you please provide detailed information about the U-Net code and the testing methodology? I am having difficulty understanding the setup based on the materials mentioned in the paper.

Specifically, I would like to know how the train/test (validation) sets were constructed, how the masks and generated images were prepared, and how many pairs were used.

Additionally, any information on the hyperparameter settings used during the U-Net training would be greatly appreciated!

Wishing you a great day! I look forward to your response.

Thank you!!

[sjtuplayer]

Hi, you can refer to https://github.com/sjtuplayer/anomalydiffusion/blob/master/train-localization.py