sunjesse / shape-attentive-unet

Code for our paper SAUNet: Shape Attentive U-Net for Interpretable Medical Image Segmentation. https://arxiv.org/pdf/2001.07645v3.pdf published at MICCAI 2020.
BSD 3-Clause "New" or "Revised" License
94 stars 18 forks source link

The threshold of the canny operator. #10

Open Max-Well-Wang opened 3 years ago

Max-Well-Wang commented 3 years ago

Thank you for sharing the source code of this excellent paper. Can you tell me the threshold set by the Canny operator in the experiment? The threshold in the source code is 10-100,But it seems to have no effect in AC17...

Max-Well-Wang commented 3 years ago

Hello author, I like your paper very much. In the past two months, I have fine-tuned the existing code based on all the parameters provided by the paper, but I have not been able to achieve the effect of your paper in the data set AC17. Are you sure you can train the paper to have such a good effect relying on the parameters of the paper and this code?

The following is my training process. origin_model_BS10.txt