search

shenqq377 / Q-Net

Python codes to implement Q-Net, a meta-learning method for few shot medical image segmentation
https://arxiv.org/abs/2208.11451
42 stars 5 forks source link

What is the purpose of generating a class slice index for setting up the experiment (i.e. classmap.json) #15

Closed WangBingJian233 closed 9 months ago

WangBingJian233 commented 9 months ago

Dear author, thank you very much for your outstanding contribution. I don't understand the function of classmap.json file. It seems that this file is not used in the subsequent code. Could you explain its function? Thank you! image

shenqq377 commented 9 months ago

These files are generated by superpixel or supervoxel segmentation. We didn't use these files in our code.

WangBingJian233 commented 9 months ago

这些文件由超像素或超体素分割生成。我们没有在代码中使用这些文件。 You mean that classmap.json is not used in the whole project, so you don't need it, right? Because I found that the code generate_supervoxels.py, which generates hypervoxels, does not use the classmap.json file.

shenqq377 commented 9 months ago

Yeah, as you may notice that we read supervoxel files (e.g., data/CHAOST/supervoxel_xx) directly in dataloader/datasets.py (Line 141~142).

WangBingJian233 commented 9 months ago

Ok, thank you for your reply! I wish you all the best!