BTCV dataset

[numanai]

Hi @MrGiovanni!

Thank you for sharing the code.

I read in the paper that you are using 47 CT scans of the BTCV dataset but they have only 30 CT scans and their labels available publicly, am I missing something here?.

Best!

[numanai]

01_Multi-Atlas_Labeling/img/img0061.nii.gz 01_Multi-Atlas_Labeling/label/label0061.nii.gz
01_Multi-Atlas_Labeling/img/img0062.nii.gz 01_Multi-Atlas_Labeling/label/label0062.nii.gz
01_Multi-Atlas_Labeling/img/img0063.nii.gz 01_Multi-Atlas_Labeling/label/label0063.nii.gz
01_Multi-Atlas_Labeling/img/img0064.nii.gz 01_Multi-Atlas_Labeling/label/label0064.nii.gz
01_Multi-Atlas_Labeling/img/img0065.nii.gz 01_Multi-Atlas_Labeling/label/label0065.nii.gz
01_Multi-Atlas_Labeling/img/img0067.nii.gz 01_Multi-Atlas_Labeling/label/label0067.nii.gz
01_Multi-Atlas_Labeling/img/img0069.nii.gz 01_Multi-Atlas_Labeling/label/label0069.nii.gz
01_Multi-Atlas_Labeling/img/img0070.nii.gz 01_Multi-Atlas_Labeling/label/label0070.nii.gz
01_Multi-Atlas_Labeling/img/img0075.nii.gz 01_Multi-Atlas_Labeling/label/label0075.nii.gz
01_Multi-Atlas_Labeling/img/img0079.nii.gz 01_Multi-Atlas_Labeling/label/label0079.nii.gz
01_Multi-Atlas_Labeling/img/img0080.nii.gz 01_Multi-Atlas_Labeling/label/label0080.nii.gz

[northwill]

At the beginning of the training phase, it seems that a '. h5' file is needed in the code, and I don't know where to get it from

for line in open(args.train_data_txt_path):
        if line.startswith('#'):
            continue
        name = line.strip().split()[1].split('.')[0]
        train_img.append(args.data_root_path + line.strip().split()[0])
        train_lbl.append(args.data_root_path + line.strip().split()[1])
        train_post_lbl.append(args.data_root_path + name.replace('label', 'post_label') + '.h5')`