Label of BraTS

[YonghengSun1997]

你好，BraTS原始数据是4类label，您是做的哪类的分割？还是几类的并集作为一类？

[zzzqzhou]

你好，我们是将所有非背景的类别都合并为一类，你可以参考这段代码

Dual-Normalization/preprocess_func.py

Line 95 in 8cac047

masks[masks != 0] = 1

[freedom199712]

I reproduced your code, trained on t2, tested on t1 but dice was only 0.2. Can you get back to me

[booooming]

同复现训练集源域t2，其余模态为测试，结果dice0.2

[koi805]

同，我的也是t2上训练，t1上测试为0.2左右，flar稍微高点，但也只有50不到，远低于论文中的结果