search

assassint2017 / MICCAI-LITS2017

liver segmentation using deep learning
388 stars 91 forks source link

预测结果显示背景是1， 分割区域是0 #25

Open zhouwenqiong opened 2 years ago

zhouwenqiong commented 2 years ago

麻烦问一下为什么按照您的步骤最后预测处理的结果和真实标签0 1 是刚好相反的，没有改动代码，如果您看到希望能收到您的回信

maosaomi commented 2 years ago

麻烦问一下为什么按照您的步骤最后预测处理的结果和真实标签0 1 是刚好相反的，没有改动代码，如果您看到希望能收到您的回信

请问您解决了吗 我也是这样的问题

SuperBianC commented 2 years ago

麻烦问一下为什么按照您的步骤最后预测处理的结果和真实标签0 1 是刚好相反的，没有改动代码，如果您看到希望能收到您的回信

you need to change the code: liver_seg = measure.label(liver_seg, 4) to liver_seg = measure.label(liver_seg, connectivity=2) It will work. That's because the package measure has been updated, and the parameters were changed.