junqiangchen

运行dataprocess里面的LUNA_node_extraction.py就可以生成0和1目录下的npy文件的

> 函数 def normalize(slice, bottom=99, down=1): > 能解释一下99和1这两个取值的含义吗？ > > Thanks！ 你好，你可以看一下使用这个变量的地方，我记得是用来处理异常值的。

> Hi, > > https://github.com/deep-voxel/automatic_pulmonary_lobe_segmentation_using_deep_learning talks about only the masks data. could please share the source for image data. > > Thanks you can download source image data from luna16...

> Dear, > I always saw your repo. This is excellent work. Please could be determined which training samples from luna16 segmentation training samples, I need to acquired. Thanks in...

具体的参数配置可以贴出来么？还有就是确保你的tf版本是多少，我的是1.14

@demoyu123 you can follow this project[https://github.com/junqiangchen/PytorchDeepLearing](url)

you can find the trained model from wechat public artical

> I have done this before with the same training data. Can you make a tutorial for the challenge "Pulmonary Vasculature Segmentation" in the same workstation. thank you,i have implemented...

@sst-LoveStudy 正样本使用全部，负样本使用部分数据

@Jonydom 您好，预处理好的数据集是指npy文件么，其实这个处理在dataprocess中有的，参考这个https://github.com/junqiangchen/PytorchDeepLearing/blob/main/dataprocess/data3dpreparewithSize.py