taozh2017 / hinet Goto Github PK

Dear taozh2017:
Thank you very much for sharing the code.
I have a confusion that want to ask you, I have accomplished the train process, and have got the discrimator.pkl and generator.pkl file.But when I test, only the mse, ssim and psnr values are displayed, how can I get the experimental images, for example, T1 + Flair to T2. After using the test, no experimental result images in any folder, what should I do?
@taozh2017

It seems that you have added the four losses at the end and have then backpropagated it through the generator. But how is the final loss getting backpropagated through the all the three branches?

🐛 BUG

According to

down_fu_2 = self.down_fu_2(down_2_0m, down_2_1m, down_fu_1m)
down_fu_3 = self.down_fu_3(down_3_0m, down_3_1m, down_fu_2m)

instead.

I downloaded the BraTs2018 dataset from the website, but it seems that the format in this dataset is nii while the files in data directory are in mat format. Could you please show me how to a file in nii format to mat format with python?

Can you provide the model you used for testing? Can you share your test commands?

网络结构

在syn_model的__init__函数中所定义的out1和out2在forward方法里面没有使用，主要是下面两个地方：

• out1
• out2

dataset类

从loadSubjectData可以看出数据是三维的，但是随后在generate_all_patches函数里面数据又是二维的

能否解释下这两个地方，先谢谢大佬了

The data folder of this repository is very large, about 100M. This will cause the cloning of the library to be very slow. It is recommended to store it in Google Cloud Disk or Baidu Cloud Disk and provide a download link.

您好，我最近复现了您的程序，但是在数据集方面有一些问题，可以冒昧的向您要一下您当时使用的数据集吗，谢谢您的帮助

Hi there, thanks for your work, it's really inspiring!
I notice that in synthesis network, you use multiple MFB to mix feature maps, and in the MFB, there is a multiple computation.
With so many feature map in this computation, how can you provide from the NaN problem when training?
Because mul especially in 3D Volumn, might cause overflow problem and I use MFB in mixed precision mode.

Thanks a lot!

I have realized the training process, but I don't know how to test. Could you tell me how to test it?

Sir, so sorry to bother you again. When I tried to run the code with the data example in data directory, there was problem in dataset.py, line 28, "img_t1 = img_t1[40:200, 20:200, :]". I print the shape of img_t1, which is (240,240). But the code in this line is to operate a 3-D ndarray, while img_t1 is a 2-D ndarray. And the paper of HiNet shows "For a 2D slice (240 × 240), we crop out an image of size 160 × 180 from the center region." So, what should I do to avoid this error, change the code or the data format? So sorry for my poor English and programming ability. And thanks again for your patient guidance.

您好，我按照您论文中描述的进行数据集的预处理，进行训练，但是一直达不到您论文中提到的指标，差距比较大，请问您愿意分享您当时使用的数据集以及您的参数设置吗，感谢您的帮助！