search

Tramac / DenseVoxNet-tensorflow

DenseVoxNet in TensorFlow
21 stars 3 forks source link

test question #1

Open kkkkkk1234 opened 5 years ago

kkkkkk1234 commented 5 years ago

hello , I am working with this job, and i use this with the data of the 'training_axial_crop_pat0-label.nii' and 'training_axial_crop_pat0.nii' ,and the number of the data is ten ,after I finished the step of 'python prepare_h5_data.py' and 'python train.py' , it works well ,but unfortunately when I go to the next step ,I found that it couldn't show any results ,I want to know if i have any wrong with this ? I sincerely hope I can get your replys , thank you very much!

Tramac commented 5 years ago

Hello, are there relevant error logs?

kkkkkk1234 commented 5 years ago

Hello,Firstly,thank you for the reply.It did't show any error,it just finished，and turn to the next command line without any outputs.So what's wrong with this ?I Followed the steps in Read Me file.Thank you for your help.Looking forward to your next reply!Thanks!

Tramac commented 5 years ago

This situation occurred because of the lack of visualization. Firstly, the prediction results were saved in fusion_Img_L2 and vote_label (both nparray format) in train.py. You can save them in any format you want, such as npy, mat, nrrd or others. And then visualize the slice in Python. I am very sorry that I don't have enough time to complete this part now, but I will add it as soon as possible. If there are other questions, please continue to post here.

kkkkkk1234 commented 5 years ago

Ok , Thank you very much ,I will try this step .And I have another question , when I doing the test step , I can't understand why the label data is still placed in the data folder?Thank you for your instantly reply!

Tramac commented 5 years ago

All data in the data folder is used for training. Because the data came from a competition website, where had no public test data, data used in test stage was also from data folder, aiming to test the code can run without bugs. In theory, the test data should come from a new data source, which did't appear in the training set. Thank you for your question, you are right!