search

arnab39 / FewShot_GAN-Unet3D

Tensorflow implementation of our paper: Few-shot 3D Multi-modal Medical Image Segmentation using Generative Adversarial Learning
MIT License
377 stars 82 forks source link

Predicted Output Groundtruth Images is 0 #10

Open chengaox opened 5 years ago

chengaox commented 5 years ago

I tried to train and test unet3D with "epoch", 90 "number_train_images", 5 but the output of CSF,GM,WM are all 0 (I have already checked the input image and label, Label unique:[0 10 150 250])

training validation testing

I also train and test proposed_model and the output image is as follows GAN_output

Ask for help！Thanks a lot！

oomiracle commented 5 years ago

Hi ! I have encoutered the same questions with you and I havn't fix it. I have noticed that the two UndefinedMetricWarning ，maybe the problem is in this place.

@ @kurodasho

TaoZhong11 commented 4 years ago

UndefinedMetricWarning

Hi, have you fixed it? I have encoutered the same questions.

ren19950405 commented 4 years ago

Hi, have you fixed it? I have encoutered the same questions.

ren19950405 commented 4 years ago

Hi ! I have encoutered the same questions with you and I havn't fix it. I have noticed that the two UndefinedMetricWarning ，maybe the problem is in this place.

@ @kurodasho Hi ! I have encoutered the same questions with you and I havn't fix it.Hava you fixed it?

hkthirano commented 4 years ago

Hi, have you fixed it? I have encoutered the same questions.

xiaohong666 commented 3 years ago

Hi, have you fixed it? I have encoutered the same questions.