search

LeeJunHyun / Image_Segmentation

Pytorch implementation of U-Net, R2U-Net, Attention U-Net, and Attention R2U-Net.
2.68k stars 595 forks source link

About the results #30

Closed loaysh2010 closed 5 years ago

loaysh2010 commented 5 years ago

Dear @LeeJunHyun , Thank you very much for your time and your help. I want to ask you about the results image. Firstly my dataset used in model look like TCGA-18-5592-01Z-00-DX1-exp 0 --- TCGA-18-5592-01Z-00-DX1-exp 0 for input and GT respectivly. the getting result look like U_Net_valid_1_GT Named : U_Net_valid_1_GT U_Net_valid_1_image Named : U_Net_valid_1_image U_Net_valid_1_SR Named : U_Net_valid_1_SR Can you please explain what's this images refer to ?!!!,, Where is the resulted segmented image?

LeeJunHyun commented 5 years ago

Dear @loaysh2010 , Thank you for your interest about my repo. U_Net_valid_1_GT also seems like grayscale. Is this right?

loaysh2010 commented 5 years ago

Dear @LeeJunHyun Thank you for your quick response and concern. Here is my Dataset samples for train and validation and test (input image and GT image ) respectively. TCGA-18-5592-01Z-00-DX1-exp 0 ==== TCGA-18-5592-01Z-00-DX1-exp 0 TCGA-18-5592-01Z-00-DX1-exp 16 ==== TCGA-18-5592-01Z-00-DX1-exp 16 TCGA-18-5592-01Z-00-DX1-exp 28 ==== TCGA-18-5592-01Z-00-DX1-exp 28

In the results, except the input image and that patch size is 5 all output is ambiguous for me. I think there is something in the code I have to adapt. all my images (both input and GT) are [1,3,256,256]. What do you think I did wrong?

loaysh2010 commented 5 years ago

@LeeJunHyun I solved the problem here. There is an error in the GT path given to the model. after correcting GT path the model gave excellent results. Thank you for your time and you can close the issue