The output image is strange : train.py or predict.py problem?

milesial / Pytorch-UNet

PyTorch implementation of the U-Net for image semantic segmentation with high quality images

GNU General Public License v3.0

9.17k stars 2.5k forks source link

The output image is strange : train.py or predict.py problem? #407

Open t19b205d opened 1 year ago

t19b205d commented 1 year ago

Hello! My English isn't so good so feel free to ask me if there is anything unclear. I trained 2classes. using dataset is CT image of lung tumor.

The mask image below is wandb/run-20221111*****/files/media/images/pred***.png background:0 ,tumor:1 are classified pred_22221_6b9f942956941162fa97

But output mask images is strange. output05003

I want to output like wandb/run-20221111_*****/files/media/images.

milesial commented 1 year ago

Hi, can you try the latest master version?

mahendran-narayanan commented 1 year ago

It still results the same

milesial commented 1 year ago

Can you share the pretrained model along with the input image that results in these masks?