jinglongdu/brain-mri-axial-t1-image-srcnn-

Using Three Layers SRCNN (Proposed by Chao Dong at all) to Brain MRI Axial T1 Image. I programmed with Python tensorflow. Training the model in Google colab.

Brain MRI Axial T1 Image SRCNN

Using Three Layers SRCNN (Proposed by Chao Dong at all) to Brain MRI Axial T1 Image. I programmed with Python tensorflow. Training the model in Google colab. Reference github project "tegg89/SRCNN-Tensorflow" and do some improvement.

How to use the code?

1. First, you need to execute utils.py, model.py and main.py code.
2. Second, you can adjust "is_train" parameter in "main.py". When you want to training model, you should set is_train in "True". When you want to testing model, you should set is_train in "False".
3. Third, after you test the image which you want by the trained model. You can see your SRCNN image in "sample ---> my SRCNN.png file".
4. Finally, you can caculate the PSNR performance by the .m file "cabips.m".

Image size and label size setting

Initial, we separate 255x255 pixel image into N 33x33 pixel sub-image. After convolution caculate, 21x21 piel sub-image will stack back to 252x252 SRCNN image.

Testing data

Testing data will be update before 2019/6.