GAN method to help covert lower resolution STEM images generated by convolution methods to higher resolution STEM images generated by Multi-Slice methods. We published the results in this Arxiv paper, you can check more details from there, Exploring Generative Adversarial Networks for Image-to-Image Translation in STEM Simulation

Calculate_NRMSE.py helps calculate the normalized root mean square error of two images generated by different methods.

To run the command you need first download the data file from Google Drive and name the folder Data. Then you can run the program with this command.

python Calculate_NRMSE.py

You need Python 3 and scikit-image,numpy and matplotlib to run the program .

To change to a different fold of images just comment currently used folder path then uncomment the previously commented folder path.

Based on Keras-GAN we have obtained the initial results which shows very promising results.