Image super-resolution using matrix valued operations
Matrix valued image super-resolution is a novel topic of research that uses the spatial dependencies when finding the mapping between low and high resolution images. This project implements this paper, and provides a proof-of-concept.
The data should be downloaded from this link, and moved to the Data folder.
We train the model using trainingPatchExtraction.m
and the test scripts are provided in testingPatchExtraction.m
.
The scripts psnrCalc.m
and ssim_wl.m
are provided for model evaluation.
The paper is provided at Single Image SISR