Super resolution using CNN based on paper titled "Perceptual Losses for Real-Time Style Transfer and Super-Resolution" (https://arxiv.org/abs/1603.08155)
Note: Dataset used is 10000 images from ImageNet dataset and was not uploaded due to the size. The training dataset consists of 10000 pairs of low resolution images (72x72x3) and high resolution images (288x288x3).
Here is comparison between upscaling using GIMP cubic transformation and super resolution
