Experimental Analysis on Perceptual Image Quality Outcomes of Single Image Super Resolution Frameworks

Numerous advancements in image super resolution in recent years have explored different deep learning approaches and frameworks based on Convolution Neural Network (CNN) to achieve enhanced perceptual results. In this article, detailed experimental and performance analysis on perpetual image quality results of Single Image Super Resolutions frameworks through multiple CNN architectures has carried out according to the reconstruction efficiency, perceptual and reconstruction accuracy. Initially background of the image super resolution is analyzed. Next, review of the image super resolution architecture based on deep learning architecture with different optimization approaches has been analyzed with respect to the benchmark dataset. Additionally different image quality assessment method has been discussed to highlight performance of the existing image super resolution frameworks. Finally these experimental findings provides the image results of basis of measures like Peak signal to noise ratio, structural similarity index and visual information fidelity to present the performance of the model.