Full reference quality assessment of downsized images

Resizing image processing tools are in vogue nowadays due to various practical reasons. The images are resized into different resolutions and scales. The quality of the image may get affected by resizing the original image. In this paper, quality of downsized images is evaluated. The proposed metric is full reference and based on edge detection, sharpness and contrast of images. Moreover, due to the limited availability of databases with perceived image quality score of resized images, the proposed metric is compared with recent no reference image quality assessment techniques (as they are size independent) BRISQUE and NIQE. Experimental results show that the proposed metric is consistent with the perceived quality and outperform the aforesaid no reference techniques.