What is Image Quality Assessment (IQA)?

Image Quality Assessment (IQA) algorithms take an arbitrary image as input and output a quality score as output. There are three types of IQAs:

• Full-Reference IQA: Here you have a ‘clean’ reference (non-distorted) image to measure the quality of your distorted image. This measure may be used in assessing the quality of an image compression algorithm where we have access to both the original image and its compressed version.

• Reduced-Reference IQA: Here you don’t have a reference image, but an image having some selective information about it (e.g. watermarked image) to compare and measure the quality of distorted image.

• Objective Blind or No-Reference IQA: The only input the algorithm gets is the image whose quality you want to measure. This is thus called, No-Reference or Objective-Blind.

A package developed for BRISQUE implementation

FULL CODE @ https://github.com/rehanguha/brisque

References

This blog is entirely referenced from the below two links. This motive for this blog is to showcase the brisque package which I developed.

• https://learnopencv.com/image-quality-assessment-brisque/
• https://www.mathworks.com/help/images/ref/brisque.html