Reduced-Reference Image Quality Assessment
Objective image quality assessment models typically require the access to a reference image that is assumed to have perfect quality. In practice, such full-reference (FR) methods may not be applicable because the reference image is often not available. On the other hand, no-reference (NR) or “blind” image quality assessment is an extremely difficult task (especially general-purpose metrics that are applicable to a wide variety of image distortion types). Reduced-reference (RR) image quality metrics provide a solution that lies between FR and NR models. They are designed to predict the perceptual quality of distorted images with only partial information about the reference images.
This page provides an implementation of the RR method proposed in the following paper:
Z. Wang and E. P. Simoncelli, "Reduced-reference image quality assessment using a wavelet-domain natural image statistic model," Human Vision and Electronic Imaging X, Proc. SPIE, vol. 5666, San Jose, CA, Jan. 2005.
You can download the software for free, change it as you like and use it anywhere, but please refer to its original source.