Abstract
The present work concerns the analysis of how demosaicing artifacts affect image quality and proposes a novel no-reference metric for their quantification. This metric that fits the psycho-visual data obtained by an experiment analyzes the perceived distortions produced by demosaicing algorithms. The demosaicing operation consists of a combination of color interpolation (CI) and anti-aliasing (AA) algorithms and converts a raw image acquired with a single sensor array, overlaid with a color filter array, into a full-color image. The most prominent artifact generated by demosaicing algorithms is called zipper. The zipper artifact is characterized by segments (zips) with an On–Off pattern. We perform psycho-visual experiments on a dataset of images that covers nine different degrees of distortions, obtained using three CI algorithms combined with two AA algorithms. We then propose our no-reference metric based on measures of blurriness, chromatic and achromatic distortions to fit the psycho-visual data. With this metric demosaicing algorithms could be evaluated and compared.
Highlights
Image quality is difficult to assess correctly for a number of reasons [1]
Due to its perceptual nature, image quality should be evaluated through a subjective assessment, and quality metrics should be designed to fit quality judgments collected by psycho-visual experiments
For each subject j-th and distorted image i-th we evaluated the perceptual distance between original and distorted images in terms of difference of assigned scores (Difference Score, DS): DSij = Soij − Sdij where Sdij represents the score assigned by the j-th subject to the i-th distorted image, while Soij the score of the reference image corresponding to the i-th distorted image; j = 1, . . . , J, denotes subjects belonging to the group of J individuals, and i = 1, . . . , S × T denotes the distorted image, with S number of reference images, and T number of algorithms to be evaluated
Summary
This makes it hard to measure it in a standardized way and allows for personal preferences. It can vary widely between different application domains. Due to its perceptual nature, image quality should be evaluated through a subjective assessment, and quality metrics should be designed to fit quality judgments collected by psycho-visual experiments. To validate automated approaches, psycho-visual scaling studies are insurmountable for image quality research [2,3]. We are interested in developing a pool of no-reference (NR) metrics to automatically assess the performance of the algorithms composing the image generation pipeline of digital cameras. A general three-step approach to design and develop these types of metrics has been given by Bartleson [4]: 1. Identification of perceptual dimensions (attributes) of quality
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
