Non-Shift Edge Based Ratio (NSER): An Image Quality Assessment Metric Based on Early Vision Features

Abstract

How to evaluate the image perceptual quality is a fundamental problem in image and video processing, and various methods have been proposed for image quality assessment (IQA). This letter presents a novel IQA metric, which is based on the image primitive features produced in the earliest processing stage of human visual system. The procedures involved in the proposed method include computing the response of classical receptive fields, zero-crossing detection, and non-shift edge based ratio (NSER) calculation. The proposed IQA metric is very simple but very effective. The experimental results on benchmark databases show that the NSER index has very high consistency with the psychological evaluation, performing much better than most state-of-the-art IQA metrics.