Towards Subjective Consistency: An Effective Objective Quality Assessment Algorithm for Binary Image

Abstract

Image quality assessment plays an important role in various image processing applications. One of the challenges to objectively assess image quality is how to design an effective scheme to achieve high consistency with the classic subjective image assessment criterion, Mean Opinion Score (MOS). This work presents a novel objective assessment algorithm for binary images by considering three factors which have great influences on visual quality of binary images, i.e., structural change caused by noise point, isolated noise points, and gathering noise points. Experimental results show that our algorithm can achieve effective objective assessment results with higher consistency with the MOS criterion.