Toward Accurate Quality Estimation of Screen Content Pictures With Very Sparse Reference Information

Research output: Journal Publications and Reviews (RGC: 21, 22, 62)21_Publication in refereed journal

1 Scopus Citations
View graph of relations

Author(s)

Related Research Unit(s)

Detail(s)

Original languageEnglish
Article number8672903
Pages (from-to)2251-2261
Journal / PublicationIEEE Transactions on Industrial Electronics
Volume67
Issue number3
Online published22 Mar 2019
Publication statusPublished - Mar 2020

Abstract

The screen content (SC) pictures, such as webpages, serve as a visible and convenient medium to well-represent the Internet information, and therefore, the visual quality of SC pictures is highly significant and has attained a growing amount of attention. Accurate quality evaluation of SC pictures not only provides the fidelity of the conveyed information, but also contributes to the improvement of the user experience. In practical applications, a reliable estimation of SC pictures plays a considerably critical role for the optimization of the processing systems as the guidance. Based on these motivations, this paper proposes a novel method for precisely assessing the quality of SC pictures using very sparse reference information. Specifically, the proposed quality method separately extracts the macroscopic and microscopic structures, followed by comparing the differences of macroscopic and microscopic features between a pristine SC picture and its corrupted version to infer the overall quality score. By studying the feature histogram for dimensionality reduction, the proposed method merely requires two features as the reference information that can be exactly embedded in the file header with very few bits. Experiments manifest the superiority of our algorithm as compared with state-of-the-art relevant quality metrics when applied to the visual quality evaluation of SC pictures.

Research Area(s)

  • Anisotropic magnetoresistance, Visualization, Microscopy, Feature extraction, Distortion measurement, Estimation, Macroscopic/microscopic structure, quality estimation, screen content (SC) picture, sparse reference, FREE-ENERGY PRINCIPLE, IMAGE, BRAIN, MODEL

Citation Format(s)