Recent approaches on no- reference image quality assessment for contrast distortion images with multiscale geometric analysis transforms: A survey

Ismail T. Ahmed, Soong Der Chen, Baraa Tareq Hammad

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

The study of Image Quality Assessment (IQA) in digital image and video processing is challenging due to the existences of numerous types of distortions such as blur, noise, blocking, contrast change, etc. Nevertheless, it is interesting to devise a metric system in order to determine the quality of an image quantitatively. Currently, most of the existing No Reference(NR)-IQA metrics focus on the quality evaluation of distorted images due to compression, noise and blurring. The related work performed in the area of NR-IQA for Contrast Distortion Images (CDI) is quite limited unfortunately. Also, most of the existing NR-IQA metrics are designed in spatial domain and very little of them are devised based on Multiscale Geometric Analysis (MGA) Transforms. Therefore, in this paper, NR-IQA metrics are classified into two groups, i.e. NR-IQA Metrics for general purpose and NR-IQA Metrics for CDI. Due to the fact that our main focus is contrast distortion, NR IQA metrics have been overviewed in both spatial and transform domains. We classify the transform domain into traditional transform and MGA transform then focusing on MGA Transforms. Subsequently, the MGA transform which is suitable for the design of NR-IQA metric used to predict the quality of CDI is proposed. The presented survey will to keep up-to-date the researchers in the field of image quality assessment especially for CDI. Also, this survey provides an outlook for future work using many combinations among MGA Transforms to access to new IQA metric for CDI.

Original languageEnglish
Pages (from-to)561-569
Number of pages9
JournalJournal of Theoretical and Applied Information Technology
Volume95
Issue number3
Publication statusPublished - 15 Feb 2017

    Fingerprint

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this