Analysis and improvement on statistical naturalness measure

Nur Halilah Binti Ismail, Soong Der Chen

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Although Statistical Naturalness Measure (SNM) can rate the naturalness of image’s contrast relatively well, but it suffers from the problem of inconsistent rating across different spatial resolution; ratings for images with identical content and contrast level but different spatial resolution might be significantly different. The statistical model SNM was developed upon sample images with certain spatial resolution and the contrast was defined using a specific block size. This paper suggests that the problem is due to the computation of contrast using fixed block size regardless of the spatial resolution of the input image. It is proposed to rescale the block size based on the spatial resolution of input image to keep the ratio of block size to image size similar to those of the sample images used in developing the statistical model of SNM. The proposed method is tested and proven statistically to be effective in reducing the inconsistency in the ratings of images across different spatial resolution such there is no significant difference among them.

Original languageEnglish
Pages (from-to)368-372
Number of pages5
JournalJournal of Theoretical and Applied Information Technology
Volume83
Issue number3
Publication statusPublished - 31 Jan 2016

    Fingerprint

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this