Publication detail

Video Quality Assessment in Experimental Retinal Imaging

LIBERDOVÁ, I. KOLÁŘ, R.

Original Title

Video Quality Assessment in Experimental Retinal Imaging

English Title

Video Quality Assessment in Experimental Retinal Imaging

Type

conference paper

Language

en

Original Abstract

This paper is focused on video quality assessment of experimental retinal video image sequences with two different approaches (automatic and expert analysis). Methods utilizing both video quality evaluations are presented and described within the paper. Evaluations consider the most frequent retinal imaging artefacts (contrast distortion, noise and blurriness). Automatic evaluating algorithm estimates the image quality based on three different parameters characterizing the image over time (signal to noise ratio - SNR, blurriness quality evaluation - BQE and contrast quality evaluation - CQE). Four different experts performed the evaluation with their subjective image perceiving and assess the degree of noise, blurriness and contrast in the images over time. The reliability of proposed algorithm for automatic retinal image video quality assessment seems to be correct for all three parameters. Medians of SNR as well as the medians of CQE are decreasing with decreasing image quality evaluated with 4 experts, and medians of BQE are increasing.

English abstract

This paper is focused on video quality assessment of experimental retinal video image sequences with two different approaches (automatic and expert analysis). Methods utilizing both video quality evaluations are presented and described within the paper. Evaluations consider the most frequent retinal imaging artefacts (contrast distortion, noise and blurriness). Automatic evaluating algorithm estimates the image quality based on three different parameters characterizing the image over time (signal to noise ratio - SNR, blurriness quality evaluation - BQE and contrast quality evaluation - CQE). Four different experts performed the evaluation with their subjective image perceiving and assess the degree of noise, blurriness and contrast in the images over time. The reliability of proposed algorithm for automatic retinal image video quality assessment seems to be correct for all three parameters. Medians of SNR as well as the medians of CQE are decreasing with decreasing image quality evaluated with 4 experts, and medians of BQE are increasing.

Keywords

retinal imaging, video quality assessment, cross correlation, noise

Released

29.08.2016

Publisher

Brno University of Technology

Location

Blansko

ISBN

978-80-214-5389-0

Book

Sborník příspěvků studentské konference Blansko

Edition

1

Edition number

1

Pages from

51

Pages to

54

Pages count

4

BibTex


@inproceedings{BUT127623,
  author="Ivana {Labounková} and Radim {Kolář}",
  title="Video Quality Assessment in Experimental Retinal Imaging",
  annote="This paper is focused on video quality assessment of experimental retinal video image sequences with two different approaches (automatic and expert analysis). Methods utilizing both video quality evaluations are presented and described within the paper. Evaluations consider the most frequent retinal imaging artefacts (contrast distortion, noise and blurriness). Automatic evaluating algorithm estimates the image quality based on three different parameters characterizing the image over time (signal to noise ratio - SNR, blurriness quality evaluation - BQE and contrast quality evaluation - CQE). Four different experts performed the evaluation with their subjective image perceiving and assess the degree of noise, blurriness and contrast in the images over time. The reliability of proposed algorithm for automatic retinal image video quality assessment seems to be correct for all three parameters. Medians of SNR as well as the medians of CQE are decreasing with decreasing image quality evaluated with 4 experts, and medians of BQE are increasing.",
  address="Brno University of Technology",
  booktitle="Sborník příspěvků studentské konference Blansko",
  chapter="127623",
  edition="1",
  howpublished="electronic, physical medium",
  institution="Brno University of Technology",
  year="2016",
  month="august",
  pages="51--54",
  publisher="Brno University of Technology",
  type="conference paper"
}