Gradient Ascent Subjective Multimedia Quality Testing

<p/> <p>Subjective testing is the most direct means of assessing multimedia quality as experienced by users. When multiple dimensions must be evaluated, these tests can become slow and costly. We present gradient ascent subjective testing (GAST) as an efficient way to locate optimizing s...

Full description

Bibliographic Details
Main Authors: Voran Stephen, Catellier Andrew
Format: Article
Language:English
Published: SpringerOpen 2011-01-01
Series:EURASIP Journal on Image and Video Processing
Online Access:http://jivp.eurasipjournals.com/content/2011/472185
id doaj-af50e6c7bf4346bfbe3a092af6d975e0
record_format Article
spelling doaj-af50e6c7bf4346bfbe3a092af6d975e02020-11-24T22:38:50ZengSpringerOpenEURASIP Journal on Image and Video Processing1687-51761687-52812011-01-0120111472185Gradient Ascent Subjective Multimedia Quality TestingVoran StephenCatellier Andrew<p/> <p>Subjective testing is the most direct means of assessing multimedia quality as experienced by users. When multiple dimensions must be evaluated, these tests can become slow and costly. We present gradient ascent subjective testing (GAST) as an efficient way to locate optimizing sets of coding or transmission parameter values. GAST combines gradient ascent optimization techniques with subjective test trials. As a proof-of-concept, we used GAST to search a two-dimensional parameter space for the known region of maximal audio quality, using paired-comparison listening trials. That region was located accurately and much more efficiently than use of an exhaustive search. We also used GAST to search a two-dimensional quantizer design space for a point of maximal image quality, using side-by-side paired-comparison trials. The point of maximal image quality was efficiently located, and the corresponding quantizer shape and deadzone agree closely with the quantizer specifications for JPEG 2000, Part 1.</p>http://jivp.eurasipjournals.com/content/2011/472185
collection DOAJ
language English
format Article
sources DOAJ
author Voran Stephen
Catellier Andrew
spellingShingle Voran Stephen
Catellier Andrew
Gradient Ascent Subjective Multimedia Quality Testing
EURASIP Journal on Image and Video Processing
author_facet Voran Stephen
Catellier Andrew
author_sort Voran Stephen
title Gradient Ascent Subjective Multimedia Quality Testing
title_short Gradient Ascent Subjective Multimedia Quality Testing
title_full Gradient Ascent Subjective Multimedia Quality Testing
title_fullStr Gradient Ascent Subjective Multimedia Quality Testing
title_full_unstemmed Gradient Ascent Subjective Multimedia Quality Testing
title_sort gradient ascent subjective multimedia quality testing
publisher SpringerOpen
series EURASIP Journal on Image and Video Processing
issn 1687-5176
1687-5281
publishDate 2011-01-01
description <p/> <p>Subjective testing is the most direct means of assessing multimedia quality as experienced by users. When multiple dimensions must be evaluated, these tests can become slow and costly. We present gradient ascent subjective testing (GAST) as an efficient way to locate optimizing sets of coding or transmission parameter values. GAST combines gradient ascent optimization techniques with subjective test trials. As a proof-of-concept, we used GAST to search a two-dimensional parameter space for the known region of maximal audio quality, using paired-comparison listening trials. That region was located accurately and much more efficiently than use of an exhaustive search. We also used GAST to search a two-dimensional quantizer design space for a point of maximal image quality, using side-by-side paired-comparison trials. The point of maximal image quality was efficiently located, and the corresponding quantizer shape and deadzone agree closely with the quantizer specifications for JPEG 2000, Part 1.</p>
url http://jivp.eurasipjournals.com/content/2011/472185
work_keys_str_mv AT voranstephen gradientascentsubjectivemultimediaqualitytesting
AT catellierandrew gradientascentsubjectivemultimediaqualitytesting
_version_ 1725711660218318848