Combining semantic tools for automatic evaluation of alternative texts

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

3 Scopus citations

Abstract

Automated accessibility evaluation will have an increasingly important role in the near future due to legal requirements to monitor the accessibility of the websites of public bodies in the European Union. However, automated evaluation tools are still limited in the scope of the conformance testing they are able to perform. Future solutions resorting to AI based techniques might help in addressing some of these limitations. In this paper we present an updated version of an algorithm that classifies the quality, from an accessibility perspective, of alternative texts for images in web pages. We evaluated the proposed algorithm on a set of 149 images and their corresponding alternative texts and found improved performance compared to the initial version of the algorithm.

Original languageEnglish (US)
Title of host publicationProceedings of the 16th Web For All 2019 Personalization - Personalizing the Web, W4A 2019
PublisherAssociation for Computing Machinery, Inc
ISBN (Electronic)9781450367165
DOIs
StatePublished - May 13 2019
Event16th International Web for All Conference: Personalization - Personalizing the Web, W4A 2019 - San Francisco, United States
Duration: May 13 2019May 15 2019

Publication series

NameProceedings of the 16th Web For All 2019 Personalization - Personalizing the Web, W4A 2019

Conference

Conference16th International Web for All Conference: Personalization - Personalizing the Web, W4A 2019
Country/TerritoryUnited States
CitySan Francisco
Period05/13/1905/15/19

Keywords

  • Algorithm
  • Alternative text
  • Automatic accessibility evaluation
  • Images
  • Semantic analysis

ASJC Scopus subject areas

  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'Combining semantic tools for automatic evaluation of alternative texts'. Together they form a unique fingerprint.

Cite this