Automatic color improvement of web pages with time limited operators

  • Authors:
  • Sébastien Aupetit;Alina Mereuţă;Mohamed Slimane

  • Affiliations:
  • Laboratoire Informatique (EA6300), France, Université Franţois Rabelais Tours, Tours, France;Laboratoire Informatique (EA6300), France, Université Franţois Rabelais Tours, Tours, France;Laboratoire Informatique (EA6300), France, Université Franţois Rabelais Tours, Tours, France

  • Venue:
  • ICCHP'12 Proceedings of the 13th international conference on Computers Helping People with Special Needs - Volume Part I
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Accessibility is unfortunately not among the main concern when developing web sites. Webmasters create mostly involuntarily numerous obstacles for people with visual impairments. That's why it becomes fundamental to identify the existing barriers and to propose solutions in order to at least diminish their impact to the user. Accessibility guidelines, as WCAG 2.0, indicate that a minimum difference of brightness, tonality and contrast is necessary to reach a minimum level of accessibility. In numerous cases, web designers ignore or just limit their choices to a low level of accessibility. For an user needing a higher level of accessibility than the one offered by the web page, the access to information may be difficult. In this context, we propose to transform the colors of web pages according to user's needs with the help of a client-side HTTP proxy. The requirements for the colors can be expressed as a fitness function. In order to recolor the page to increase accessibility, it's enough to minimize the fitness function. Trying to find a minimum can be a time consuming task not appropriate for real time recoloring. Finding a minimum can be considered as a search with varying time limits. In this article, our objective is to compare different search methods and their performance under time limit: the search can be interrupted at any time. The studied methods are a random search, different types of pseudo gradient descend and an adaptation of the API metaheuristic. Finally, the different methods are compared.