Simplifying and Improving Effectiveness of Image Description for Accessibility using Sample Cues
Universal design (UD) of information communication technology (ICT) is a basic principle that confirms the accessibility of ICT products and services to avoid discrimination and to support the equal ability participation in the society to enhance democracy. An image is a non-textual Web content which demands an equivalent image description to be accessible and usable for the users who have different abilities. Previous research suggests that the significant number of images exist on the Web which do not contain accessible and usable text descriptions. Literature also suggests that describing an image is a complex task for the Web workers and normal users. To the knowledge of this study, there is almost no research available in image accessibility which incorporates different types of images. This research’s aim is to fulfill the existing gap of not having the proper solution to help Web users in writing usable image descriptions. This study investigated the effectiveness of providing sample cues in making image description comply with the set of guidelines suggested by National Center for Accessible Media (NCAM) through an online experiment. Three different types of images: graph, map, and general image are provided in three different steps with no cue, random cue, and similar cue respectively. Text descriptions, Likert-type ratings on the compliment to NCAM guidelines, and time data are collected from two separate groups—participants and the judgement group. The experimental result is analyzed using several statistical tests. The results from this research suggest that providing the similar sample cue is effective writing an accessible image description in case of all the three types of images. Similar sample cues make it easier to write a description which follows the NCAM guidelines. The random sample cues are better than not having any cue. However, there is no significant difference in time it takes to write image descriptions in the three cases.