System of recommendation and automatic correction of web accessibility using artificial intelligence

Abstract

This work presents the development of a Web system that allows the identification, evaluation and automatic correction of Web accessibility barriers associated with multimedia elements. The analysis carried out by the tool takes into account the conformance level A of WCAG 2.0 standard. Web system takes as input the URL of the Web page to be evaluated to connect and send the evaluation request to three Web Content Analysis APIs: OAW, Tenon, and Achecker. The results of the evaluations are sent to the artificial intelligence services of Google, to obtain an adequate description of multimedia items that are not correctly labeled. Finally, the system proposes a holistic automatic correction of the website source code and allows the result to be exported. To test the effectiveness of the tool were evaluated 54 websites, in different sectors such as government, education, finance, etc. The results show an average increase of 2.57% in web accessibility conformance level, reaching a maximum increase of 24%.