An AI Solution for Web Accessibility and Images Classification
This article details the research on web accessibility conducted at Capgemini's SogetiLabs. We introduce our project aimed at developing an automatic accessibility audit tool for website images. Our AI solution for web accessibility focuses on distinguishing between informative and decorative images in line with RGAA (Referenciel General d'Amelioration de I'Acessibilite) recommendations and then generating alternative text for informative images. To achieve this, we have established a comprehensive processing workflow. Additionally, we present initial experiments in image classification using Convolutional Neural Networks (CNNs) and YOLO's (You Only Look Once) model.
Authors
- Laura Noreskal
- Guillaume Feuilloley
- Simon-Pierre Charbel
Institutions
- Capgemini (France) (FR)
Publication Details
- Published
- 2024-10-14
- DOI
- https://doi.org/10.1109/ipta62886.2024.10755614
- Citations
- 1
- Primary Topic
- Digital Accessibility for Disabilities
- Type
- article
- Field-Weighted Citation Impact
- 2.19