Appendix for the paper: A Reproducible Methodology for Evaluating WCAG-Aligned Alt Text Generation by General-Purpose Generative AI

This record provides the supplementary materials for our paper evaluating how mainstream, general-purpose generative AI models generate WCAG-aligned alternative text under a standardized protocol. It includes the 37-image gold-standard reference set, the study webpage (HTML+images with intentionally empty alt=""), the print-friendly PDF version, model outputs, and ROUGE scoring spreadsheets/figures used for analysis.