Empirical web accessibility evaluation for blind web users

Abstract

The project intends to produce a novel, accurate and holistic system that assesses the accessibility of an entire website based on a sample of webpages. The project also strives to bring statistical rigour to the sampling and clustering of webpages. By orchestrating pre-existing evaluative tools, alongside developer testing, we can produce a combined accessibility assessment. A major benefit to the accessibility field is the effective, large-scale evaluation of large websites. The approach strives to improve automated testing coverage by combining evaluative tools, and minimise inaccuracy through effective clustering and sampling.