At the moment WCAG validation is a legal requirement, while PDF/UA is not (unfortunately). In this talk we make an attempt to perform a gap analysis between these standards, taking PDF/UA-1 and Matterhorn 1.1 protocol as a basis and analyzing the differences.
In particular, we suggest parts of PDF/UA-1, which are not a part of WCAG (eg PDF/UA metadata identification) and, on the opposite, parts of WCAG, which are not a part of PDF/UA (eg, detailed requirements for the text contrast).
Finally, we describe how veraPDF approaches WCAG validation. This includes interfaces and algorithms for contrast checks and semantic appropriateness checks, based on table recognition heuristics.