What I meant was scanning the DOM of a page for text equivalents and showing the actual parts of the page, as well as hiding the irrelevant ones.