Ever since the early days of search engine optimisation the whole process in every case was to remove image navigation and replace it where possible with text links or at least Alt tag images. This has been the only way to inform the search engine crawlers what the link is and what to expect on the end of it.
This could all change. In June 2007 Google lodged a patent to a complicated search technique to identify and interpret text within images. This is effectively making it possible for the search giant to read and understand image text.
It is an interesting concept but on reflection if you have ever used a text image converter, you know the kind where you place a document in your scanner and it processes the scanned image into a word document, all I can say is I hope they have been working on accuracy.