Will Google extract text from images?
Summary:
Today there still does not exists a system based on Optical Character Recognition (OCR) for crawling images from the web. But that would be a very useful thing so probably Google will work in that direction. It’s a lot of work to do to make such a system work so it could take some time to implement but one day it will definitely exists.
Matt's answer:
That would be a really big undertaking. I think it would be fun. You know, we’ve joked around the pool table about “Wouldn’t it be great if we crawled the web, found all the images, and ran OCR on all the images on the web?” That really would be a lot of work. I think it would be a fun idea, but I’m not sure that you should count on that in the short term from Google.
by Matt Cutts - Google's Head of Search Quality Team