For example, it has Google’s image recognition algorithm can now describe an entire scene being captured in an image like “A person riding a motorcycle on a dirt road.” Check out this horrifyingly awesome power here.
If it can do that, it could easily detect and index the words found inside an image. The company has simply decided not to though.
To be clear, it seems as though Google has decided to not index text-based content found in non-text-based media because the payoff would not be worth the extra effort (computing power).
NOTE: You’ll notice images and videos both have their own separate algorithms within their respective parts of Universal Search Results.
To learn more about Rich Media (how to properly and improperly implement, as well as how to find it on your site), signup for SEO Pro here