I was surprised too. I don't see why Google would spent cpu time on this, and I can't think of a way one would be able to validate that claim.
CPU time is cheap, and it could be useful ranking information. Compared to an H1 tag, a graphical header is harder to create, so it shows a certain level of investment.
If the words in the graphic aren't repeated elsewhere in the page, then the page author may be naive in the ways of SEO. But that could be good, because the savvy SEO people are all trying to pull one over on the GoogleBot.
Further: Google has an intense interest in OCR, adopting the open-source 'Tesseract' project, spending millions on scanning first catalogs then books and journals, and most recently announcing they are OCRing bitmaps in PDFs:
http://googleblog.blogspot.com/2008/10/picture-of-thousand-w...
Finally: any reasoning based on Google being miserly with cycles is going to be wrong. (They invest a lot in efficiency, yes, but that's so that they can spend cycles freely to collect data.)
It's possible they ran an experiment and found text in embedded/header images was no better than inline text. (I doubt they would find such a thing, because generally indicators of sustained effort -- careful design, site longevity, good writing -- are also indicators of site quality.) But there's zero chance the CPU cost or scale deterred Google from testing the idea.