I disagree. Sure, "beauty" isn't a very robustly-defined term, but the headline is reasonable. The most important point is that people's views of themselves are objectively wrong, because the identify modified images of themselves. Conversely, people can reliably pick the correct image of others. Finally, this depictions differ in that the self-image has characteristics society deems attractive, for our purposes, "beauty". I don't think beauty is quite the best term, but I wouldn't call it wrong.
Also, I think your proposed study would undermine the whole point of this study, since someone who didn't realize they were looking at themselves wouldn't have the cognitive impetus to distort reality.
edit: I just realized I may have been missing part of your argument. Would you be more satisfied if the title were "You are less beautiful than you _say_ you are"? This study itself doesn't offer any real evidence that people actually believe they look like the picture they say they look like. However, the other cognitive studies linked in this article, including some other similar studies I've read about, make me think it's extremely likely.