This wouldn't stop fraudsters using stolen ID cards, since people are surprisingly bad at checking photos. In a 1995 study University of Westminster researchers issued 44 students with four picture ID cards each. The cards carried a variety of photos, including a simulated "old" photo of the holder (with a different hairstyle, or addition or removal or glasses or a beard, as one might have on a five-year-old ID card) and one chosen to look like the holder from a hundred random photos of different people (as a criminal would choose from a stack of stolen ID cards). Experienced supermarket cashiers couldn't reliably tell whether students were using an "old" card or a "stolen" card. This experiment was done under optimum conditions, with experienced staff, plenty of time, and no threat of embarrassment if a card was rejected; shop assistants' real-life performance would be worse. As a result of this study, no UK credit card company now puts the holder's photo on credit cards.

R. Kemp, N. Towell, G. Pike, "When seeing should not be believing: Photographs, credit cards and fraud" Applied Cognitive Psychology Vol 11(3) (1997) pp 211-222 http://www3.interscience.wiley.com/cgi-bin/abstract/11942/ABSTRACT