Photographs vary widely in technical quality. Pictures provided by amateur photographers for identification purposes may be color or black-and-white, printed on glossy or matte paper, damaged by sunlight or otherwise stained, or improperly exposed. The use of such photographs in an image-processing system mandates normalization of brightness and contrast in order that the resulting copies, produced on a laser printer, can be used to recognize the subject of the photograph. In a batch environment, it is also necessary that photo processing be completed in a short period of time. This paper describes a sub-second photograph-processing algorithm.