A procedure to estimate the distance between two facial features on a standard mug-shot photograph based on analysis of non-standard photograph



Journal Title

Journal ISSN

Volume Title



The problem considered in this thesis, grew out of attempts to apply a pattern recognition approach to the retrieval of 'look-alikes' from a mugfile to forgery cases. In forgery cases, we usually have a photograph taken at the time the forged check was cashed of the person cashing it. Unfortunately, photographs of forgery suspects are not all taken in the straight-on position (front bust view), therefore horizontal and vertical distances on these non-standard photographs do not correspond to distances on a standard mug-shot photograph. This thesis presents a procedure to estimate the distance between the pupils of the eyes on a standard photograph using measurements made on a non-standard photograph. The procedure can be extended to other horizontal and vertical measures and the precision should be adequate for application to the forgery problem.