Search

Previous research on composite systems has cast doubt on their potential to produce a good likeness. Poor memory of the target's face and problems in translating a verbal description into a visual mental image were examined as possible reasons for this. A computerized composite system was used to construct facial images of famous individuals. Results indicated that a reasonable number of composites were identified overall. Composites produced directly by the “operators” tended to be more accurate than those produced via “describers”. In addition, composites produced via describers, but not those directly by operators, were less accurate when created from memory than when a photograph was provided. This suggests that verbal descriptions, as well as facial memory, may limit composite accuracy. [ABSTRACT FROM AUTHOR]