On the costs and benefits of display format in a video-based observation task
The role of colour and motion information in 'active vision' tasks is of both practical and theoretical interest. In real world interactions, Levin et al. (2002) have shown that detection performance for unexpected 'person-changes' can be remarkably poor (even as low as 12%). In the recognition of unfamiliar faces from CCTV footage, somewhat surprisingly, Bruce et al., (1999) found no advantage for colour presentation over monochrome black and white (B&W). However, subsequent experiments have shown a benefit for colour over B&W static natural scenes in a recognition memory test (Wichmann, et al., 2002). This latter finding chimes with beliefs held by the general public and many CCTV user groups, but in this faith in colour justified?