Use this URL to cite or link to this record in EThOS: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.440549
Title: The forensic identification of CCTV images of unfamiliar faces
Author: Davis, Jeremy Paul
Awarding Body: Goldsmiths, University of London
Current Institution: Goldsmiths College (University of London)
Date of Award: 2007
Availability of Full Text:
Access from EThOS:
Access from Institution:
Abstract:
Government and private crime prevention initiatives in recent years have resulted in the increasingly widespread establishment of Closed Circuit Television (CCTV) systems. This thesis discusses the history, development, social impact and the efficacy of video surveillance with particular emphasis placed on the admissibility in court of CCTV evidence for identification purposes. Indeed, a verdict may depend on the judgement by members of a jury that the defendant is depicted in video footage. A series of 8 experiments, mainly employing a single-item identity-verification simultaneous matching design were conducted to evaluate human ability in this context, using both photographs and actors present in person as targets. Across all experiments, some trials were target absent in which a physically matched distracter replaced the target. Specific features were varied such as video quality, the age of participants, the use of disguise and the period of time between image acquisition and identification session. Across all experiments performance was found to be error prone, even if the quality of the images was high and depicted targets in close-up. Further experiments examined jury decision making when presented with CCTV evidence and also whether extensive examination of images would aid identification performance. In addition, evidence may be presented in court by facial structure experts in order to verify the identity of an offender caught on CCTV. Some of these methods were discussed and a software package was designed to aid in the identification of facial landmarks in photographs and to provide a database of the physical and angular distance between them for this purpose. A series of analyses were conducted and on the majority of these, the system was found to be more reliable than humans at facial discrimination. All the results are discussed in a forensic context and the implications for current legal practices are considered.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.440549  DOI:
Share: