1
Identification and Surveillance of Facial Images: Progress and Problems

Tim Valentine and Josh P. Davis

Eyewitness testimony is older than the law. Even today, with sophisticated forensic science, eyewitness testimony forms the bedrock of many criminal cases. Whenever a witness gives testimony in court, jurors, judge(s) or magistrate(s) are faced with two basic questions: Is this witness giving an honest account? If so, can their account be relied upon as accurate? There are many reasons why a witness may deliberately give false testimony or identify a defendant they know to be innocent. The witness may be seeking revenge, have been intimidated into giving a false account, or be motivated to deflect blame away from the true culprit. Legal procedure is designed to expose a dishonest witness. In an adversarial system, for example in the UK, US, Canada, Australia and New Zealand, the defence have the right to test the testimony of prosecution witnesses through cross-examination. Equally the prosecution cross-examines witnesses for the defence. Cross-examination has been described as "the greatest legal engine ever invented for the discovery of truth".1 It is intended as a method to expose a dishonest witness, but psychological science shows that cross-examination is ineffective in distinguishing reliable eyewitnesses from those who are honest but mistaken (e.g., Valentine & Maras, 2011; Zajac & Hayne, 2003, 2006).

Case Study

"George Davis is innocent" daubed on walls and bridges was a familiar sight around London in the late 1970s. The graffiti referred to a man convicted for an armed robbery, which targeted a wages delivery at the office of London Electricity Board (LEB), Ilford, on April 4, 1974. Acting on information received, two plain-clothed policemen were watching the building. Two guns were carried by the robbers, and as they made a desperate getaway, one of the policemen at the scene was shot in the leg, and several motorists were hijacked at gunpoint.

At trial in March 1975 the prosecution primarily relied upon identification evidence by the police officers at the scene, and by police and other witnesses at other locations as the robbers switched vehicles during a dramatic car chase. Blood samples, recovered from the scene of a crashed getaway car, did not match any of the defendants. George Davis was the only one of four defendants to be convicted. He was sentenced to 20 years in prison. The conviction was upheld by the Court of Appeal in December 1975.

There was a high-profile campaign against George Davis' conviction, which involved much graffiti around London. The campaign gained notoriety when the Headingley cricket pitch was dug up during an England v. Australia test match, preventing play from continuing. In May 1976 the Home Secretary took the exceptional step of exercising the Royal Prerogative of Mercy to release Davis without referring the case back to the Court of Appeal. The Home Secretary deemed the conviction to be unsafe because of doubts over the police evidence, but Davis was not held to be innocent.

In 1977, George Davis was caught in the act of an armed robbery on the Bank of Cyprus. He pleaded guilty and was sentenced to 15 years in prison. He was released in 1984 but convicted of armed robbery for a third time in 1987.

George Davis' conviction for the armed robbery of the LEB wages office was quashed by the Court of Appeal on May 24, 2011 - 37 years after the original conviction. The principal grounds were concerns about the reliability of the identification of Davis from a live identity parade (lineup) by the two police officers who witnessed the robbery. Most notably the prosecution had not disclosed that one officer, PC Grove, had previously identified a different man from police photographs. Prior to the identity parade in which George Davis was identified, the investigating officer had told PC Grove that he had been mistaken in his identification of the photograph. Confidential government papers are normally subjected to a 30-year embargo, after which the papers are held by the Public Record Office. In 2006 the embargo of the papers relating to the Home Secretary's 1976 decision to free George Davis was extended by 20 years.

Sources:

Davis v R. (2011) EWCA Crim 1258 (24 May 2011). Retrieved from http://www.bailii.org/ew/cases/EWCA/Crim/2011/1258.html

Wikipedia entry for George Davis (robber). Retrieved from http://en.wikipedia.org/wiki/George_Davis_(robber)

Courts have long acknowledged that a mistaken eyewitness may give convincing identification evidence. The extraordinary case of Adolf Beck, twice wrongly convicted on the basis of mistaken eyewitness identification, described in the case study in Chapter 6, resulted in the Criminal Appeal Act (1907) which established the Court of Criminal Appeal in London (Bogan & Roberts, 2011). Widespread concern about the reliability of eyewitness identification evidence in a number of English cases during the 1970s led the British government to set up an enquiry into eyewitness identification evidence (Devlin, 1976; see case study). Despite legal reforms in the UK since the 1970s, studies of police identification procedures have shown that a third of all identifications from live parades (Valentine, Pickering, & Darling, 2003; Wright & McDaid, 1996) and 40% of all positive identifications from video lineups are known to be mistaken, as the witness selected an innocent volunteer foil or filler (Horry, Memon, Wright, & Milne, 2012). The Innocence Project (2013) in New York has produced incontrovertible evidence of the devastating impact of mistaken eyewitness identification in the US. Over the last 20 years more than 300 prisoners have been exonerated by DNA evidence that proved they were actually innocent of the crimes of which they were convicted. The crimes were serious, mostly rape and murder, because physical evidence from which a DNA profile can be obtained is most likely to be available and collected in serious violent crimes. Mistaken eyewitness identification was the leading cause of wrongful conviction, and occurred in nearly 75% of cases.

In addition to the strong evidence of the high risk of mistaken eyewitness identification, research also demonstrates that approximately 40% of witnesses fail to identify anyone from a lineup. In many cases the witness may not have an adequate memory of the culprit. It may be that in an unknown proportion of these cases, the culprit was not included and the witness was making the correct decision. However, justice is served by developing procedures that both reduce the likelihood of an innocent suspect being identified, and enhance the likelihood that the actual perpetrator will be identified. Identification failures may leave a guilty suspect free to offend again.

The problem of distinguishing accurate from inaccurate identification is at the heart of Forensic Facial Identification. In the chapters that follow, distinguished scholars grapple with the problems of identification of suspects by eyewitnesses, from CCTV imagery, and identification of deceased victims from reconstructions of their facial appearance in life.

In many criminal investigations, the first problem the police may face is to identify a suspect. This issue is addressed in Part 2. If an eyewitness is available, the first step will be to interview the witness and in the course of that interview obtain a description of the offender. In Chapter 2 Fiona Gabbert and Charity Brown evaluate the relationship between the completeness and accuracy of a witness' description of the perpetrator and the likelihood that a subsequent identification from a lineup will be accurate. This is a difficult issue for the psychology of eyewitness identification because research findings are contradictory. Most researchers accept that, contrary to common-sense expectations, there is little relationship between the quality of a witness' verbal description of the perpetrator and their identification accuracy. On balance, laboratory research shows that a witness who gives a detailed description is no more likely to be able to identify the offender than a witness who can provide only a brief description. From their analysis of the literature Gabbert and Brown show that it is the inclusion of incorrect details in a verbal description that adversely affects identification accuracy. Therefore interview procedures that produce detailed descriptions by encouraging witnesses to provide details of which they are unsure are likely to impair eyewitness memory for the perpetrator and may increase the risk of a mistaken identification. Informed by this analysis, Gabbert and Brown provide practical advice for the employment of appropriate procedures most likely to obtain accurate descriptions from witnesses. These guidelines should help to safeguard the quality of any subsequent eyewitness identification.

Having obtained a description of the offender(s), in the absence of other evidence, the police may ask a witness to create a facial composite or likeness from memory. The image can then be publicized in the hope that somebody will recognize it as an individual they know and will provide a name to the police. In Chapter 3, Charlie Frowd reviews the development of techniques and methods used to construct facial composite images. This field has shown remarkable development in recent years. In 2007 the best recognized images were facial sketches produced by skilled police artists,...