Identifying novel markers of concealed face recognition

Lead Research Organisation: University of Stirling
Department Name: Psychology


Suppose you are being questioned by police about a serious crime. They show you a picture of the victim, or of your accomplice, either of whom you should know only if you were involved. You therefore deny knowing them. The aim of this project is to find ways to help police detect such lies, by advancing our understanding of the markers of recognition. The proposed experiments are specifically about recognising faces, not detecting lies more generally.
Our approach is known as the concealed information test (CIT). This is widely used in Japan to uncover guilty knowledge about a crime - something such as a murder weapon, which only the person who did it should know. So, for example, a suspect might be presented with a series of pictures of possible murder weapons. When the true weapon appears, a guilty suspect produces a rapid recognition response that is hard to control. Used carefully, it has been shown to be a useful source of information, unlike standard lie-detector tests which are very error-prone. There is, however, little work on using the CIT with faces.
We plan to use eye-tracking, together with other measures such as skin conductance, facial expressions and vocal cues to provide objective evidence of recognition. There are systematic differences between the patterns of eye movement on familiar and unfamiliar faces. For example, recognition causes the first fixation to a familiar face to be longer, while pupils also tend to dilate. Part of our work will be to look for ways to combine different measures into one overall indicator of concealed recognition.
We plan to study two other, novel approaches to detecting recognition of faces, which rely on the differences in how we process familiar and unfamiliar faces. We can recognise familiar faces even in very poor quality images, perhaps quite small and blurred. However, to be confident that we do not recognise a face, we need to see it rather clearly. One of our methods starts with images that are very blurred and gradually makes them clearer. The other presents a very blurred image that has a clear 'spotlight' determined by where the 'suspect' is looking. In both cases, people who are trying to conceal recognition reject familiar faces as unknown too soon. Genuinely unfamiliar faces take longer to reject.
Pilot work indicates that all three of our proposed methods show promise. The research will investigate how to make them work well, for example what size the 'spotlight' should be in our third approach. We then aim to identify the most reliable measures. We need a method that is robust in two different ways. First, it should produce a signal that can be detected in a single person - and across as many different people as possible. Second, it should be resistant to attempts to conceal deception of the sort that make traditional lie-detectors so unreliable. Again, our pilot data are promising.
Our results should help understand better how our brains recognise faces. During the project, we shall develop methods for analysing eye-tracking data, which will be released for other researchers to use. We shall also collect more face images that will be added to our existing face database, which is also available to other researchers.
The ultimate aim is to develop a method that will help with the detection of crime and breaking criminal and terrorist networks. We shall therefore work with the relevant department in the Home Office, and with police in Scotland, to ensure that our results can be applied successfully. We shall also collaborate with researchers in Japan and the Japanese National Research Institute of Police, where the CIT has been successfully applied in daily field practice for many years (though rarely with faces due to a lack of research) and is admissible as an evidence-based protocol in courts. Knowledge exchange and collaboration between UK and Japanese researchers is key to the originality and success of this proposal.

Planned Impact

The goal of this research is to establish robust approaches for identifying when someone recognises a familiar face but denies knowledge of their identity. It therefore has the potential for direct application in the interviewing of suspects, by police and other security forces, when dealing with this sort of deception.
We have established formal interest in our work with a range of stakeholders at the regional and international level:

1. The Home Office Centre for Applied Science and Technology (Marek Rejman-Greene, Senior Biometrics Adviser)

2. Scottish Police (DCI Mark Bell)

3. The Japanese National Research Institute of Police Science (NRIPS, Chiba) and experts in research and the use of the Concealed Information Test in field practice (Professor Shinji Hira).

Stirling has a unique MSc in the Psychology of face perception; the project will afford students on this and at final year undergraduate level the opportunity to engage with applied research.

There is considerable general public interest in both face recognition and the detection of crime. We plan to engage the public directly through science centres and festivals. The national press reflect the public interest and we shall promote findings through press releases, writing articles for The Conversation and through social media such as Twitter.


10 25 50
Description EAPL 2018 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Other audiences
Results and Impact Talk at European Association for Psychology and Law annual conference: Detecting concealed face recognition with blurred faces
Year(s) Of Engagement Activity 2018
Description Explorathon 2018 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Public/other audiences
Results and Impact Public engagement activity, piloting some experiment ideas and explaining our work to those who took part.
Year(s) Of Engagement Activity 2018