© University of Kent - Contact | Feedback | Legal | Cookies
The University of Kent, Canterbury, Kent, CT2 7NZ, T +44 (0)1227 764000
Forensic Imaging Group
It is the job of the forensic scientist to establish a link between the perpetrator and the crime based on scientific evidence. Forensic science has traditionally been associated with the chemical sciences. Now rapid developments in digital and computer technology are beginning to change the scene. Advancements in CCD technology have made high resolution digital photography affordable and common place, which makes 'the digital image' an appealing medium for capturing evidence (for example CCTV). Digital photography, unlike analog photography, can be analysed relatively easily using image processing algorithms. For instance, the clarity of an image may be improved by removing noise and increasing the image contrast. Other virtues of digital images, of which there are many, include the possibility of image compression for data storage and fast transmission. It is well known within the image processing community that the FBI use wavelet image compression to store and transmit fingerprints (external link).
The forensic imaging group was founded by Dr. Chris Solomon in March 2002. The aim of the group is to investigate and develop the use of digital image processing as a tool for the forensic scientist.
The UKC Forensic Imaging Group specialises in the analysis of face images whilst retaining a more general interest in other digital image processing applications. Why analyse human faces? We can infer a wealth of information from a person's face such as age, race, gender and even mood. Since everyone's face is unique (with the possible exception of identical twins) we can also use it as means to identify an individual. Humans are social species and as such have evolved an exceptional ability to distinguish between different faces even under adverse viewing conditions (see example below). Since we use this ability on a daily basis it is easy to dismiss it as a trivial task. This misconception couldn't be further from the truth. In fact, the face recognition capabilities of the human brain provide the ultimate benchmark to which developers of artificial face recognition systems aspire. Artificial face recognition sytsems struggle to provide reliable identification under unfavourable viewing conditions such as poor lighting, pose variations or occlusions. These issues have been addressed by the members of this research group in the past and continue to feature in current research projects along with many other aspects of face analysis including face identification and face encoding.
Human Vision vs Computer Vision
If you view this image at a distance of approximately 3 metres
you should be able to see the face famous of a famous physicist.
Place your mouse cursor over the image to reveal the original photograph.
This image illustrates the remarkable face recognition capability of the human brain. It also underlines the formidable task faced by computer vision experts who work in the field of face recognition.