When it comes to the fight against crime, new software and technologies are proving essential in helping police and security forces identify suspicious individuals as well as actual perpetrators, aiding their prosecution. British universities are at the forefront of developing such software and systems, ranging from enhanced close circuit television (CCTV) units to new forensic techniques – and even ways of spotting people acting suspiciously before a crime occurs.

At Portsmouth University, researchers are using sound cues to identify potential crimes in progress. If a particular noise is identified, this can trigger CCTV cameras to turn towards the source, making sure any crime is captured and alerting control room operators. While CCTV surveillance can help identify criminals after a crime has occurred, perhaps recording someone making their escape down a nearby road, cameras do not always catch the act in progress. By modifying software currently used to identify visual patterns, the researchers at Portsmouth aim to allow the technology to pick out unusual noises and train cameras in this direction. ‘We are looking at sound and the signatures in sound – you can identify an abnormal noise such as a scream by looking at the distinctive  shape of the sound wave,’ explained Dr David Brown of Portsmouth’s Institute of Industrial Research, who is leading the project.

Over the next three years the university will create software that will fit an artificial intelligence or AI template to the waveform and use fuzzy logic if the fit is not exact. For example, the sounds of a car window or a bottle breaking have different waveforms but the same generic shape.

The software will work alongside CCTV-based human-motion-analysis systems that have been developed at the same institute. These can identify movements such as whether a person is punching another or simply reaching for something instead. The potential market for software incorporating the algorithm created from the research could include local councils, private security firms, car parks, shopping centres, football stadiums  and public transport. Eventually, the team hopes to have generated algorithms that can be incorporated into a commercial software suite, with each generation of algorithms becoming more sophisticated as the project progresses. The system could one day become so intelligent that certain words would trigger the camera’s response.

Gun fighting
Elsewhere, Brunel, Cranfield, Brighton, Surrey and York universities are creating  new bullet-tagging technology to give forensic teams a new tool in the fight against gun crime. The tags, which measure 30 microns in diameter, are applied to gun cartridges by being embedded in special cartridge coatings. The tags primarily consist of naturally occurring pollen, a substance that evolution has provided with extraordinary adhesive properties, together with a cocktail of other chemicals. The precise composition of the coating can be varied subtly from one batch of cartridges to another, enabling a firm connection to be made between a particular fired cartridge and its user. The coating attaches itself to the hands or gloves of anyone handling the cartridge and is said to be very difficult to wash off. Project partners include the Forensic Science Service and defence firm BAE Systems.

The current success rates for DNA profiling using evidence from gun cartridges are only around 10 per cent because of the difficulty in retrieving fingerprints or DNA from cartridge surfaces. However, the team has developed a method of trapping forensically useful amounts of DNA on gun cartridges. It involves increasing the abrasive character of the cartridge case with micro-patterned pyramid textures, or adding an abrasive grit, held in place by a thin layer of resin, to the cartridge base. This rough surface is able to retain skin cells from a thumb as it loads a cartridge into a firearm. The technology has been designed to avoid damage to the DNA captured that is caused by temperatures generated as the gun is fired.

Self control
Finally, security at border control points is being tackled using a computer system that can detect guilt. Bradford University is working to develop a system capturing facial expressions, eye movement and pupil changes that would then be analysed by computer-based intelligent algorithms to pick out potential drugs smugglers. It will also find out whether facial blood flow, which is hard to control, could also be used in the profiling system to pick out smugglers who are good actors.  The team is collaborating with the Home Office and HM Revenue and Customs, as well as defence firm Qinetiq. ‘We will be visiting airports  later this year to collect data on how border control checks are carried out,’ said Dr Hassan Ugail, head of visual computing research at Bradford’s School of Informatics. ‘By doing this we will be able to see what sort of system is really needed.’

Experienced officers often become good at spotting suspicious individuals based on their behaviour. However, the sheer volume of people passing through ports and airports makes it difficult to closely examine each individual in turn, and up until now, teaching a machine to do this has  been hard. If successful, the system could have potential far beyond border control applications. It might, for example, be used in police interrogations and interviews.

Research into facial recognition taking place within the School of Informatics is feeding into Ugail’s work. ‘My reseach work mainly focuses on 3D face recognition,’ says Xia Han, a student from Daqing, China, who is studying for a PhD in facial recognition technology at Bradford, and whose work will aid  the system in recognizing people efficiently despite different facial expressions. Similar research is being carried out by another PhD student, Sokyna Al-Qatawnah, from Karak, Jordan. ‘Face recognition research  using automatic techniques emerged  in the last two decades. One reason  for this growing interest is the wide range of possible applications for  face-recognition systems. Another reason is the emergence of affordable hardware, such as digital photography and video, which have made the acquisition of high-quality and high-resolution images much more ubiquitous,’ she said. ‘I chose this field because I think the interest in 3D face recognition will be increased in the near future, opening the door for my career.’

As crime fighting becomes more futuristic, it is clear that universities are rising to the task of making sure that the authorities are being given all the help that they can in order to identify wrongdoers. Meanwhile, as governments continue to prioritise security, this looks set to be an area where research opportunities will grow.
 

Julia Pierce

Julia Pierce is a freelance science journalist based in Cornwall. She has written for titles including the New Scientist, The Independent, the Daily Mail and The Engineer.