UK: COMPUTERS VERSUS CRIME. - A fast fingerprint recognition system, graphical information technology for analysing crime patterns, and virtual reality - they are all being brought in to help the police with their enquiries.

Last Updated: 31 Aug 2010

A fast fingerprint recognition system, graphical information technology for analysing crime patterns, and virtual reality - they are all being brought in to help the police with their enquiries.

By Jane Bird.

Anyone who has been burgled is likely to have received a visit from their local police scene-of-crime expert in search of fingermarks. The UK is one of the biggest collectors of fingerprints in the world. Around 900,000 useful marks are collected each year from the one million sites visited. These can be compared with existing databases to find out whether known criminals are involved and to spot links between crimes.

The problem is that the searches and comparisons are expensive, time-consuming and slow. This is because many of the local police-force collections are manual or only semi-automated. The national collection, meanwhile, which is computerised, is housed at New Scotland Yard and administered by the National Identification Service (NIS). Police forces have to check the fingerprints of anyone who is charged against the national database to confirm their identification or to add a new record, but waiting for the NIS to perform this routine can take weeks.

That is all about to change, thanks to National Automated Fingerprint Identification System (NAFIS), the 100 million computerised fingerprint database currently under development.

NAFIS will hold the six million records of known criminals expected by the year 2000, along with two million unidentified marks from the scenes of crimes. Ultimately, each police force in the country will be able to access it via a graphical, user-friendly computer screen. The system will allow them to scan in marks that they have picked up from crime locations, or from people who have been arrested, and produce a short-list of matches within a few seconds.

'NAFIS will be the biggest, most advanced, fingerprint recognition system in the world,' says Mark Goulding, the project's director. Larger numbers of fingerprints may be held in systems in California and Tokyo, he admits. 'But NAFIS will be more flexible and powerful, for example in its ability to handle unidentified scene-of-crime marks.' The database, which will occupy four tera (million million) bytes, will comprise one of the largest magnetic disk stores ever built.

The system works by using algorithms to search for approximately 100 minute details in the ridges of each fingerprint. This method produces a far more accurate short-list than previous fingerprint recognition systems which look for larger characteristics such as loops, whorls and so-called tented arches. 'Despite the huge size of the database, you are more or less guaranteed that, if the print is on file, it will be in the top few,' Goulding says. In fact, the technical specification requires that the target print, when present, should be listed among the top three in 99% of cases, and top in 98% of cases.

The database will help the police establish links between crimes by matching fingermarks found at separate sites. It should also improve management control by providing an audit trail of individual investigations. The system has been designed so that in future police will be able to use mobile fingerprint scanners to carry out on-the-spot checks at the scene of a crime. They could also use mobile scanners to identify individuals arrested in the street, find out whether they have criminal records and, if necessary, charge them there and then.

NAFIS is due to begin a pilot implementation next year in eight forces and is expected to be fully operational by mid-1998. It will then begin being rolled out to the remainder of the 43 police forces in England and Wales. The project should be complete by April 2001, when it will also be available to forces in Scotland and Northern Ireland. By the following year, it is expected to have paid for itself in the improved productivity of fingerprint officers.

The system exemplifies one way in which technology can be exploited by the police to make sense of vast quantities of data. This is the key to how it can help fight crime, according to detective superintendent Ken Grange, the newly-appointed police adviser to the Home Office. Grange's role is to set policy and co-ordinate the development of technology for police forces across England and Wales so that they get optimum benefit from investment.

Having spent nearly 10 years in the flying squad tackling organised crime, Grange says that one of the frustrations of being a police officer is knowing that you could probably identify the criminal or solve the crime if only you knew where to look for the information. 'The chances are that out there somewhere is the answer to the question. The problem is how do we got hold of it?'

One of Grange's prime goals is to find more effective ways of using large volumes of data. Take phonecalls from the public, for example. This is one of the main sources of information for the police, but the way calls are currently dealt with means that links and patterns are easily missed.

'It may be 12 hours between someone phoning in to report a door-to-door caller posing as a bogus official and someone else reporting a burglary,' says Grange. 'By that time a different officer will be on duty and there may be nothing obvious to connect the two events.'

Improved data-handling techniques would help the police to identify links and at the same time take account of much greater quantities of data.

'There is plenty of it about,' Grange points out. Databanks are being compiled all the time on everything from mobile phone calls and electronic cash machine withdrawals, to registers of parking permits, store cards and firearms holders. 'Suppose we are investigating a murder and know a red Sierra from north London is involved,' says Grange. 'Instead of sending a team of police officers on house-to-house enquiries, we might be able to narrow the search enormously by using this type of information.'

Another way that computers can improve police efficiency and effectiveness is by helping them track the evolution of a crime on the screen. This is where graphical information systems (GIS) come in, with their colourful displays of maps and diagrams. GIS make it possible to replicate events on maps and to plot scenarios based on sightings of suspects or vehicles in different places. 'As the story evolves, the picture changes, and if there are several suspects, you can draw links on screen. This is a tremendous benefit,' says Grange.

GIS technology is also invaluable for spotting patterns in the time, location and method of crimes. Leicestershire Constabulary is pioneering on-screen crime pattern analysis as part of a criminal intelligence system it has commissioned from ICL for use by its 2,500 police staff. The system is due to go live this year. 'We collect masses of data such as where and when a crime was committed, what was done and what the modus operandi was,' says John Hughes, Leicestershire Constabulary's IT manager. 'The process of collection is extremely laborious, but technology makes it a worthwhile exercise because we can then make sense of what we have found.'

Police forces have for some time used text-based systems to search, for example, for all burglaries at a particular time, in a specific area or involving the same methods. The advantage of GIS, is that it overlays the data on Ordnance Survey maps. It is much more useful being able to see the data on a map of the county, says Hughes. 'The bobby on the beat can then spot, for example, whether crimes are being committed in a particular area at a certain time of night, and can plan either to be there to catch the criminal or to have a sufficiently strong presence on the street to act as a deterrent.'

Leicestershire Constabulary's system will also make it possible to track the return on investment of deploying officers in a particular area by measuring the crime clean-up rate. Section heads can then make informed decisions about where and when to send out their forces. The constabulary hopes it will lead to significant improvements in the detection of Leicestershire's 100,000 crimes a year.

Virtual reality is another technology with tremendous potential to help the police, says Grange. 'In a hostage or siege situation, for example, VR can be of tremendous value if the police need to gain entry to a building.' VR can recreate the interior of a building on-screen using data from video cameras, floor-plans and construction documents. This enables officers to familiarise themselves with the layout of passages and potential routes for entry or escape before they go in. 'This is invaluable because, with all the planning in the world, intelligence can go awry,' says Grange.

'Somebody might open the wrong door or park in the wrong place.' By giving officers the chance to practise their tactics, VR minimises the risks.

Not wishing to be overtaken by business and commerce, the police have even begun using the Internet in their fight against crime. Grange cites TV programmes such as Crimewatch as an example of how the public's appetite for crime could be mobilised in cyberspace. 'We might use the Internet to circulate details of missing persons and vehicles, or stolen property,' he says. 'But as with other sources of information, it would have to be updated regularly if it were going to be of any use, so we have to be prepared to put lots of effort in.' The danger of approaching a wider public on the electronic highways, he admits, is that you are a lot more visible when you get things wrong. 'We don't want to fall flat on our faces.'

In the course of the next year or so, Grange will evaluate a host of other technologies. An automatic, number-plate recognition system, for instance, has been developed by ICL. Its potential uses include monitoring bus lanes, checking for invalid Road Fund licences and spotting stolen vehicles by identifying their number-plates. Similar digital recognition technology is also being used to improve accuracy in the visual identification of criminals. 'By examining physical characteristics such as the shape of someone's face, the position of the eyes, nose and mouth, or the overall bone structure, height and posture, such systems can help identify people who are masked, wearing disguises, or whose image is only poorly captured on video cameras,' Grange says.

But putting together complex IT systems for the police is not easy. In the case of NAFIS, for example, Goulding initiated the procurement in 1992, yet it was not until the beginning of 1995 that he signed the final contract.

One reason for the delay was the time involved in explaining to the 18 companies that initially expressed an interest in bidding precisely what was involved. The operational requirement ran to seven volumes and listed 3,000 criteria. Even though most of the components were available off-the-shelf, it is no easy task putting them together.

Much of NAFIS's complexity is due to the processes involved. For example, extensive cross-referencing is necessary so that, if someone is acquitted, the police can be sure that they have deleted all references to that person on the system. 'Also, we wanted to specify something that would work for the UK and for criminal justice agencies around the world,' says Goulding.

'We don't want to be on our own because the system would then become too expensive to maintain.' Another problem is that, as with all high-tech projects, police systems are seldom completed without a hitch. A consortium of police forces involved in an interim fingerprint system which has been dropped is currently suing IBM for £10 million for allegedly unsatisfactory work, while IBM is counterclaiming for twice that sum. Lawsuits have even erupted between two companies collaborating on preparatory work for NAFIS.

When eventually the disputes are over, the technical and management issues resolved and NAFIS is fully installed, the police will have gained an excellent tool. But don't get it out of proportion, warns Grange. A tool is all it is. Nobody will take the machine's word for it - police will still perform manual checks on all fingerprint identification. 'Technology is an aid, not the answer,' Grange says. 'There are no "whodunit?" buttons. Solving crime is always done between the ears.'.

Find this article useful?

Get more great articles like this in your inbox every lunchtime