10 Jan Safe and sound
Timothy Compston discovers that there is a real buzz around sound analysis for security and public safety
The volume of security-related applications where sound is playing a role is certainly on the rise, whether that be on the biometrics front in the form of voice recognition to allow someone to securely access their bank account; in smart homes to detect breaking glass or to identify and locate gunshots in an urban area.
One capability we are hearing more and more about is voice recognition to improve the experience of customers when accessing services and to tackle fraud. To put the trends here into perspective, a report issued by Research and Markets on the biometric market in the Middle East is predicting growth of CAGR of 19.35% up to 2019. Voice recognition, says Research and Markets, is expected to gain more traction for verification in banking, specifically for mobile banking, as in the view of the report’s authors it is ‘affordable and user-friendly’, and, crucially, does not require any hardware investment.
In terms of the Middle East institutions that are trailblazing voice recognition, Research and Markets singles out Abu Dhabi Commercial Bank (ADCB) which has already deployed biometric solutions for secure and easy customer identification. In fact, last year ADCB received a prestigious award from Asian Banking and Finance for the introduction of its voice biometrics initiative, thought to be a first for the Middle East region. ADCB VoicePass allows customers to be authenticated by the bank’s system using a customized ‘vocal password’ for their phone banking accounts.
Speaking at the time of the award win, Abdul Shakeel Abdul Hameed, head of retail banking at ADCB, said: “The success of ‘VoicePass’ is a clear indication that our customers appreciated the ease of authentication combined with a high-level of security afforded by biometric technology.” In addition, ADCB believes that VoicePass underlines the bank’s commitment to serve multi-cultural communities with the voice biometric program now available in English, Arabic, and Hindi.
Over in the UK, a major bank driving down the voice security technology road is Barclays, which pioneered the technology to a limited number of customers in 2013, and claims that it was the first bank in Europe to do so. In the summer of 2016 Barclays announced that it was introducing voice security technology for all its UK customers, so passwords would no longer be needed by callers. Stressing the security advantages that come with voice recognition, Barclays’ pointed out that a person’s voice is made up of over 100 characteristics, based on the physical configuration of the speaker’s mouth and throat. Crucially, the bank claimed that voice recognition, as a highly personal form of identification, should create a greater barrier to fraud.
Issues with fraud
When it comes to voice recognition there can, of course, be no room for complacency on the security front. Earlier this year the UK’s BBC News Channel highlighted how a reporter on its Click programme – Dan Simmons – set up an HSBC account and signed up to the bank’s voice ID authentication service but the bank then let his non-identical twin, Joe, access the account via the telephone after he was able to mimic his brother’s voice, albeit after seven failed attempts. Following the investigation HSBC told the BBC that it would ‘review’ ways to make the ID system more sensitive.
Of course, it is highly unlikely that twins are going to attempt to access their siblings accounts for nefarious reasons but looking ahead another issue may emerge from the advances being made in the development of synthetic speech. Worryingly, it may, in the foreseeable future, be possible to replicate artificially someone’s voice from a small number of phrases.
Given this ever-changing threat landscape, vendors offering voice recognition are now looking to further ramp-up the security and capabilities of their solutions and Nuance Communications – a leader in the field – is no exception here. Recently, the company announced plans to reduce the risk of consumer fraud with a new suite of biometric security solutions.
Once up and running the Nuance Security Suite is designed to apply deep neural networks (DNN), as well as advanced algorithms, to help detect synthetic speech attacks – with the latest release including a 45 to 55% improvement in this area. The Security Suite also integrates facial and behavioural biometrics to enhance fraud prevention. Speaking at the launch, Brett Beranek, director of product strategy – biometric security – at Nuance was keen to underline the proven security track record of the vendor’s biometrics technology: “Already this year, around the world, some 150 million people have made more than one billion successful authentications using Nuance biometrics technology, with not one reported act of fraud.”
Turning to other sound security solutions, with attacks involving active shooters in urban areas on the up – and the potential for incidents to take place across multiple locations – law enforcement and security professionals are keen to look at technologies that can help to distinguish gunshots from other sounds.
US vendor SST Inc. – a leader in gunshot detection and location technology – has developed its ShotSpotter technology. One version of this, ShotSpotter Flex is helping law enforcement agencies in cities across the US and Cape Town, South Africa, by directing police to the precise location of illegal gunfire incidents. The company says that ShotSpotter Flex operates through multiple ‘collaborative’ acoustic sensors that activate when gunfire occurs outdoors. The sensors and software triangulate and pinpoint the precise location of each round fired in a matter of seconds. SST has an Incident Review Centre (IRC) where an acoustic expert analyses the data and qualifies the incident before an alert is issued.
Beyond helping to tackle gun violence in cities, SST has also rolled out an acoustic gunshot detection solution – ShotSpotter SiteSecure – to act as part of a layered perimeter defence for critical infrastructure sites, such as energy production and transmission facilities, and is built around a small number of fence or pole-mounted networked sensors.
Moving on to other solutions, mobile and rapidly deployable gunshot detection systems are starting to gain more traction. Many of these have already proven their worth detecting shooters in conflict zones like Afghanistan and Iraq, with Boomerang a case in point. This features an array of microphone sensors that are fitted to a mast on a vehicle. Since 2003 Boomerang has gone through several iterations and provides visual and auditory – ‘voice’ – information on the position, range, elevation and azimuth of a shooter. Interestingly, Boomerang III models – from Raytheon BBN Technologies – that employ passive acoustic detection and computer-based signal processing to locate a shooter in less than a second – were deployed on telegraph poles and tripods during the 2014 Boston Marathon to ramp up security, one year on from the terrorist attack there.
Regarding the wider trajectory of gunshot detection, it is also now being integrated with other security platforms. For its part, Genetec – a leading provider of open architecture security and public safety solutions – now supports ShotSpotter gunshot detection in its unified IP security platform – Security Center. In addition, Genetec’s Citigraf, a new public safety decision support system, that has been created for citywide law enforcement and public safety agencies, features a correlation and analytics engine that instantly detects and displays relevant information from disparate systems, including gunshot detection technology.
Giovanni Gaccione, law enforcement practice leader at Genetec says that a key advantage with Citigraf is the way it is able to provide a more comprehensive view of a situation, for example, around a gunshot: “Having a system tell you where a gunshot happened is great but really what is required is what is happening in that area or has happened in that area. We look to tie-in to sensors and databases so that when a gunshot comes in the system can start automatically trying to find out what happened prior to an event or in the prior life of that event.”
Gaccione goes on to say that Citigraf is even able to work with license plate (automatic number plate) readers allowing vehicles in the area, at the time of a gunshot, to be identified and even listed by the speed and the direction they are travelling in.
With all the talk of smart homes, sound solutions are also starting to make their mark in the domestic security arena. A case in point is Cambridge UK-based vendor, sound recognition company, Audio Analytic which offers, what it describes as, ‘embeddable software sensors’. Demonstrating the shape of things to come here, a smart lightbulb ‘Sengled Voice’ won the 2016 CES Best of Innovation in the Smart Home category. Essentially, this is a light bulb with an integrated microphone and speaker that allows the detection of noises, like glass breaking or a smoke alarm, which are then analysed by Audio Analytic’s ai3 technology and communicated to a homeowner via a mobile alerts.