As the road network continues to get busier with more vehicles on the road now than ever before, the number of people turning to the rail network has had a noticeable boost especially in large cities such as Liverpool and London. In response to this steadily growing interest Rail Operators have increased their focus on improving the overall passenger experience, whilst attempting to find a solution to manage the increasing demand.
Some areas that receive ongoing interest is the ticketing process and passenger entry through barriers at stations. With such large volumes of passengers, the traditional manned booths are no longer sufficient on their own, which is why there is so much attention on finding new technological solutions. Smartcard solutions, such as TfL’s Oyster Card, and the new support of contactless payment or smartphone applications, such as Apple Pay, have begun to make a difference but they are by no means the final solution.
Having recently explored video analysis/recognition technologies for a variety of applications, the recent developments by Bristol Robotics Laboratory caught our eye. Bristol Robotics Laboratory have been working alongside Oyster Card developer Cubic Transportation Systems on a new innovative solution that if proven to be reliable could see commuters using their face in place of traditional tickets. The system works by using rapid flashes of infra-red lights to help a camera capture the shape, orientation and even texture of a person’s face, which can then be run through a database of known customers so as to identify who they are.
As with any system that involves the transfer of money, some consideration must be given for the potential of identity theft and other security risks. In an era where social media is king it has never been easier to obtain an image of someone’s face, and it’s not as easy to change your face as it is to change your PIN number. Having thought ahead, a solution has already been put in place to drastically minimise this risk.
By using high speed infra-red lights the system is not only capable of measuring the distance between various points on the human face but also sensing shape in space, essentially allowing three dimensional imaging. By gathering this 3D information it is impossible to fool the system with a printed image of a person’s face, in fact the only way you could fool the system would be to carry around an accurate 3D representation of your face.
Having a solution that minimises the risk of identity theft will certainly help minimise the concerns of the general public, but it is still very unlikely that wide spread acceptance will be a rapid process. As the registration process for this new system will most likely have to include a thorough three-dimensional scan of each user’s face, which is unlikely to be a lightning fast process in itself, it will be important for the system to be clearly faster than current solutions otherwise its popularity with the public will take a serious hit.
Issues with public opinion and security aside, anyone who has experience of working with video analysis and identification systems will tell you just how fickle they can be. Take the unfortunate incident from the release keynote of the new Apple iPhone X for example, where the new facial recognition technology did not work quite as intended. This begins to beg the question of how effective would a facial recognition system actually be.
According to members of staff from Bristol Robotics Laboratory their solution is currently running at 95% accuracy, however this is under lab conditions without the risk of unexpected events. There will be a multitude of potential issues that could be presented to the system, of which only a finite number could be picked up in the lab testing process. If you take just a 30-second pause, immediately some potential issues spring to mind:
- Facial coverings – be it for medical (surgical masks, eye patches and bandaging) or religious (hijab, niqab or burqa) purposes
- Facial Alterations – be it purposeful (tattoos, plastic surgery, growing out facial hair) or unintentional (scar tissue, facial reconstruction, aging)
- Large crowds – stations are busy places and if the plan is to use this technology in place of barriers it will need to be able to handle lots of people in a small space
- People of different height – unfortunately the human race does not come in a standard size and as the plan is for the system to replace the access barriers at stations it will need to be able to cope with people at all different heights
- Changes in light – a problem experienced by all video analysis and recognition systems depending on camera placement the system may experience changing light levels
The only way to find out if the system will cope in the real world is to have it set up and running in the real world. The short-term plan is to initially install a small number of ‘fast-track’ lanes for use by a select number of passengers who will have their faces scanned as part of a registration process, with plans to expand availability dependant from the results of this initial stage.
Looking into the future if we can unlock our phones and pay for our fares with just our face what else could we control with a look. Could our face become the key to our cars, reducing the possibility of car theft, how about clocking our time spent at the office automatically with cameras watching the entrance and exits. The possibility for this technology is endless, but I seem to remember there being a book forecasting such a future where the life of its characters is far from a blissful utopia.
Big Brother Is Watching You.