Face Identification Briefing Paper
Face Identification software is a computer application which uses a digital representation from a picture, video frame or 3D scan to identify or authenticate a person’s identity. Generally this is done by matching corresponding facial features to those stored in a database. The software has a wide variety of applications as a strong security identification and can be used in combination with other biometric technology to make protected systems more secure.
Major technology companies like Apple, Google, Samsung, Facebook, and Amazon have begun realizing the impact that facial recognition software can have on their existing security infrastructure.Apple has been attempting to add movement capabilities to the system. The subject of facial scan can now be talking or moving during the scan. This allows facial recognition to be combined with other biometric security measures like voice recognition. With the ability to scan moving subjects and identify individuals in a crowd, facial recognition demonstrates its uses beyond personal security.
Face identification software can use either a 2D/3D image or video feed to create a digital representation and identify a face. There exists variance among different algorithms but the overall concept is the same. Every face has several “landmarks”, the system will flag these as “nodal points”. A human face can have up to 80 of these points. They represent areas of interest on the face which the system measures. Some examples of these measurements would be, distance between the eyes, width of the nose, depth of the eye socket, and more. These measurements will be stored in a database. When the system scans a face it will compare all these measurements to those in the database. Face Identification systems can also account for the orientation in different pictures not being uniform. For example, if an individuals face has been scanned using an image, the system will be able to recognize the real individual when scanning the person or a video feed. Facial recognition of 3D video feed can be performed even when the face is turned to a 90 degree angle.
Face identification offers another form of biometric identification systems. Several vendors have been leveraging face identification as a security measure either for their clients or for internal use. Its application is not limited to a form of security but it is also being used in healthcare and retail. Although 2D based facial identification may not be as accurate as other forms of biometric technology like fingerprint, it does have its advantages. The subject of a face scan does not have to know they are being scanned, which demonstrates the software can be used in large crowds to identify threats quickly.
Several vendors in today’s market are leveraging face identification for its many applications. For example, Amazon has been developing a system which allows users to pay for their items using an actionable image (selfie). A user can use a selfie in which they are moving or saying a particular phrase, as a password to verify their identity when paying for an item. The logic behind having the customer say a phrase or make a movement, is to try to overcome the possibility of the system being fooled by scanning a 2D image of the person.
After purchasing Face.com in 2012 Facebook began using facial recognition technology to connect users with their photos. When a user uploads a photo, the software will automatically suggest other people to be tagged. It will also group people together showing more content for people who are tagged in photos with the same user.
Apple has also released its new FaceID technology to be used in its new iPhone X. The technology uses a sensor on the front of the phone with two modules. One module projects a grid of 30,000 infrared dots onto the user’s face. The other scans the pattern, and either denies or grants access. The 3D facial map is stored locally in the device’s processor making it less susceptible to infiltration. Apple believes so much in the technology that it is said to move away from fingerprint biometric technology and to use FacialID exclusively in all of its new portable devices.
Canadian Government Use
The GC began using facial recognition technology during passport verification procedures. The project was brought forth due to concerns that individuals wanted by the Canadian Border Service Agency (CBSA) would falsify travel documents to exit the country. After correctly identifying 15 suspects during the 2016 trial, the CBSA released a statement that the GC plans to unveil un-maned kiosks at airports. These kiosks will use facial recognition in order to clear passengers. The overall shift to un-maned kiosks is intended to bolster security while improving congestion at airports. The Primary Inspection Kiosk (PIK) program has been in development since 2015. Portuguese company Vision-Box has installed 130 Kiosks in Toronto’s Pearson International Airport. The Kiosks are designed to take biometric data in two phases, facial recognition, and fingerprint biometrics. The kiosk will also be able to obtain iris data, a feature reserved for people travelling under the NEXUS program.
The use of facial recognition in border security demonstrates its broader use in preventing identity theft and fraud. At the moment driver’s licences in several provinces including Ontario, BC, and Manitoba contain facial recognition-ready photographs. With its wide array of applications, the GC will have to measure the privacy implications as the more prevalent facial recognition technology becomes.
Implications for Departments
SSC would benefit more as a user of this technology than a provider. For example, the technology could allow users to unlock smartphones or desktops using Face Identification rather than passcode or fingerprint.
If facial recognition technology (creating models out of millions of faces) were to be adopted by one or multiple departments, the computational needs to produce such models would be extensive. From this perspective if SSC were to support such a technology work it could aid in providing some of the computational power required either in the form of private or public cloud accelerated artificial intelligence or machine learning computing.
The largest challenge with facial recognition technology is privacy. One method to mitigate privacy concern is to store biometric data locally to prevent data loss and inappropriate cross-linking of data across systems. The data is better stored as biometric features rather than images of people’s faces (like storing a password as a hash value instead of the actual password itself). This inhibits the misuse of the image for unauthorized reasons. Also if this technology were to be used to authenticate identities in multiple areas there may be large amounts of hardware required. The use of this technology also requires users to view the use of such technology as a form of security instead of as a privacy concern.
Content to be added by each departments