Video surveillance and access control technologies are developing and improving every year. The most significant focus is now on biometric identification, and electronic keys or pass cards are no longer the norm. There are several types of biometric authentication, and we will discuss them in detail in this article.
Additionally, readers will learn about the sectors in which the technology is used most actively. In addition, of course, we will tell you about the prospects for the future development of BI tools.
What Is Biometric Identification?
Biometric identification (BI) is the process whereby a user presents a unique parameter and then automatically compares it with the entire database. The subject of verification can be a fingerprint, retina, full-face image, or another similar attribute. Some systems have built-in emotion recognition by Recfaces, which increases the accuracy of control technology to almost 100%.
This form of access control is convenient for end-users because the information necessary for authentication is always with them. Moreover, biometric access control considered the most reliable form because identifiers cannot transferred to third parties or copied. Generally, the most significant advantages of BI include:
- the system is considered very easy to use;
- the technology provides time-saving in searching for and using the identifier;
- the required parameter cannot be lost or stolen;
- the parameter necessary is almost impossible to fake.
Unfortunately, BI systems are quite expensive, so not every company can afford them. It is also worth considering that scanners or readers can fail due to mechanical damage. It is vital to create the safest possible conditions for their installation for a long operation period. For example, if a system purchased to control the building entrance, the company needs to protect it from adverse weather conditions (rain, hail, snow).
Modern BI Methods
Modern biometric authentication based on two methods:
- Static method — recognizes the physical parameters of a person (fingerprints, distinctive characteristics of the iris, the retina image, a thermogram, face geometry, hand geometry, and even a genetic code fragment).
- Dynamic method — analyzes the characteristic features of user’s behavior, which are demonstrated when performing any ordinary daily action (signature, keyboard handwriting, showing certain emotions, or voice).
Typically, the static method is considered to be mainstream in the global biometric security market. Dynamic authentication and combined information security systems occupied only 20% of the market. However, in recent years, dynamic protection methods have been developing. For instance, keyboard handwriting and emotion recognition are of particular interest for network technologies at the moment.
The main types of static authentication include face-ID, iris imaging, and fingerprinting. Let us consider them in more detail.
Face authentication can be 2D or 3D:
- Two-dimensional (2D) face recognition has used for a very long time, mainly in forensic science. However, 2D face recognition is far from perfect. Statistically, the probability of false positives with this authentication varies from 0.1 to 1%. In general, the technology captures facial features such as the contours of the eyebrows, eyes, nose, lips, etc.
- Three-dimensional (3D) face recognition uses algorithms that are more complex. The scanning procedure takes about 20-30 seconds. At this moment, the face can be rotated relative to the camera. It forces the system to compensate for movements and form facial projections with a clear definition. Scanners use unchanged characteristics such as the depth of the eye sockets, the shape of the skull, eyebrows, the height and width of the cheekbones, etc. The system will subsequently recognize the face even with a beard, glasses, scars, or headgear. In total, from 12 to 40 features of the user’s face and head used to build a template.
Iris authentication based on recognizing the unique features of the eye. This method’s reliability of control helps to distinguish between the left and right eyes of a person. This technology practically eliminates errors and failures during authentication because the iris does not change much throughout life.
Fingerprinting is the most popular biometric authentication technology based on scanning and recognizing fingerprints. Law enforcement agencies actively support this method to attract electronic samples to their archives. Such authentication carried out relatively quickly since the system does not require recognition of each line of the pattern. It is enough to detect coincidences in large-scale blocks and analyze bifurcations, breaks, and other distortions of lines.
Dynamic authentication methods include:
- Voice recognition. This technology conducts the identification of a person using only one microphone, which connected to a recording device. The huge advantage of this method is the right to perform hidden authentication. The user may not always be aware of additional verification, which means that it will be even more difficult for attackers to gain access.
- Emotion recognition. Active facial features considered as unique as fingerprints and the iris. Therefore, the scanner can read not just an image of a face but also its position in a certain emotion.
- Handwriting recognition is yet another sufficiently accurate verification method. The user needs to write something on the panel using the stylus. The check lasts up to 15 seconds, after which access opened or denied.
Present BI Application
Nowadays, biometric identification systems used not only for access control in large companies, where information leakage is a significant problem; many BI methods are relevant in forensic science and medicine too. For example, if the only piece of evidence is a recording of a telephone conversation, then voice recognition will come in handier than ever.
BI is also widely used at airports in many countries to authenticate employees while crossing restricted areas. These systems also performed well during experimental use with ATMs in:
- The USA;
Considering the development of this area, it is worth noting that static methods will still be popular going forward, but they will have to change. For example, the International Subcommittee on Biometrics Standardization (IS0 / IEC JTC1 / SC37 Biometrics) has recently been developing a unified data format for human face recognition based on 2D and 3D images. Most likely, these two methods will be combine into one biometric authentication method.
The biggest potential direction is the creation of combined BI devices. Such developments considered the most reliable in terms of protecting information using given biometric indicators. Obviously, it is much more difficult to falsify several indicators at once than one alone. The most reliable combinations are:
- iris recognition + fingerprinting;
- fingerprinting + hand thermography;
- emotion recognition + fingerprinting.
Combined BI systems are much more efficient than mono-modal solutions. This is confirmed by many studies, including the experience of a bank, which first installed a user authentication system by facial recognition (error rate due to low-quality cameras was 7%) and by voice (error rate due to background noise was 5%). As a result, the security department achieved almost 100% efficiency.
Thus, the conclusion is that biometrics is the key to full-scale data protection or obtaining the necessary information. Such systems are relevant for many industries and will continue to evolve.