Biometrics-Unique Identity Verification System

Introduction to Biometrics

Today's world is working under the umbrella of digitization. Everyone is nowadays predominantly performing his/her major and minor activities on the digital platform. People are surrounded by computing devices such as smartphones, iPads, wearable devices, home automation and many others. With the advancement of technologies, every individual is now storing his/her data, whether personal or professional, in digital format, which can easily be accessed anytime and anywhere. With the increase in popularity of the digital platform, it has also opened up the gateway to modern ways to commit crimes, known as cyber-crimes. Thus, to overcome this issue one has to identify modern methods to provide security. Here, security does not mean to secure only digital data but also means to provide security to protect digital devices from damage, attack or unauthorized access.

Major crimes on the digital platform happen due to the loss of identity, that is by gaining unauthorized access. In previous decades, a lot of research has been done to develop strong authorization methods like strong password systems, secure ID card systems, two-way authentication systems, 3D secure systems and many more. Every system has its limitations and flaws like stealing secret passwords/IDs, brute-force attacks, remembering strong and long passwords and many others. Further research on security comes up with a more reliable and better solution: biometrics (Narayanan and Chhabra 2013).

Biometrics is a process of measuring biological characteristics of an individual for recognition, authentication and verification. These biological characteristics carry some unique features that may help anyone to recognize the right person and then provide access controls to them. Every human being has various biological characteristics that are unique, distinct and are measurable, and in addition that higher uniqueness provides high security (Narayanan and Chhabra 2013). This will help in developing highly reliable systems in comparison to the traditional methods used for authentication. At present, a large amount of data and information is available which should be kept secure; hence, biometrics is widely used for securing them as it provides a high degree of accuracy in terms of recognizing an individual. In terms of physical and behavioral characteristics, humans are all different from one another. With the help of these characteristics, one can analyze a person’s attributes, which will help an individual to stand out from the rest (Jain et al. 2004).

In simple language, one can say that biometrics is an autonomous recognition method for an individual based on his/her physical or behavioral eccentricities. On the other hand, biometric systems is a modern technology, which helps in recognizing a person's physiological and/or behavioral characteristics, verifies them, identifies an individual and distinguishes them from others. In a human body, there are several biological characteristics available that can be measured easily, but not every characteristic can be used as a biometric feature. Any biological trait falling under the category of biometrics must satisfy the following properties (Jain et al. 2004):

• • Robustness: This property indicates that the characteristic should not change over time, that is, it should be permanent. This is because if it changes with time, it will be difficult for a system to recognize the unique property of an individual. Along with that, for the matching criteria, it should be adequately invariant. The reason behind this is that if there is high intra-class variability, there will be difficulty in distinguishing between individuals.
• • Availability: Secondly, the characteristic should be universal, in other words every individual should possess that characteristic. Until and unless availability or universality of the property are not followed, one cannot include that characteristic as a biometric.
• • Collectability: Another important property is that the characteristic should be easy to acquire and able to be measured quantitatively.
• • Distinctiveness: This property concerns the uniqueness of the characteristics. In the broad population, there must be large variation in the characteristics among the people. In simple words, one can say that two or more people must carry such characteristics that make them sufficiently different from each other.

While working with these properties, one has to consider some other issues also related to biometrics. These issues are related to the computational point of view: circumvention deals with how the system can easily be bypassed (spoofed) by falsified approaches; performance deals with accuracy and speed of the system for recognition; and acceptability deals with the number of people who accept the use of a particular biometric identifier in their day-to-day life. Hence, for biometrics systems, the above-mentioned properties must be satisfied by the biometric characteristic chosen (Jain et al. 2004).

Authentication vs Verification

In terms of security, the very first step is concerned with verification, followed by authentication of the user. With the advancement of technology and the initiation of biometrics, security has been enhanced a lot. However, people generally use the terms authentication and verification interchangeably; actually, both words have different meanings. The difference between the two is revealed in Table 12.1 below.

Types of Biometrics: Physical and Behavioral Biometrics

Based on the various characteristics of human beings, biometrics are divided into two categories, namely physical biometrics, which are those that need physical contact

TABLE 12.1

Difference Between Verification and Authentication.

between sensors and the human body to capture the features for identification and verification. They focus on the physical characteristics of an individual. Secondly, behavioral biometrics are those in which no physical contact is required to capture the features for identification and verification. They focus on the behavioral characteristics of an individual.

Biometrics uses a sensor, which behaves as a transducer that is used to convert the user's biometric traits, such as iris, face, fingerprint and so on, into an electrical signal. These sensors come in two different types: physiological and behavioral. Figure 12.1 below illustrates the basic working of the biometrics system. These systems are used in various applications in this digital world.

FIGURE 12.1

Block diagram of a biometrics system. (Based on Hernandez Alvarez et al. 2009)

12.3.1 Physical Biometrics

Physiological biometrics works with humans to identify physical characteristics. It measures the unique patterns of the body for identification. Different types of physiological biometrics commonly used nowadays are:

• • Fingerprint recognition: Fingerprint biometrics have been considered the most popular and effective authentication technique for the identification of human identity as they work based on a natural truth that no two individuals have common fingerprints (Eng and Wahsheh 2013, Alsaadi 2015). It is also one of the oldest ways of confirming one's identity. Previously, ink and paper were used to take a thumb impression and to verify the human being; however, fingerprint biometrics is an inkless method, which senses the ridges and valleys on a surface tip of the finger.
• • Face recognition: This is a passive technique that depends on a unique shape like the position of eyes, chin and lips, and the distance between features, cuts and so on (Koong et al. 2014). It is widely accepted and used to identify a person from a digital image or a video source. Face recognition can be done in two ways: static recognition based on a static camera photograph, and real-time recognition based on video camera (Satone and Kharate 2012). The accuracy of the face detection system depends on various factors such as makeup, aging effects, motion estimation and low-quality videos.
• • Iris recognition: This is a method for confirming a person's identity from a random pattern of the iris. The iris is a muscle in the human eye that controls the amount of light entering the eye. Infrared light is used to take a high-precision photograph of the iris and this is then compared with the existing templates (Как et al. 2010). It is highly accurate and really hard to spoof but it is very expensive to set up.
• • Hand geometry: It is one of the oldest techniques which was first mentioned in the 1970s. The concept uses the measurement of different traits of human hands like palm thickness, length of the hand, measurement of fingers and so on. The advantage of the system is that it is widely accepted and easy to implement but it has a disadvantage as well in that it is not highly unique.
• 12.3.2 Behavioral Biometrics

Behavioral biometrics is a branch of science that utilizes human behavior (like habits and movements) to recognize a human being. Some of the popular behavioral biometric techniques are:

• Voice recognition: This is also known as speaker identification where the voice template is matched with different stored voice templates to identify a human being.

It first gained attention in 1952 with the development of a system capable of identifying spoken figures. In the last decade, the voice recognition hardware devices gained a new boost with the introduction of devices like Alexa and Google home. According to one of the surveys conducted by Strategy Analytics, the sales of voice- enabled devices will exceed 15 million by 2020.

• Keystroke dynamics: Keystroke dynamics depend on the style in which one writes with a pen or types with a keyboard. The rhythm in which the individual performs the said actions is measured to identify the human being. It is based on the earlier concept of the telegraph.