32 Recent trends in Personal Identification

Introduction to Biometrics:

The classical methods for personal identification depending mainly on personal probabilities are being increasingly challenged. Accurate and efficient recognition have become a vital  requirement for identification because of the intensified need of individualization and diversities of criminal activities. In recent years traditional methods of establishing a person’s identity are replaced by a sound scientific and justifiable protocol. Biometrics offers a unique and reliable system of identification with high confidence. Biometric recognition overcomes the loopholes that are found in traditional identification system. It is considered as a fundamental shift in the way the persons are identified.

Biometrics is the science of establishing the identity of an individual based on the physical, chemical or behavioural attributes of a person, (Jain et al., 2007). Biometric system uses a variety of physical or behavioural characteristics, including, finger print, face, hand/finger geometry, iris, retina, signature, gait, palm print, voice pattern, ear, hand, vein, colour, odour or the DNA information of an individual to establish identity. (Jain et al., 1999 and Wayman et al., 2005)These characteristics are known as traits, indicators, identifiers or modalities in the biometric literature.

Functioning of a Biometric System:

A biometric system is a pattern recognition device that acquires physical or behavioural data from an individual, extracts a salient feature, set from the data, compares this feature set against the features set stored in the database and provides the result of the comparison. Therefore, a biometric system is composed of four modules:

• Sensor module: This component acquires the raw biometric data of an individual by scanning and reading. For example, In case of fingerprint recognition, an optical fingerprint sensor may be used to image the ridge pattern of the fingertip. The quality of raw data is influenced by the scanning or camera device that is used.
• Quality Assessment and Feature Extraction module: For further processing, the quality of the acquired raw data is first assessed. The  raw data is subjected to signal enhancement algorithm to improve its quality. This data is then processed and a set of salient features extracted to represent the underlying trait. This feature set is stored in the database and is referred as template. For example, the position and orientation of minutiae in a fingerprint image is extracted by the feature extraction module in finger print biometric system.
• Matching and Decision making module: In this module, the extracted templates are then matched against the stored templates and a matching score is given. On the basis of the matching score, the identity of a person is validated or ranked.
• System Database module: This module acts as a storage of biometric system. During the enrolment process, the template extracted from raw biometric data is stored in the database along with some biographic information (such as name, address etc.) of the user.

Identification and Verification:

Biometric  system can  be  classified  into  two  main  categories  on  the  basis  of application  mode: Verification and Identification

• In the identification mode, the biometric system identifies an individual by searching the templates of all the individuals whose identification details are stored in the database. In this process, the system conducts a one to many comparisons to prove the identity of a person.
• In the verification mode, the biometrics information of an individual, who claims certain identity is compared with his own biometric template stored in the system database. This is also referred to as one to one comparison.

Characteristics of Biometrics:

The selection of each biometric trait depends on the variety of issues besides its matching criteria. Jain et al (1999) have identified seven factors that determine the suitability of a physical or behavioural trait to be used in biometric application.

1.   Universality: Every individual who is using the biometric application must possess the trait.

2.   Uniqueness: The trait must show a sufficient difference across individuals comprising the population.

3.   Permanence: The given biometric trait should not change significantly over a period of time.

4.   Measurability: The trait should be easy to get and digitize and should not cause inconvenience to the individual. It should also be amenable to process further in order to extract features from the acquired data.

5.   Performance: The recognition accuracy and the resources acquired to achieve that accuracy must meet the constraints imposed by the individual.

6.   Acceptability: Individuals who will access the biometric device should be willing to undergo for their biometric test.

7.   Circumvention: It refers to the ease with which the trait of a person can be imitated or copied by using artifacts (e.g. fake fingers in case of physical and mimicry in case of behavioural traits). The biometric system should be immune to the circumvention.

Physical Biometrics:

A biometric which recognizes an individual on the basis of his physical characteristics is referred as physical biometric. Examples of physical biometrics include face, fingerprint, hand geometry, palm print, iris and DNA. A brief introduction to these biometrics is given below:

• Face: Facial appearance is used as a primary means for recognizing humans. Face recognition is the most acceptable tool of biometrics because of its non-intrusive method and naturalness. The most popular approaches to face recognition (Stan and Jain, 2005)are based on either (i) the location and shape of facial attribute, such as eye, eyebrows, nose, lips, chin, and their spatial relationships, or (ii) the overall (global) analysis of the face image that represents a face as a weighted combination of a number of canonical faces. In order for a facial recognition system to work properly, it should automatically (i) detect whether a face is present in the acquired image (ii) locate the face, if it is present; and (iii) identification of the face from a general viewpoint (i.e. from any pose) under different lighting conditions. The face recognition system is also influenced by variations in facial expressions.
• Fingerprint: Fingerprinting has a long history as an important tool of biometrics. A fingerprint is the pattern of friction ridges and valleys on the surface of a fingertip. The formation of fingerprint is determined during first seven months of foetal development. Each individual has a unique pattern of fingerprint. Even the identical twins do not have the similar pattern of finger prints. The types of information that can be collected from a fingerprint friction ridge impression include the flow of the friction ridges, the presence and the absence of features along the individual friction ridge path and their sequence, and the intricate detail of a single ridge. The process of fingerprint feature extraction typically starts by examining the quality of the input image. Multiple fingerprints of a person (e.g. ten digits used in  integrated automatic fingerprint identification system) provide additional information to allow for large scale identification involving multiple of identities. The process of fingerprint feature extraction typically starts by examining the quality of the input image. The local ridge orientation is then used to tune and filter parameters for image enhancement and ridge segmentation. From the segmented ridges, a thinned image is compared to locate the minima features. The machine representation of a fingerprint is critical to the success of the matching algorithm (Bolle et al., 2004).
• Hand Geometry: Hand geometric biometric system consists of a number of measurements taken from the human hand, including its shape, size of palm, and the lengths and widths of the fingers. This technique is simple, easy to use and inexpensive. The authentication accuracy of hand geometry is not influenced by the environmental factors such as dry weather but show variations during the growth period of children. In addition, an individual’s jewellery (e.g. rings), or limitations in dexterity (e.g. from arthritis), may pose challenges in extracting the correct hand geometry information. The size of hand geometry biometric is large and it cannot be embedded in small screen devices. A number of other authentication system is also available that are based on the measurements of a few fingers such as index and middle finger. These devices are smaller than the hand geometry devices and are easy to use.
• Palm prints: The palms of the human hands also contain unique pattern of valley and ridges. The area of palm is much larger than the area of a finger, and as a result, palmprints are expected to be even more distinctive than fingerprints (Zhang et al., 2003). Palms of the human hands consists of many unique features that can be used for personal identification such as principle lines, wrinkles, ridges, minutiae points, singular points and texture. These features can be extracted at different image resolutions. For civil and commercial applications, low resolution palm print devices are more suitable due to their small size and shorter computation time
• Iris: Iris texture is a unique and reliable biometric trait for personal identification. The iris is the annular region of the eye bounded by the pupil and sclera (white part of the eye). It is formed during foetal development and stabilizes during first two years of life. The accuracy and speed of currently deployed iris-based recognition system is promising and support the feasibility of large scale identification system based on iris information. It is also possible to detect contact lenses printed with a fake iris (Daugman, 1999). Every individual has a unique pattern of iris and even the identical twins do not have the similar patterns
• Retina: Retina recognition seeks to identify a person by comparing the pattern of blood vessels in the back of the human eye. Retina is protected in an eye itself and it is not easy to change or replicate the retinal vasculature.

The retinal biometric system requires the eye to be placed close to the sensor, therefore it also implies cooperation of the subject. This system uses a camera with an electromechanical sensor which measures from a short distance (<3 cm) the natural reflective and absorption properties of the retina. A 7 –mW bulb illuminates the retina and the vein pattern is recorded with visible light and near infrared. (Bolle et al., 2004)

• Ear: Ear were originally a part of the Bertillon system of human identification. The shape of the ear and structure of cartilaginous tissue of the pinna are distinctive. Matching the distance of salient points on the pinna from a landmark of the ear is used as a method of recognition. Ear recognition is not as reliable as face recognition. The combined recognition based on ear and face offers improved accuracy in identification.
• DNA: DNA (Deoxyribose nucleic acid) identification (National Research council, 1992 & 1996) is often cited as ultimate biometrics, since DNA codes identify information in a digital form that is available from every cell in the body. The basis of DNA identification is the comparison of alternate forms (alleles) of DNA sequences found at identifiable points (loci) in nuclear genetic material. For identification, a set of loci is examined to determine which alleles are present at each chromosome. Any difference between enrolled and test samples indicates identity difference, while consistent identity between the samples indicate identity or a coincidence whose probability can be determined (Bolle et al., 2004).

This method has some drawbacks: (a) contamination and sensitivity, since it is easy to steal a piece of DNA from an individual and use it for an un-disclosed purpose (b) no real-time application is possible because DNA matching requires complex chemical methods involving expert’s skills, (c) privacy issues since DNA sample taken from an individual is likely to show susceptibility of a person to some disease (Delac and Grgic, 2004).

Behavioural Biometrics:

Behavioural biometrics is based on the behavioural trait of an individual. The most common behavioural biometrics is signature, voice, gait and key stroke recognition. Brief introductions to this biometrics are given:

• Signature: Signature is a simple, concrete expression of unique variations in human hand geometry. The way a person signs his or her name is known to be characteristic of that individual. Although signature requires contact with the writing instrument and an effort on the part of the user, they have been accepted in government, legal and commercial transaction as a method of authentication. With the proliferation of PDAs and tablet PCs, online signature may emerge as the biometric of choice in these devices. Signatures are a behavioural biometric that change over a period of time and are influenced by physical and emotional conditions of a subject. In addition to the general shape of the signed name, a signature recognition system can also measure pressure and velocity of the point of the stylus across the sensor pad (Jain et al., 2007).
• Voice: Voice is a combination of both physical and behavioural biometric characteristics (Campbell, 1997). An individual’s voice is based on the physical characteristics such as vocal tracts, mouth, nasal cavities and lips that are used in creating a sound. These physical biometric characteristics are invariant but the behavioural aspect of human speech changes over time due to age, medical conditions e.g. cold, emotional state etc. Voice is also not very unique and appropriate in large scale identification. It has two approaches: Text dependent (recognition based on the fixed predetermined phrases) and text independent (recognition is independent of what a person is speaking). A text independent system is more difficult to design than text dependent voice recognition system. Speaker recognition is most useful in telephone based applications but the voice signal is typically degraded in the quality by the communication channel.
• Gait: Gait refers to the peculiar way one walks and it is a complex spatio-temporal biometrics. It can be used to identify a person from a distant point. Therefore, this biometric is appropriate in surveillance scenario where the identity of a person can be superstitiously established. Recognition based on gait is somewhat newer biometrics and needs to be researched in detail. Gait is a behavioural biometric and is influenced by a number of factors such as body weight, walking surface, footwear, nature of clothing etc.
• Keystroke: It is believed that each individual types on a keyboard in a unique way. This biometric is also not very distinctive and unique in identification but assists in recognition of an individual by offering sufficient discriminatory information. Keystroke pattern is also influenced by emotional state, keyboard position, type of keyboard etc. Advantage of using keystroke behaviour for recognition is that it can be easily observed unobtrusively as that person is keying the information. This biometric permits “continuous verification” of an individual’s identity over a session after the person logs in using a stronger biometric such as fingerprint or iris.
• Odour: Each object spreads around an odour that is characteristic of its chemical composition and this could be used for distinguishing various objects. This would be done with an array of chemical sensors, each sensitive to a certain group of compounds. Deodorants and perfumes could lower the distinctiveness. (Delac and Grgic, 2004)

It is obvious that no single biometric is the “ultimate” recognition tool and the choice depends on the application. A brief comparison of the above techniques based on seven factors described in section 2 is provided in Table 1.

Table 1: Comparison of Various Biometric technologies (Jain et al., 2004)

Multimodal Biometrics System:

This system allows the integration of two or more types of biometric recognition and verification systems in order to enhance performance requirements of identification. These systems are more reliable due to the presence of multiple biometric evidences. A multimodal system could be, for instance, a combination of fingerprint verification, face recognition, voice verification and smart-card or any other combination of biometrics. This enhanced structure takes advantage of the proficiency of each individual biometric and can be used to overcome some of the limitations of a single biometric (Delac and Grgic, 2004).

Applications of Biometrics:

Biometrics is being increasingly used in several different applications. The major applications are given below:

Table 2: Application of Biometrics

Summary:

Traditional system of identification is limited in their ability to address various identity issues. The advent of biometrics has served to overcome the shortcomings of traditional personal identification issues. It refers to an automatic recognition of a person based on her physiological and behavioural characteristics such as face, fingerprint, hand geometry, palm print, iris, retina, ear, DNA, signature, voice, gait, keystroke, odour etc. A broad spectrum of establishments can engage the services of a biometric system including travel and transportation, financial institutions, health care, law enforcement agencies, and various government sectors. Similarly, the concept of multi-biometrics is also improving the identification performance of biometric based recognition. In short, biometric recognition will have a great influence on the way we conduct our daily business in near future.