Biometrics

The oldest method of authentication and so is largely accepted as a biometric, is signature. There are two subtypes of signature verification systems, static and dynamic. In the later subtype, speed, velocity, pressure, angle of the pen and the number of times the pen is lifted from the pad, is be measured while in static subtype only the image of the signature is used. The dynamic signature verification is more secure and reliable than static signatures. Shortcomings of signature biometrics include inconsistency, for example, signatures lack permanence, which means that may change under the influence of illness, emotions, age, etc. These systems render performance only in verification mode and not in identification mode.

Fingerprint verification has the advantage that no two individuals possess the same fingerprints, not even identical twins, though suffers from a few disadvantages. Dry, wet, damaged or dirty skin may affect the quality of the fingerprint. With fingerprints, the attacking technology is as easy as the defending technology, this fact has been proven by many successful security attacks. Tsutomu Matsumoto of the Yokohama National University victoriously counterfeited numerous fingerprint based biometric systems into accepting fake fingers made of gelatin gaining an 80% success rate. It is also difficult to acquire fingerprint features for some classes of people like manual laborers or elderly people. In spite of these shortcomings, modern fingerprint recognition algorithms are researched because of their applicability and entailment. For instance, almost always in forensic scenarios latent prints are the only trace for fraud identifications.

As palmprint comprises wider area than any other biometric traits, it can be used even in fallacious conditions like burns, boils, cuts, dirt and oil stains on palms. Also when fading of palm texture occurs due to lot of physical work with hands.

Hand Geometry systems measure demographic aspects like thickness and width of the palm, length and width of fingers, and so on. Because of its adaptability, ease of measurement and storage, hand geometry based systems are highly acceptable than finger-print based systems. But because of its biometric properties it is only suitable for verification and fails in identification mode. There is also a high number of false positives. It always has to be associated with some other biometric trait for perfect authentication.

Hand Vascular Pattern makes use of infrared light to produce an image of a person's vein pattern in their face, wrist or hand, as veins are stable through one's life. The vein pattern recorded by any device like video camera is used as a personal code which is acutely laborious to duplicate. The fact that the use of this biometric needs no physical contact with the sensor and that it provides notable convenience and no performance degradation even with scars or hand contamination makes this physiological biometric a reliable one. 

From facile edge-based algorithms to advanced pattern recognition methods, a wide range of techniques have been proposed for face recognition. Numerous existing face recognition techniques succeed with frontal faces of similar sizes and even with distorted facial images. While in reality, this presumption may not hold good as human face is dynamic in nature hence has a high degree of variability in its appearance, making face detection an intricate problem in computer vision. Factors such as changing hairstyles, beard, mustache and aging only make righteous face recognition more difficult. Bruce Schneier, in his book "Beyond Fear" calculated the math and stated that if a face detection system is 99,9% accurate, still it would generate 10.000 false alarms for every single real terrorist in 10 million civilians.

Iris is a thin annular structure around the pupil of the human eye. Its complex pattern is constructed of many idiosyncratic features such as fibers, freckles, furrows, arching ligaments, ridges, serpentine vasculature, rings, rifts and corona. All these establish a distinctive signature for human authentication. Patterns in human iris have abundance of invariance. Iris patterns emerge during the eighth month of the fetal term and remain stable throughout the life time of an individual. On the lines of precision, report successful authentication across millions of cases without a single failed test. Given its non-invasive nature and affordable hardware solutions, iris based authentication systems have become an indispensable tool for many high-security applications. Apparently, both iris and retina recognition would not work for visually-impaired people and people suffering from serious eye illnesses.

Retinal scanning based systems use infrared illumination while acquisition and compare images of the blood vessels in the back of the eye, the choroidal vasculature. Apparently, retina recognition not only is futile for people suffering from serious eye illnesses but also raises privacy issues in case of misuse of acquired data.

Complex Eye Movements is a very recent biometric trait which was brought to light by Komogortsev et al. and Kasprowski et al. They carried out considerable research on CEM and established commendable results. In Complex Eye Movements are combined with Oculomotor Plant Characteristics where a mathematical model for eye and its associated muscle movement is established when eyes respond to a stimuli. This Biometric is still in its infancy but seems to be a non-counterfietable Biometric. But in regard to its ease-to-acquire and easy-to-use, justification is still void.

Voice verification identifies myriad characteristics of a human voice like frequency, nasal tone, cadence, inflection to recognize the speaker. Voice recognition systems take advantage that they do not require expensive input devices and can even accomplish the recognition task in the background while the person speaks without explicitly forcing the users to spend time to do the same. But like all other biometrics, voice systems have their fair share of shortcomings, for instance, record and play attacks in fixed-text models, also some people might skillfully duplicate/imitate others' voices. Voice of an individual may also change with age, illness, mental state, etc.

Gait is the pattern of movement of the limbs of animals, including humans, during locomotion over a solid substrate. Examples of gaits include jogging, running, walking, jumping, sitting down, picking up an object or climbing stairs. According to the performance of gait recognition systems is below what is required for use in biometrics as this biometric recognition system is confounded by the following factors viz, terrain, injury, footwear, any kind of training to the human body, passage of time.

Body Odor is a contact-less biometric that confirms a person's identity by analyzing the olfactory properties of the human body scent. Cambridge university has developed electronic sensors to gather the human odor, usually from the non-intrusive areas, such as the back of the hand. Each human smell is made up of chemicals known as volatiles. Each chemical of the human odor is extracted by the biometric recognition system and converted into a unique data string. But privacy of the individual will be compromised while using this biometric as body odor carries an amount of sensitive personal information.

DNA 99.7% of human DNA is shared. 0.3% (1 million nucleotides) is variable and so is unique. These variable regions, called Short Tandem Repeats (or STRs), can be examined to distinguish one person from another. DNA samples can be isolated from a sample such as saliva, blood, hair, tissue or semen. But it suffers from the following complications: (i) DNA matching is not done in real-time, a physical sample must be taken unlike other biometric systems which use an image or a recording, (ii) invasion of civil liberties, (iii) storage of DNA and (iv) extraction and process time. DNA based biometric system cannot be easily simulated but is invasive and arduous to setup.

Excerpt from: "Bio-inspiring Cyber Security and Cloud Services: Trends and Innovations", by Aboul Ella Hassanien, Tai-Hoon Kim, Janusz Kacprzyk, Ali Ismail Awad  - January 1, 2014 - Springer