Introduction

Overview of the Previous Article

Our last article reviewed what Multimodal Biometric Solutions are all about, and what the implications are for deploying such a type of security system. Essentially, this type of configuration involves implementing two or more Biometric modalities in such a fashion that it provides multiple layers of security for the corporation.

Remember, one of the oldest rules of network security is never to rely on just one means of defense. More than one layer is needed. In this regard, two types of Multimodal Biometric Solutions were also reviewed: 1) Asynchronous, and 2) Synchronous.

With the former, two or more Biometric modalities are used in a sequential fashion; and with the latter, two Biometric modalities are used together, at the same time, to confirm the identity of an individual.

The Biometric Technologies of Today

There are two main classifications of Biometric Technology:

  • Physiological based Biometrics;
  • Behavioral Based Biometrics.

The types of modalities which fall underneath these categories are as follows:

Physiological based Biometrics

  • Fingerprint Recognition:

    This examines the minutiae which are found in the breaks and discontinuities located in the whorls, valleys, and ridges of the fingerprint.

  • Hand Geometry Recognition;

    With this, the geometrical features of the hand are examined. This includes analyzing the distance between the knuckles and the fingertips, and the distance to the palm.

  • Iris/Retinal Recognition:

    These two modalities are often confused with one another, and in reality, they are quite different. The retina is located in the back of the eye, and the iris is in the front of the eye. With the former, the pattern of blood vessels which form the retina going into the optic nerve are identified, and with the latter, the vector orientations of furrows and freckles in the iris are examined.

  • Facial Recognition:

    Like that of Hand Geometry Recognition, this modality examines the distances between the prominent features of the face, which includes the eyes, eyebrows, nose, lips, and chin.

  • Voice Recognition:

    This examines the changes in the inflections and pitch in one’s voice, as he or she speaks.

  • Vein Pattern Recognition:

    Also like Retinal Recognition, this modality identifies the pattern of blood vessels just underneath the palm or the fingertip.

Behavioral based Biometrics

  • Keystroke Recognition:

    This measures the rhythmic pattern when typing on a computer keyboard. For example, variables include the time which elapses when the keys are held down and released, the various typing sequences, the speed of typing, etc.

  • Signature Recognition:

    This modality examines the mannerisms in which a signature is signed. This includes the way the pen is held, the pressure which is held from the hand onto the pen while the signature is being signed, any time intervals during the signing process, etc.

The Biometric Technologies of the Future

The above mentioned Biometric Technologies are currently available in the marketplace today. But, just like other security tools, there are advancements being made to Biometrics every day. This includes enhancements to the hardware, the Software Development Kits, and APIs, and even the mathematical algorithms themselves.

There are three potential Biometric modalities which are under heavy research and development at present. They are as follows:

  • DNA Recognition;
  • Earlobe Recognition;
  • Gait Recognition.

DNA Recognition

The goal of DNA Recognition, as the name implies, is to confirm the identity of an individual based upon the unique strands of DNA which he or she possesses. There are two kinds of DNA which exist in all living human beings:

  • The Nuclear DNA:

    This is the contribution of the genetic components from both the mother and the father which makes up the DNA of the offspring. This type of DNA is found in the blood, semen, saliva, and even in the bones of an individual.

  • The Mitochondrial DNA:

    This is the genetic contribution from the mother only and is found in the hair and teeth of an individual.

To create the Biometric Template, the repetitive patterns of the four base pairs are utilized from either of the DNA mentioned above components. The base pairs are Adenine (A), Cytosine (C), Guanine (G), and Thymine (T). Specifically, it is the unique features from the uncoded portion of the DNA sequence which is captured.

This uncoded portion can also be referred to as a “Short Tandem Repeat”, or an “STR” for short. A highly differentiated STR exists when two or more base pairs appear as a distinguishable and repetitive pattern. After this is collected, the DNA sample is subsequently subdivided into much shorter segments, and then placed onto a nitrocellulose filter.

This specialized filter consists of various fluorescent dyes and is then x-rayed. The resultant image becomes what is known as the “Digital DNA Fingerprint.” This theoretically becomes the Enrollment Template and is stored in a DNA Recognition device. The same process would then have to repeat to create the Verification Template.

Earlobe Recognition

The scientific methodology behind Earlobe Recognition can be considered to be the same as that of Hand Geometry Recognition. In fact, using the structure of the ear as a means to identify an individual goes back all the way back to the 19th century. Also, the first ear classification system was proposed in 1964 by the scientist Alfred Iannerreli.

Earlobe Recognition involves measuring the distances between the prominent landmarks of the ear, as well as other unique features. Current research is primarily focused on building and refining the concepts of the “Eigenear”. This involves examining the structure of the outer ear, such as the depth, the angular structure, and the overall crescent shape.

With the use of the of the Eigenear, various 45 degree and 3-D angular models of the outer ear can be constructed. There are two specific techniques which are being examined to create the Biometric Template of the ear:

  • PCA:

    With this, the raw image of the outer ear is cropped and then scaled back to a regular size. In the end, two major features of the outer ear are captured, specifically the triangular fossa and the antitragus. From this point, the unique features are then extracted to create both the Enrollment and Verification templates.

  • 3-D Analysis:

    Under this technique, the image of the outer ear is captured (which includes both the depth and the color type). The helix and the antihelix are also carefully analyzed, from which the unique features are extracted to create both the Enrollment and Verification Templates.

Gait Recognition

Gait Recognition is the study of the unique ways in which people walk. On a theoretical level, it can be considered to be both a Physical and Behavioral based Biometric. In fact, it is a subset of the field known as “Kinematics,” which is the study of motion.

The unique features which can be extracted from an individual’s stride include the various angles, lengths, and speeds when he or she is walking on a plane. At present, there are two techniques which are being examined for capturing these particular traits. They are as follows:

  • Static Shapes:

    With this, a silhouette image of the individual walking is extracted from a series of video frames taken by a CCTV camera. From this, a digital map is composed, which is also known as the “Gait Signature.”

  • Dynamic Shapes:

    This type of technique makes use of an accelerometer, which is placed on the individual’s leg. From here, the unique features of the stride can be captured and analyzed, at three different vector based distances.

Ethical Hacking Training – Resources (InfoSec)

After unique feature extraction, the next step is the creation of both the Enrollment and Verification Templates. There are three methodologies which are being closely reviewed by researchers. These can be described as:

  • Direct Matching:

    A specific Gait Sequence is calculated by comparing the unique features identified in the stride of the individual, which are also associated with certain reference data. At this point, the least minimal distances are then computed and compared with one another.

  • Dynamic Time Warping:

    The statistical closeness between two sequences of gait movements is calculated and compared with one another. In this technique, the examined variables include the time, speed, and the various stride patterns of the individual.

  • Hidden Markov Models (also known as HMMs):

    With this, various statistical profiles are created to examine the probability of similarity between the different shapes which appear in the walking stride of the individual.

Conclusions

It should be noted that these “Biometric Technologies of the Future” are still in the mid stages of research and development. Some are further along in this process than others, and none of them have proven to be viable security solutions yet which can be procured and implemented. Each of them has their own strengths and weaknesses which can be outlined as follows:

  • DNA Recognition:

    The largest advantage is that the DNA of an individual possesses the richest and most unique features, and as a result, it will provide the most irrefutable proof in confirming that particular identity. In fact, if DNA Recognition makes its way to be a viable modality, it will be considered to be the ultimate Biometric of all. But despite this, it also possesses its own set of weaknesses as well. For example, the analysis which is required in DNA Recognition can take hours. To be a proven technology, this has to come down to a matter of two seconds or less in performing the Verification and/or Identification transactions. Also, unlike the other Physical Biometrics, DNA is prone to rapid degradation if exposed to the outside environment. Because the use of DNA is often correlated with law enforcement and the judicial system here in the United States, it is very much prone to privacy rights issues and claims of civil liberties violations.

  • Gait Recognition:

    One of the strongest advantages of using the unique walking stride of an individual is that it can be used in a very covert fashion. For example, if it proves to be a viable Biometric modality, it will be a very valuable security tool in both international airport settings and other large-scale transportation hubs. It can also be considered to be a non-contactless type of Biometric, which will be a catalyst in its acceptance rate with society in general. But, the walking stride can be greatly affected by such variables as the type of surface the individual is walking on, extra weight gains or weight loss that the individual may experience over time, older age, physical ailments, and even diseases. In other words, the walking stride is not considered to be a stable feature.

  • Earlobe Recognition:

    Regarding strengths, various scientific studies have shown that most individuals do possess a unique ear structure. Also, while the ear is exposed to the extremities of harsh weather (especially in cold temperatures), it is still considered to be a rather stable component of the human body. But, collecting a quality based raw image of the ear can be a challenge, especially if jewelry (such as earrings) and hearing devices (such as a hearing aid, or a Bluetooth device) are worn, or even if the lighting conditions are poor from the external environment. There is no direct contact required with the ear to capture the raw images, thus like Gait Recognition; it too can be considered a non-contactless type of Biometric (on a theoretical basis).

When these three potential modalities are compared to each other on a spectrum, it is Gait Recognition which has advanced the furthest regarding research and development. Thus, it holds the greatest potential to enter the marketplace and serve security applications in the short term.

Sources

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3285485/

http://nar.oxfordjournals.org/content/32/22/6673.full

http://thesis.library.caltech.edu/1720/2/(02)ch1_Title_text.pdf

http://www.cse.msu.edu/~rossarun/pubs/RossEarRecognition_IEEECOMP2011.pdf

http://www.wired.com/2010/11/ears-biometric-identification/

https://www.researchgate.net/publication/303671134_Human_Identity_Recognition_Using_Ear_Lobes_Geometric_Features

http://pages.cpsc.ucalgary.ca/~boyd/papers/biometric-summerschool.pdf

http://airccse.org/journal/jcsit/0210ijcsit4.pdf

http://ttic.uchicago.edu/~karthik/gait.pdf