Biometric Facial Recognition Database Systems

by Robert E. Vanaman

 For millenniums, humans have availed themselves of the body characteristics of one another such as face, speech, gait, size, etc. to recognize each other. This instinctual differentiation has evolved into the science of Biometrics. Biometrics is nothing more than a form of bioinformatics1 that applies biological properties in the identification of individuals (Animetrice, 2008a). Facial recognition (FR) is a subset of the broader science of Biometrics. A biometric systemis effectively a pattern recognition system that operates by acquiring biometric data from an individual, and extracts a feature set from the acquired data for comparison purposes (Jain, Prabhakar, & Ross, 2004). The information needed for recognition is acquired by a sensor, and is converted into a digital format (, 2011a). This digitized representation of a feature, in this case a face, is then compared to a “biometric template” (, 2011b) or a “gallery” (Bronstein, Bronstein, & Kimmel, 2004) stored in a database. This paper will delve into the Facial Recognition Database Systems (FRDBS) currently in place and cover predictions for future use, exploring the processes and methodology employed therein, specifically addressing FRDBS methodologies and techniques employed in capturing, storing, and comparing scanned images.

Historical Background

The term biometrics is derived from the Greek words “bio” meaning life and “metrics” implying to measure (, 2011c). Although the advent of modern (Read: computerized) biometric systems have been possible only in the most recent times, the concept of using body characteristic to distinguish one another was conceived hundreds, perhaps thousands of years ago. As evidence of this assertion, handprints found in caves estimated to be over 31,000 years old, are believed to represent “un-forgeable signatures” (Renaghan, 1997) of their authors. Evidence that fingerprints were used in Babylonia in 500 B.C. can be found in the clay tablets of the day used to record business transactions (International Institute of Hand Analysis, 2005).

In Europe, with the emergence of a more mobile population in the mid-18th century, justice systems were tasked with tracking and cataloging first-time, verses repeat, offenders across a larger sampling of populous as well as in a growing geographic area. A formal system was needed, and devised to accomplish this. First there was the Bertillon system in France, which measured various body dimensions. This approach was called anthropometrics. The second method, which is still in use today, was a card file of indexed fingerprints, which was discovered to be exclusive on individuals, based on the patterns and ridges found on fingers (, 2011c). Today, biometric identification and verification systems have proliferated beyond 19th century fingerprinting and palm printing indexed card files (one of the first scientific application of biometrics) to include 21st century computerized DNA, ear, gait, hand geometry, hand vein, iris, keystroke, odor, retina, signature, and speech recognition processes, as well as FRDBS (Jain, Prabhakar, & Ross, 2004).

It should be noted early in our discussion that FR systems are broadly divided into two distinct classifications: authentication and recognition (Bronstein et al., 2004). Authentication is the process of comparing a claimed identity with a designated template in the gallery. Here, the FR algorithm is tasked with only a one to one (1:1) relationship comparison. Furthermore, in an authentication scenario, the enrolled individual is “assumed to be collaborative” (Bronstein et al., 2004). In stark contrast, the processes involved in recognitions are far more rigorous. Here, the FR algorithm must compare the subject with all the templates in the database. This entails a one too many (1:M) relationship matching, a far more arduous undertaking. Compounding this undertaking is that a collaborative demeanor on the part of the subject is highly unlikely.


Qualifiable Constraints

Before the techniques and methods of the comparisons of images to templates can be examined, an understanding of what qualifies as a measureable biological biometric needs defining. Any number of “human physiological and/or behavioral characteristics” (Jain, Prabhakar, & Ross, 2004) can be engaged as long as they comply with the following four constraints:

  • Universality: everyone has this characteristic or trait.

  • Distinctiveness: One individual’s characteristic or trait must be sufficiently dissimilar from other individuals.

  • Permanence: the characteristic or trait should be persistent over time with respect to its measurable biometric criterion.

  • Collectability: characteristic or trait can be analyzed – measured – quantitatively (Jain, Prabhakar, & Ross, 2004).


As with most theoretical scenarios, if a practical field based operation is to be deployed, several other constraints which could place limitations regarding the system’s practicality must be addressed. These would include, though not limited to:

  • Performance: this criterion is composed of two elements. First, the ability of the system to make accurate comparisons; that is to minimize false matches and false non-matches. Second, the speed at which the operation renders an acceptable and reliable result.

  • Acceptability: this is a measurable indicator of the extent that a particular biometric system is embraced – or at minimum – tolerated in the public’s daily life.

  • Circumvention: a reflection of the robustness of the system’s ability to resist unlawful criminal acts of evasion (Jain, Prabhakar, & Ross, 2004).

To summarize, a feasible biometric scheme should meet or exceed the specified recognition precision and speed parameters, be innocuous to the scanners and ones being scanned, be sufficiently benign to be adopted by the selected group of the populous intended for its use, and be sufficiently vigorous in thwarting various devious methods of attack on the reliability of the

system (Jain, Prabhakar, & Ross, 2004). Table 1 is a synopsis for a comparison of the aforementioned Biometric recognition processes and their corresponding effectiveness in respect to the constraints.


Advantages and Disadvantages

Why favor FR as the viable approach for identification over other biometric methods? First, FR is stealthy; it does not require the active participation of the participant, and may be acquired from a distance. Concisely, FR is “unobtrusive and discrete” (Animetrice, 2008a) and recognized as a natural, non-intimidating and universally accepted biometric identification method (Bronstein et al., 2004). Second, the infrastructure necessary for the system to operate is already widely in place, and relatively inexpensive. Security cameras are a common fixture from airports and retail establishments to ATMs and private residences. Lastly, a multitude of private companies have stored “photo ID records” (Animetrice, 2008a) and virtually every governmental and intelligence agency has vast quantities of surveillance photos and videos in legacy database repositories.

Historically the fundamental drawback to FRDBS has been the 2D personality of the stored image. These two dimensional (2D) representations of a three dimensional (3D) face introduce unacceptable failure rates. A 2D view measures only the height and width, with a corresponding measurement between facial features. Additionally, faces reflect light casting shadows, faces change with expressions or head pose, “facial hair, the use of cosmetics, jewelry and piercings” (Bronstein et al., 2004) all have their influence, thereby creating a different visual persona under different circumstances (Animetrice, 2008a).

A partial solution to this quandary came after the 2002 Defense Advanced Research Projects Agency (DARPA) and the National Institute of Standards and Technology (NIST) Face Recognition Vendor Test (FRVT) initiative that concluded that a 3D FR system was the solution. 3D FR systems address the issues of variable circumstances and can be implemented without the use of specialized cameras, thereby utilizing existing infrastructure equipment (Animetrice, 2008a).

3D FRDBS Design

3D FR Systems is a comparatively fresh approach in the science of Biometrics. It, in many ways, breaks with the long-term approach of attempting to interpret human facial features by employing the traditional human visual recognition methodology – which is utilized in 2D models – that have yielded such limited success. These results impart, to the attributed graphic flat representation of the compound curved, 3D facial surface (Bronstein et al., 2004). 3D facial geometry systems accomplishes its mission by representing the internal anatomical structure of the face as an alternative to the external representation utilized by 2D models that are susceptible to various environmental considerations (Bronstein et al., 2004). Consequently, 3D FR systems are insensitive to diverse illumination backgrounds and the interfering effects of cosmetics and head pose.

A fundamental postulate of 3D FR systems is that once facial features are modeled as isometrics2 in the context of Riemannian geometry3 facial surfaces are regarded as deformable objects. The advantage of this approach results in the inherent geometric characteristics of the facial surface becoming “expression-invariant” (Bronstein et al., 2004). This hypothesis of facial expressions being invariant once modeled as isometrics, juxtaposed upon the framework of Riemannian geometry, can be proven quantitatively by locating a set of attribute points on the facial contour, and calculating how the expanse between them is altered due to facial expressions. In 2004, Alexander and Michael Bronstein, in concert with Ron Kimmel, conducted an experiment that placed “133 markers on a face and tracked how the distances between these points change due to facial expressions” (Bronstein et al., 2004). They concluded that the divergence from of the geodesic4 distances due to facial expressions is insignificant (which justifies our model), and that the discrepancy is less half than the applicable change of the Euclidean model. This signifies that the isometric model portrays the character of facial expressions over as twice accurately, when compared to the inflexible 2D systems (see graphic 1) (Bronstein et al., 2004).


Normalized histogram of the absolute error of the geodesic (solid) and the Euclidean (dashed) distances.


A 3D Prototype

At the Geometric Image Processing laboratory a prototype 3D FR system, based on the expression-invariant representation of facial surfaces was developed and tested. This prototype is capable of operating in the 1:1 authentication mode, and in the more demanding 1:M recognition methodology. An AMD Opterron64 workstation, utilizing a Microsoft Windows XP operating system, was employed in the data processing and data analysis, as well as managing the graphical user interface for commands and visualization (Bronstein et al., 2004).

The data processing and analysis was disseminated into several discrete phases. First, the face is scanned in three dimensions, producing a shroud of data points corresponding to the facial surface. This pseudo-image is then “cropped, smoothed and subsampled” (Bronstein et al., 2004). The next steps involve a feature detector that isolates a few critical standard points for comparison. These points then result in a computed geodesic mask. Finally, the facial surface undergoes a canonization5 utilizing a multidimensional scaling6 (MDS).

The conclusions presented in this study confirmed that this 3D FR system, relying on models as isometrics in the configuration of Riemannian geometry, and then processed – massaged – in a perspective of canonization incorporating MDS, comprehensively out performed traditional 2D and 3D systems. This system’s expression-invariant nature asserts its relevance in commercial applications, which pre-conceives a less then collaborative subject, in various environments and situational circumstances. Additionally, a high level of public acceptability would exist due to the clandestine nature of the scanning process. A further benefit of utilizing a canonical forms methodology is their irreversibility. Canonical forms cannot be reversed -engineered back – to a facial surface features that would be recognizable in traditional human identification images. It essentially hides the actual identification of the subject in the database gallery. This feature has a prominent benefit in commercial systems, where the security of biometrics data is of paramount importance.

Storage and Retrieval

An Elementary System

One form of FRDBS is utilized primarily in criminal and forensic identification, but has intelligence and broader-security applications in which subjects need to be authenticated against passport or other recognition, documentation, or credential photographs (Forensic Science Communications, 2008). Here the FBI, in conjunction with the Magna Science AdventureCentre in Rotherham England, by means of digital stereo photography (Geometrix FaceVision FV802 Series Biometric Camera) and 3-D laser scanning (Cyberware 3030PS Head and Neck Scanner) equipment captured 30 craniofacial anthropometric landmark sites (see appendix A. Table 2) (Forensic Science Communications, 2008) on the 3D images of 3115 volunteers. These images were captured with the “Geometrix FaceVision system which (is) based on eight digital cameras” (Forensic Science Communications, 2008). The 30 landmark sites are mapped on one of the photographs in Image 1.


The Geometrix database contained 3115 folders and was a combination of a Microsoft Excel 2003 spreadsheet where each row (tuple) contained attributes (columns) including the age, gender, and ancestry of the subject and including the 3D Cartesian coordinates7 of each of the landmarks. There was untouched stereophotographic image data, including eight JFIF images that corresponded to each of the eight cameras. The 3D facial wireframe models (see appendix A. figure 1) (Forensic Science Communications, 2008) and the textured-mapped surface data images (see appendix A. figure 2) (Forensic Science Communications, 2008) were stored in three formats, DXF, AutoCad and Virtual Reality Modeling Language (VRML) (Forensic Science Communications, 2008). As of the publication date of this article, the Geometrix database was considered the largest such collection in existence (Forensic Science Communications, 2008).

An Advanced System

A more advance in FRDBS has recently emerged in Japan. This hybrid-system utilizes both 2D and 3D facial images superimposed (Clement & Marks, 2005). Although this application is primarily a 1:1 authentication system, it promises future refinements to allow for 1:M recognition. The key refinement in this approach is that the two facial images can be oriented by the rotation of the 3D model. Furthermore the system draws upon morphometric8 matching using facial outlines and anatomical landmarks. This methodology as before, provides results based on empirical, numerical data. Commercially available software packages thus far are limited to frontal image comparisons under current two dimensional technology. In contrast, this robust system for FR comparison allows for “severely disadvantageous angles” (Clement & Marks, 2005) of comparisons thereby rendering it effective in a structured database and real-world environment.

In the storage phase of the FRDBS, a software package called “3D-Rugle3 for Face-To-Face” (Medic Engineering, Kyoto, Japan) was used for “automatically adjusting the facial orientation of all the 3D images in the database” (Clement & Marks, 2005). Once this procedure was accomplished, all the 3D images were stored within the system and normalized in the orientation to the “Frankford horizontal9 plane”. Next, fourteen orienting landmarks were mapped on the 3D facial image, and their discrete distances and coordinates were calculated and stored (see appendix A. figure 3) (Clement & Marks, 2005). Correspondingly, the 2D facial images were subjected to identical treatment (see appendix A. figure 4) (Clement & Marks, 2005). Again, discrete distances and coordinates were calculated and stored. Once specific parameters for rotation angle and facial dimensions were plotted from the anatomical landmarks, comparisons from the two images could be accomplished by superimposing a single 2D image upon multiple 3D images (Clement & Marks, 2005).

In the retrieval phase of the process, a “Commercial off-the-shelf software (COTS)” (Kendall, & Kendall, 2011) call “FaceList” developed by OMRON Co. of Japan, was substantially modified in its robustness when tasked in appraising variations in facial ordination. This package was then employed to identify the target subject’s 2D image against the stored 3D images in the FRDBS. This customized software is largely founded on the graph-matching methodology. Here, images are first overlaid with a course matching algorithmic graph of fixed parameters, followed by progressively finer matching functions of dynamic design. Once an acceptable graph is defined and selected, it contains fifty nodes with the distances between each and their illuminated grid properties calculated and accounted for (see appendix A. figure 5) (Clement & Marks, 2005).

FRDBS Future

A stunning example of FRDBS practical application at work now, and destined to affect the future, is the development and deployment of a new biometric security system, Broadway 3D, that recently concluded tests at Moscow’s Sheremetyevo International Airport. The system, designed and build by Artec Ventures10, represents a multi-magnitude jump in 3D surface imagery. Broadway 3D boasts numerous advancements that until now were merely items on biometric scientists’ wish lists. First, the system “is highly automated and minimizes the need for human supervision” (HSNW, 2010). Second, during a single one month testing period, over 3,500 individuals were scanned and processed with their 3D images stored in a database gallery. These scans resulted in 100% accuracy rates when templates were compared to database images; no false/positives, no false/negatives. Third, the quality of recognition is unaffected by glasses, hairstyles, facial hair, cosmetics, or head pose. Additionally, due to the controlled environment in which the imagery is captured (the subject passes through a turnstile where the 3D camera emits a light pattern at parallel angles to the face), the illumination considerations are mitigated. The monumental advancement in Broadway 3D’s success rate can be attributed to several factors. First, the 3D image, is composed of 40,000 data points (Image 2) (Artec ID, 2011), compared to a few dozen in previous FR systems. This allows for a vastly more accurate mapping, and thus, a far more precise comparison algorithm of a subject’s template to the stored gallery image in the database (see appendix A. Figures 6-10) (Artec 3D, 2011). Second, the recognition time is a phenomenal one second, with enrollment time in the database of less than three seconds, and the exposure time for the subject of a mere 0.2ms. These unheard of time frames resulted in a throughput capacity, utilizing a turnstile, of up to 30 people per minute. Lastly, the diminutive size of the equipment cannot be over emphasized. The scanning unit measures a five feet four inches, by nine inches square, and weighing a little over thirty-five pounds. This enables the system to be portable, thus flexible, making it deployable and practical in the extreme (Artec ID, 2011).


Having overcome numerous shortcomings of previous FRDBS earns Broadway 3D compliances with all of the qualifications necessary for a measurable biological biometric. The human face has universality, distinctiveness, permanence, and collectability. Additionally, it meets the requirements for a field practical system by exhibiting outstanding performance, with its contactless identification and diminutive size it merits public acceptability, and the systems robustness prevents circumvention measures.


From prehistoric handprints discovered upon cave walls, which bare mute witness to their creator’s existence and demise, too millisecond facial recognition systems at 21st century airports, the science of biometrics has changed the path of human history, and is destined to change the direction of humanity’s future. From politicians, to security specialists, to administrators, all struggle with the real-world consequences and ramifications of biometrics in seeking a balance between an individual’s right to privacy, against that same individual’s right to safety. These juxtaposed apprehensional quandaries will be abolished as the science of biometric facial recognition systems become more reliable and less detectable. As with all modern day sciences, FRDBS are waiting on advancements in computer technology for advancement in their given field. The latter must precede the former. Additionally, as the boundaries of what constitutes an acceptable, viable, measureable, biological biometric shrink, due to the speed, accuracy, and the progression of the system’s miniaturization – all of which continue to astonish – the public’s acceptance will to continue to escalate at an accelerated rate.

Final perceptions: as biometric identification processes become a larger part of everyday life, clearly social concerns, legal issues, and ethical considerations will rise to the forefront of contemporary debates. As the miniaturization and stealth of biometric identifications systems continues to make their omnipresence diminish, their proliferation will only intensify. What is clear is that their governance must be a global endeavor. Any attempt to govern identification technologies will be an international effort, or will be ineffective (Challenge Liberty & Security, 2005). Lastly a word of caution, biometric identification system’s influence on individual freedoms and individual safety can be benevolent or nefarious. It is mankind’s responsibility to remain a faithful Argus of this proto-science. Unleashing its promise for humanities betterment, rather than succumbing to its potential for the demise of both mankind’s liberties and mankind’s protection.



Completed in June of 2013, and deployed at Sochi’s International Airport, Elektronika, LLC – a Russian security system integrator – has developed and implemented an integrated security system for the 2014 Winter Olympics. The system “provides a wide range of functional capabilities including security monitoring and alarm situation detection, as well as its rapid response” (Artic ID, 2013). Over 550 high-definition surveillance cameras, along with checkpoints and automatic alert detection cover Sochi’s entire premises. This unified monitoring system is feed into single dispatch center collecting data from the entire airport tract. Although utilizing the cerebral capabilities of various technologies, the system is built on a singular software platform – Elektronika’s ESM. One of these technologies is a biometric access control system based on 3D facial geometry recognition.


The Artec Group’s (formerly Artec Ventures) Broadway 3D facial recognition system was selected by Elektronika, and is currently operational at Sochi International. As expounded earlier, geometry of a human face is one of the most precise biometrics that is available, it possesses a universality, distinctiveness, permanence, and is unobtrusively collectible; moreover, it is near impossible to fool or fake (Jain, Prabhakar, & Ross, 2004, Artic ID, 2013). Additionally, Broadway 3D was selected for its safety and fast performance. The system is designed to prohibit both the access of any unregistered individual or unauthorized employee. The latest version of Broadway 3-D, “is capable of identifying a person while walking, wearing hats or sunglasses and can also decipher between identical twins” (designboom, 2013). Furthermore the system’s high throughput with registration taking no longer than two seconds and with a 60 person per minute rate of scan is quite capable of handling critical rush-hour volume. Below, is a YouTube link to a short video depicting how Artec Group’s Broadway 3D facial recognition operates at Sochi:


Since August, 2011 when this paper was first penned, biometrics and specifically facial recognition database systems have moved out of the secretive world of governmental security, and into the mainstream realm of retail technology and consumer sciences. This migration includes inroads into such realms as shoplifter identification and various types of market basket analysis; the latter, a mainstream of data mining prospecting techniques (Greg, 2014). Biometric gurus FaceFirst of Camarillo, California are already providing security protection to retailers by identifying known shoplifters when they enter the facilities facial recognition surveillance security zone. This system sends text messages and emails notifications to select individuals within the organization sounding an alert that an identified shoplifter has entered the premises. This software is envisioned to recognize the “bad guy” unaided, and promptly send the appropriate message to the appropriate authorities. However, other biometric facial recognition systems are used for more lucrative purposes by lodgers and retailers.


A touted capability of the NEC Corporation is their “V.I.P. identification software”. This innovative facial recognition system is designed for hotels, casinos, and other facilities “where there is a need to identify the presence of important visitors” (Singer, 2014). FaceFirst intends on adding this capability to the existing capability of their current system. Joseph Rosenkrantz, the CEO of FaceFirst explains “Just load existing photos of your known shoplifters, members of organized retail crime syndicates, persons of interest and your best customers into FaceFirst”. Then “Instantly, when a person in your FaceFirst database steps into one of your stores, you are sent an email, text or SMS alert that includes their picture and all biographical information of the known individual so you can take immediate and appropriate action.” (FaceFirst, 2014). Additionally, for the V.I.P. shopper and high rollers, the software autonomously notifies these individual sending personalized offers to their smart phones, tablets, and other Web enabled electronic devices (Singer, 2014).


There is a caveat to employing commercial facial recognition technology, although it “has the potential to provide important benefits and to support a new wave of technological innovation, [it] also poses consumer privacy challenges” (Singer, 2014). Legislation is currently pending, having been introduced by the White House and in conjunction with the National Telecommunications and Information Administration, to draft and enact baseline federal consumer privacy legislation. This legislation would require opt-in consent for consumers.

Essentially, the same privacy concerns, concerning DNA sequencing – that of measuring biological patterns unique to individuals – are at the heart of whether a person has a right to control who has access to his or her biometric data; and how, when, and where it can be used (Singer, 2014).




Animetrice. (2008a). Biometrics and facial recognition. Retrieved from

Animetrice. (2008b). Bioinformatics. Retrieved from (2011). Geodesic distances. Retrieved from

Artec 3D. (2011). Interactive 3D. Retrieved from

Artec ID. (2011). Broadway 3D. Retrieved from

Artic ID. (2013). Broadway 3D Face Recognition System is installed at International Sochi Airport. Retrieved from (2011a). Introduction to biometrics. Retrieved from (2011b). Biometrics frequently asked questions. Retrieved from (2011c). Biometrics history. Retrieved from

Bronstein, A., Bronstein, M. & Kimmel R. (2004). Three-dimensional face recognition. Department of Computer Science, Technion – Israel Institute of Technology, Haifa 32000, Israel. Retrieved from

Challenge Liberty & Security. (2005). Ethical and social implications of biometric identification technology: Towards an international approach. Retrieved from

Clement, J. G., & Marks, M. K. (2005). Computer-graphic facial reconstruction. Burlington, MA: Elsevier Academic Press.

Deleon, V. B., Lele, S. R., & Richtsmeier, J.T. (2002). The promise of geometric morphometrics. Yearbook of Physical Anthropology, 45, 63-91. doi:

Designboom. (2013, September 7). 3D facial recognition airport security at Sochi 2014 Olympics. Retrieved from
Encyclopedia Britannica. (2011). Riemannian geometry. Retrieved from

FaceFirst. (2014). Retail. Retrieved from

Forensic Science Communications. (2008, April). The Magna database: A database of three-dimensional facial images for research in human identification and recognition. Retrieved from

Greg, A. (2014, February 2). Facial recognition software to be used to track spending habits and send offers to customers’ cell phones when they enter stores. Mail Online. Retrieved from

HSNW. (2010). Largest Moscow airport testing of facial biometric system. Retrieved from

International Institute of Hand Analysis (2005). Dermatoglyphis. Retrieved from

Jain, A. K., Prabhakar, S. & Ross, A. (2004). An introduction to biometric recognition. IEEE Transactions on Circuits and Systems for video Technology, Special Issue on Image – and Video-Based Biometrics, 14(1). doi : 10.1109/TCSVT.2003.818349

Kendall, K. E., & Kendall, J. E. (2011). Systems analysis and design (8th ed). Upper Saddle River, NJ: Prentice Hall.

Merriam-Webster. (2011). Cartesian coordinate. Retrieved from

Renaghan, J. (1997). Etched in Stone. Zoogoer 26(4). Retrieved from

Singer, N. (2014, February 1). When no one is just a face in the crowd. New York Times. Retrieved from

Stackoverflow. (2011). What does the term “canonical form” or “canonical representation” in Java mean? Retrieved from

TheFreeDictionary. (2011). Isomety. Retrieved from

Top Rhinoplasty. (2011). Frankford horizontal. Retrieved from


Appendix A


Table 2




g Glabella The most prominent midline point between the eyebrows
sl Sublabiale Determines the lower border of the lower lip and upper border of the chin
pg Pogonion The most anterior midpoint of the chin
en Endocanthion (l, r) The point at the inner commissure of the eye fissure
ex Exocanthion (l, r) The point at the outer commissure of the eye fissure
p Pupil (l, r) Determined when the head is in the rest position and the eye is looking straight forward
pi Palpebrale inferius (l, r) The lowest point in the mid-portion of the free margin of each lower eyelid
se Sellion The deepest landmark located in the bottom of the nasofrontal angle
prn Pronasale The most protruded point of the apex nasi
al Alar (l, r) The most lateral point on each alar contour
c’ Highest point of columella (l, r) The point on each columella crest, level with the tip of the corresponding nostril
ls Labiale superius The midpoint of the upper vermillion line
li Labiale inferius The midpoint of the lower vermillion line
sto Stomion The imaginary point at the crossing of the vertical facial midline and the horizontal labial fissure between gently closed lips, with the teeth shut in the natural position
ch Cheilion (l, r) The point located at each labial commissure
sa Superaurale (l, r) The highest point on the free margin of the auricle
sba Subaurale (l, r) The lowest point on the free margin of the ear lobe
pa Postaurale (l, r) The most posterior point on the free margin of the ear
obi Otobasion inferius (l, r) The point of attachment of the ear lobe to the cheek


Figure 1












Figure 2












Figure 3












Figure 4












Figure 5












1 The application of computer technology to the management of biological information. Specifically, it is the science of developing computer databases and algorithms to facilitate and expedite biological research (Animetrice, 2008b).


2 “A function between two metric spaces (such as two coordinate systems) which preserves distances. A rotation or translation in a plane is an isometry, since the distances between two points on the plane remain the same after the rotation or translation” (TheFreeDictionary, 2011).


3 “A non-Euclidean geometry in which straight lines are geodesics and in which the parallel postulate is replaced by the postulate that every pair of straight lines intersects” (Encyclopedia Britannica, 2011).


4 In mathematics, it refers to the shortest line – distance – between two points on a curved or flat surface. For our purposes, the two points rest on a Riemannian manifold and it is the geodesic that connects them (, 2011).


5 “A process for converting data that has more than one possible representation into a “standard” canonical representation. This can be done to compare different representations for equivalence” Stackoverflow. (2011).


6 “A generic name for algorithms that compute the canonical form by minimization of the stress with respect to” (Bronstein et al., 2004) examining likenesses or divergence in data.


7 “Any of three coordinates that locate a point in space and measure its distance from any of three intersecting coordinate planes” (Merriam-Webster. 2011).


8 Morphometrics “by definition, involves the quantitative study of form. However, the measures we collect to study form contain information pertaining to a combination of size and shape” (Deleon, Lele, & Richtsmeier, 2002).


9 “A plane used in craniometry that is determined by the highest point on the upper margin of the opening of the ear canal and the low point on the lower margin of the left orbit and that is used to orient a human skull or head” (Top Rhinoplasty, 2011).


10 “Artec Ventures founder Art Yukhin, was a member of the team of researchers that first invented facial recognition technology in 1999” (HSNW, 2010).



September 4, 2014
© HAKIN9 MEDIA SP. Z O.O. SP. K. 2013