Posts Tagged ‘Facial scan’

The World of Biometrics



For more than 4000 years, man has devised the means for securing important objects. Way back in the 1400 BC, the Egyptians devised a simple yet effective lock and key mechanism to protect the Pharaohs from grave robbers. Not only were the tombs secure, but traps were devised which would fill a room with sand if incorrectly unlocked. Now in the 21 century we still search for the strongest protection means possible – with biometrics being the key (pardon the pun) for the 21st century.

So what is biometrics? Well it is the automated use of physiological or behavioral characteristics to determine or verify identities (from Biometrics, 1999). It can be used for many applications from simple door locks to access control to a computer.
The traditional authentication methods were the use of passwords and lock-key mechanics. Yet these methods without luck can be easily cracked using simple brute force. However biometric data cannot be guessed or stolen in the same fashion that passwords and locks can. They can be broken under certain conditions though BUT today’s technology is highly unlikely to be fooled by a picture of a face, an impression of a fingerprint or a recording of a voice. Though biometrics seems like the solution to all our problems, it may not be for all applications based on factors such as cost, risk and privacy. What biometrics does is bring the greatly simplified authentication process  and in most cases, increased accountability, and the fact that they are next to impossible to be forgotten or fooled.

So let us know move on to describing the main types of system that have been developed. These systems are based on behavior and physiological characteristics and are almost always done automatically by computers due to the increased processing speed. In this article I shall separated biometrics into 7 types those being:

finger-scan                retina-scan
facial-scan                voice-scan
hand-scan                 signature scan

There are 2 main goals in biometrics, that being to identify (The computer finds the identity of a person without that person first claiming an identity) and to verify (The computer verifies a given identity that the person claims to be) and are used to restrict access to either physical or logical access.

So how does biometrics work? Well to put it simply, the user first creates a template which is stored in a database. The next time the user tries to access the system, the new template is compared to the stored one and accessed granted if the score is within a confidence method as the data will never completely match again. When the user submits some biometric information such as a finger print, it is not usually stored as an image due to the size, think about it – over 10 million finger prints would require a tremendous amount of space, but as a template with features such as curves extracted out. This storage of data has some people worried however what they have to understand is that it would cost enormously to save all the raw data in its current form. Templates are small files that have distinct characteristics for each user.

Finger Scan
Finger scans are probably the most common form of biometric authentication. Police stations all over the word use it to identify criminals, a reason why some people refuse to use it under normal circumstances – they see themselves like criminals. The basic process is as follows:
> Image Acquisition – The finger print is acquired by some sort of device such as chip based cameras or ultra sonic imaging. Hight resolutions are needed in order to see the detail clearly.
> Image Processing – The captured finger print is then converted to black and white to simplify the process and are thinned as much as possible to bring out the features clearly. Image processing such as erosion can be used. After this a template is extracted with swirls, loops and ridge endings all being stored.
> Template Processing – The template(s) are processed to store the correct information.
Many finger print technology now checks for blood flow so that an intruder cannot just cut off someones finger. The popular discovery channel show, Mythbusters, have shown that these systems can be fooled with simple cut outs of photocopied fingers and molds by simply licking the mold which inhibits an electric charge.

The advantages of finger scans are the facts that it can be used in a wide range of environments with already existing proven technology that is easy to use and cost effective. However there are a small percentage of users that would have trouble using the system (EG: The Polynesians are known to have feint finger prints). Accuracy may also decrease over time.

Facial Scan
Facial scanning technology extracts features from a users face such as distance between eye sockets and the placement of these features. Nose shape and cheekbone structure can also be used. But using the mouth as a feature is not usually done due to the fact that growing a beard can affect the accuracy. There are several different approaches to face recognition which I will not all go into. One example is the Eigen Faces in which information is broken into principal components that are used to derive the template used for matching. The general algorithm used is to first train the system using a variety of poses of a person. Eigen vectors are then determined. The recognition stage involves calculating the distance between features for example and storing them in a template.

The advantages of facial recognition is that existing image capture technology can be used and no physical contact is needed. However there are numerous problems with this system in the fact that the environment can deeply affect the accuracy due to changes in lighting among others.  Also as one ages these features can become less distinguished.

Iris Scan
The iris scan is one of the most accurate and secure systems around. The processing done involves finding the iris’ outer edge and then finding the inner edge of the iris. It performs by dividing the iris into a number segments and a      count is kept of the frequency of the features as well as their position. The process though can be difficult if one has very dark eyes which makes feature extraction difficult. With the outer and inner parts determined, the system them extracts features such as color tones.

The advantages of this system is that it provides a high level of accuracy and is generally stable over long periods. However due to hardware costs this option may not be viable to everyone. This is due to the complex image capturing devices and algorithms.

Retina Scan
Retina scans are even more secure than iris scans as it involves features within the user’s eye such as blood vessels. But due to the complex task of mapping the blood vessels, specialized equipment is needed making this option again complicated.

The advantages are the high accuracy and stable physiological trait. However as the equipment is specialized it is expensive and the user may be discomforted due to the eye technology.

So what method is right for you? well that all depends on what you are protecting. If you a protecting a computer lab of US$200,000 worth of equipment then a US$1 million dollar retina scan system may not be what you are looking for. Selecting the right system depends on your needs. There are many examples of these system in operation today from banks in Puerto Rico using finger prints for authorization, to cameras in casinos protecting their interests by recognizing frequent gamblers. One thing is for sure, biometrics is here now and is here to stay.