In the security world, Positive Identification is one of the most important keys. common identification tools such as Cards or Codes are easily transferred or duplicated with or without the knowledge of the owner. using Biometric Identification ensure that the person identified by their specific unique Biological Attributes that separate them from others.
Buyer-Beware: When using Biometric technology, not all products are the same. In analyzing the Biometric information there are many components and proprietary algorithms used to improve performance. There are many technologies used to overcome attempts to “defeat” technologies and use tools to cheat the readers, these technologies are affecting the price of the product. In Biometrics, the difference between Cheep product and Good product can mean a lot.
The following is a review of current technologies available in the Biometric Identification world as well as their advantages and disadvantages:
Hand-Geometry: also called Hand-Punch. this is the oldest biometric technology in existence. It’s principle based on identifying person by the shape of their hand. The underlying assumption is that no two hands are similar. Hand Geometry readers require accurate positioning of the palm and fingers, this achieved by aligning the hand to posts on the reader.
Hand-Geometry is one of the least accurate methods hence need another measurement of identification such as card or PIN. It also need long time for verification making it a very bad candidate for Access Control.
Hand-Geometry technology fail when used with persons under 25 where their hands grow as well as with major changes in hand shape due to weight gain/loss or illnesses such as arthritis.
Fingerprint: The most common used biometric technology. It’s principle based on identifying person based on Fingerprint. Person fingerprint is a unique identification used even in courts to positively identify a person. It’s accuracy considered as up to 99.97%, but, it rely on the Algorithm of the device itself. Fingerprints are analyzed, in most cases, the same as forensic scientist are showing on TV series. The algorithm is looking for “minutiae” or ridges in the fingerprint pattern. This model converted to “templates” that represent the fingerprint. Depending on the vendor, many fingerprint readers do not store the fingerprint as picture, just a representation of the minutiae relative positioning saving space and providing better privacy and security.
Fingerprint readers are split in to reader use and reader technology. Reader use is either by Swiping the finger across a narrow strip (like hand held scanner) or Full Finger where the finger being positioned on the reader surface. The “swipe” readers usually cost less and take less space, but, they are more difficult to use. in the reader technology there are multiple technologies from the most common Optical where the scanner is taking a picture of the finger, Capacitance and Thermal with much thinner sensors (iPhone), RF and Ultrasound and even combination of technologies to ensure accuracy and prevent false identifications of fake fingers.
Fingerprint considered as one of the most accurate and most reliable since fingerprints never change, from birth to death, they stay the same. They change size as we grow, gain/lose weight, but, the relationship between the minutiae locations stay constant. With the right algorithm, the enrollment of a two years old will work even after retirement.
Fingerprint will not always is the right one. Special attention is required to the technology provider. Many products in the market use very low quality components with a very basic algorithm. Cheap product usually indicate quality. Low quality will not provide the expected reliability, accuracy and security. The fingerprint technology also have it’s natural limitations. People who work in an environment where their fingers are being injured frequently (mechanic shop, metal fabrication) will not be able to use this technology. It is recommended to enroll people with at least 2 fingers to ensure use in case one of the fingers has a paper-cut or injury.
Iris / Retina – The human eye is another organ that never change from birth to death and is unique to each person. The accuracy can get to the same level as fingerprint, 99.97%, yet the identification is more complicated. It is based on taking a very high resolution picture of the person eye and comparing it to enrolled picture in the database. Most readers require the subject to stand still during the identification process, what make it difficult for Access Control. In last few years, with the advancement of high resolution cameras, some products appeared, claiming to read iris from distance allowing identification before getting to the door. Iris readers are the most expensive readers in the biometric access control realm!
Iris technology biggest advantage is the very high accuracy without any contact with the sensor.
Iris technology may fail in cases where there are eye infections and diseases.
Face Recognition – This non-contact technology is becoming more and more common. It is based on identifying a person based on facial structure. There are two leading methods, one is based on analyzing the whole face and the other is analyzing the center from the nose-tip to the eyebrows. Where the full face is considered more accurate, the half-face is faster. Face Recognition is not as accurate as Fingerprint or Iris but use the same concept used in fingerprint analysis. By taking multiple pictures, the reader create a template containing the relationships between specific facial characteristics such as the eyes, nose, ears, eyebrows, jaw bones etc. In order to ensure usability and fast identification, many manufacturers need to overcome facial expressions and light conditions. Many face recognition products require stable environment (light and colors) to be able to differentiate the face from the background and may require standing still for few seconds. Depending on the product, it may or may no fit Access Control systems.
Face Recognition systems are very dependent on Algorithm and technology quality. Low cost units may be fooled by a picture or fail to recognize in light changes. Better units will work in any light condition (using IR illumination) and utilize 3D or other technology to identify real-face.
Due to lower accuracy it is recommended to use Dual-Factor ID by adding PIN or Card.
Vascular – Based on analyzing the vain structure in the palm of the hand, assuming it is unique to each person. This scan is being done by emitting Near-IR light through the palm of the hand. This non-contact method will work only on live hand, hence, difficult to defeat. It is used in many cases in conjunction to Fingerprint by illuminating the finger blood vessels to verify the finger is “live”.
The technology is still in early stages and not commonly used due to high price and slow identification process. It is problematic where people have blood pressure problems, arthritis, diabetes and other illnesses that affect blood vessels and bones structure.
Voice – Voice recognition is used in many applications not related to Access Control. Many call centers use the technology to identify customers calling, pulling their file by the time they finish their greetings. Some others are using the technology to identify stress, mood, lying and more. This technology has not been proven as reliable for security purposes yet.
Behavioral – A person walk pattern is in many ways unique. There are systems used to identify persons using multiple indicators including their measurement, walking pattern, posture. These systems are usually very expensive and require long line of unobstructed path (30ft.). This technology is also used to identify certain patterns observed as “threat” such as hesitating or over determine or angry behavior.
DNA – Probably the most accurate identification method. Currently, it has many obstacles, the need of a “sample” require removal of body cells or fluids, and technology today require several hours if not days for an analysis.