VEIN PATTERN RECOGNITION BIOMETRIC SYSTEMS By Sanya-Isijola, Ademuyiwa 36641
ABSTRACT
A Wide variety of biometric system exist today, such biometric systems such as Iris, gait, odour ,fingerprint, retina, voice, face etc. are used as a means of enhancing security. Vein or vascular pattern recognition in biometric systems is an emerging technology that is been gradually adopted in some parts of Asia. This report is will discuss how vein pattern recognition works, its applications, advantages, disadvantages and mode of operations.
INTRODUCTION
For many years, Fingerprints have been used as an acceptable means of identification because of its uniqueness and individuality; these ridges found on every finger are never the same for any two individuals. For this reason, fingerprint obliteration is done by those who fear being identified, arrested, deported or convicted, with the sole aim of altering their identity completely. Offenders exploit various finger obliterating techniques (mutilation) in order to change their identity but this hardly goes undetected by law enforcement. With advances in technology, offenders are more likely to get away with fingerprint alterations. In some rare occasions, some individuals may not have any fingerprint at all as a result of a medical condition. For example, Naegeli syndrome and Dermatopathia Pigmentosa Reticularis (DPR) which Is caused by defects in the protein Keratin 14. [1] Despite the distinctiveness and preference for use in identification and authentication, fingerprints are easily left unintentionally unintentionall y everywhere. Various methods have been used to create dummy fingerprints (eg Gelatin fingertip duplication) from fingerprints that were unintentionally deposited on glass cups, window etc and have successfully fooled a biometric system. In Vein pattern recognition, these issues discussed are overcome. Vein patterns are not left around unintentionally; vein pattern cannot be obliterated or easily forged and not distorted by any disease. Reproduction of the vein line pattern is almost impossible, vein pattern of inanimate
bodily parts become useless after a few minutes due to the increasing deoxidization of the tissue. Even though the vein size may change as individuals grow, the shape of the pattern remains distinct . Like fingerprint, vein pattern is unique and stable. Vein pattern is the vast network of blood vessels underneath a person’s a person’s skin. Vein pattern can be taken from different body parts but current recognition technologies concentrates on the hand (Finger, palm, backhand and wrist). Vascular pattern differs throughout the human body, for example, the pattern of the left hand differs from the right of the same person and these patterns are not altered by aging. [2]
1.0 The Vein pattern biometric system
The recognition process involves capturing vein patterns with the use of infrared technologies (infrared light with wavelength between 700nm-1000nm). When the hand is placed on a scanner, infrared light passes through the tissue and the rays are absorbed by the red blood cells (Haemoglobin). The infrared- sensitive camera sees the shadow of the veins as black lines and the rest of the hand structure is seen as white. The extracted vein template is then compared with the previously stored patterns and a match is made. [6] The concept and components of the hand vein recognition is the same with a typical biometric system, the vein pattern biometric system has the following components:
Sensor module: used for data acquisition
Feature extractor module: used to extract features from the acquired data
Matching module: used to compare the extracted feature vectors against that stored in the database. Decision making module.
1.1 Types of Palm vein scanning technology
There are two main types of palm vein scanning technology; they are Far-infrared and Nearinfrared.
1.1.0 The Far-infrared scanning technology:
It has two main short comings namely Image contrast and limited details. Image contrasts: It has been proved from experiments that vein extraction using Far-infrared technology can be very difficult because the temperature of vein and its surrounding are similar, thus making image contrast similar. Limited details: This is another problem associated with the Far-infrared technology because only major veins in the palm are displayed making it inappropriate for high security applications. 1.1.1 The Near-infrared scanning technology:
The Near-infrared technology overcomes the issues associated with Far-infrared scanning technology. Small vein images can be captured and are visible because the infrared rays have been absorbed by deoxidized haemoglobin in the vein vessels, thus making the veins to appear in black pattern by reducing reflection. Liveliness of user is easily determined because the moment the deoxidized haemoglobin stops flowing through the vein; the pattern seizes to be visible. [3]
2.0 Thermal hand vein pattern verification system
The thermal vein pattern verification system involves the following stages:
Image Acquisition
Image enhancement
Vein pattern segmentation
Skeletonization
Matching
2.0.1 Image Acquisition:
When a finger is placed into the device, infra red light is shown on the finger, this makes the veins visible while the IR camera captures the image, this process is called Image Acquisition and it is done by the Sensor module. Certain factors affect the vein structure visibility, they are:
Thickness of the skin
Degree of venous engorgement
Levels of subcutaneous fat
Surface features such as moles, scars, pigmentation and hair
These factors can be overcome by the sensitivity and quality of the capturing device.
2.0.2 Feature extraction:
Features of the data acquired by the sensor module are extracted. The image acquired is in gray scale and the image is enhanced by noise removal, illumination dispelling and normalization.
2.0.3 Vein pattern segmentation
This process separates the vein pattern from the image background by using locally adaptive thresholding, this is done by calculating threshold at each pixel. Dilation and morphological erosion is done to clean up images after thresholding.
2.0.4 Skeletonization:
Due to change in vein size as human grows, the shape of the vein pattern is the main feature used for later recognition. The pattern’s shape is derived b y extracting it’s it’s skeleton and pixels isolated during this stage are cleaned by a process known as Pruning. [6]
2.0.5 Matching:
The pattern generated is converted into a matching data (binary images) and this binary image is compared with the data or template stored in the database. The mismatch ratio, Rm is calculated to check correlation between the stored template and the data acquired. [4]
Figure 1: Hand vein pattern verification system model
3.0 Security Evaluation of biometric systems systems
The figure below (Figure 2) shows 8 different points of a possible attack on a biometric system. Attacks such as Spoofing attacks, Replay and transmission attacks and code modification can be used at different stages of biometric systems.
3.0.1 Spoofing attacks
This is done at point 1, attacker presents sensor with fake biometric with the intent to fool the system. Based on the fact that veins reside internally, it is more difficult for attacks because enrolled users do not leave biometric data unknowingly in a place like it is accidentally done with fingerprints. Dummies or imitation of body part cannot be used to fool the palm vein recognition system because the palm vein recognition method detects liveliness in the subject (flowing blood).
3.0.2 Replay and transmission attacks
Points 2, 4, 7 and 8 are susceptible to different kinds of replay and transmission attacks. Through testing, it was observed that the signal between sensors and computer can be intercepted, hence compromising the raw data and giving the attacker an authentic user’s biometric image. image. Attackers can reuse the authentic user’s biometric image at a later time. This attack is feasible with different biometric system types but further testing is yet to be conducted on the possibility of this attack been feasible on the palm vein system.
3.0.3 Code Modification
Security mechanism of the system may be bypassed by attacker modifying code at point 3 and 5. Template modification/introduction Attacker can add his/her own template or modify an existing user’s template in database.
3.1 Security improvement of biometric systems
3.1.1 Multimodal biometric systems
Multimodal biometric systems are biometric systems that identify or verify users by using multiple/several independent biometrics. This makes the possibility of a successful attack more complicated, thus enhancing security.
3.1.2 Combination of password and biometrics
A combination of the biometrics with the use of a password is used together. Biometric data provides the login information and an additional password is needed to access the system. [3]
Figure 2: Biometric systems point of attacks
4.0 Application of Palm vein recognition systems
The vascular pattern recognition system design is been developed by Fujitsu, Hitachi, Bionics and several other organizations still have their own in progress. [7] The palm vein biometrics can be used for login control, security system, banking services etc. It has been adopted in various countries in Asia e.g. Japan. People often object to any contact oriented form of biometrics due to sanitary reasons or association of the biometric system with criminal activities such as Fingerprints, this factor helped foster the adoption of this technology in a country like Japan. [8]
4.1 Case Studies 4.1.1 Bank of Tokyo-Mitsubishi, Tok yo-Mitsubishi, UFJ th
The bank of Tokyo-Mitsubishi is the 10 biggest bank in the world and the largest in Japan. The bank has been using palm vein authentication biometrics for over 3 years and has currently installed in 5,000 branches of the UFJ bank across Japan. The technology was adopted for ATMs which is the most demanding customer-facing solutions. Account holders enroll for the system with their palm and the data acquired (vascular information) is stored in a smart card given to them. To access account via ATM, card is
inserted, PIN is typed and palm is held over sensor on ATM. It is well accepted by the customers and has been used by over a million people without a single incident reported.
4.1.2 Carolinas Health Care System, Charlotte N.C
Due to the occurrence of medical identity theft where an imposter uses an authentic patient’s identity to incur bills, Patients identity remains a high priority issue for most hospital. A palm vein solution was designed to enhance the retrieval of patient’s of patient’s records and identification of patients of Carolinas health care system. Hence eliminating human error and enhancing identity theft protection. [5]
5.1 Advantages of palm vein recognition biometrics
Hidden characteristics are used as biometric features
It is difficult to forge for intruders because blood vessels are hidden within the body.
Palm vein recognition is not affected by dryness or roughness of skin or by physical injury on surface of the hand. Palm vein recognition has an FRR of 0.01% and FAR of less than 0.00008%. This is based on experiments conducted on 70,000 individuals in Japan. This technology is non-intrusive; it does not involve physical contact between the user and system.
5.2 Disadvantages of palm vein recognition biometrics
There numerous factors that can affect the quality of the captured image. They are body temperature, ambient temperature, humidity, unevenly distribution of heat, heat radiation, nearness of vein to surface, camera calibration and focus. Palm vein recognition is Invasive because it creates apprehension amongst users that it can be a painful process. It is still relatively expensive and not available for mass production yet. [3]
CONCLUSION
Palm vein pattern recognition is a convenient and easy to use biometric technology with high security and accuracy level. The technology is gaining momentum but it is still expensive and relatively untested because it has not yet been marketed globally.
References
http://news.nationalgeographic.com/news/2006/09/060922-fingerprints.html [1]
http://www.findbiometrics.com/article/320 [2]
Kenneth Wong, Thomas Lai, Bosco Lee, Analysis of palm vein biometric system [3]
L.Wong, G.Leedham and S.Y. Cho, Infrared imaging of hand vein patterns for biometric purposes [4] http://www.handresearch.com/news/lending-a-hand.htm [5] Lingyu Wang and Graham Leedham, A thermal hand vein pattern verification system.[6] Europe Biometric portal, Trend report 2007 [7] Biometrics in public by Fujitsu [8]