WO2009028926A2

WO2009028926A2 - Apparatus and method for volumetric multi-modal hand biometrlc identification

Info

Publication number: WO2009028926A2
Application number: PCT/MY2008/000092
Authority: WO
Inventors: Weng Kin Lai; Chen Change Loy; Sock Yin Tai; Hock Woon Hon
Original assignee: Mimos Berhad
Priority date: 2007-08-29
Filing date: 2008-08-29
Publication date: 2009-03-05
Also published as: WO2009028926A3

Abstract

The present invention relates to an apparatus and method for volumetric multi- modal hand biometric identification characterized in that wherein the present invention comprises of an illumination source for emitting near infra-red (0.75-1.4 μm in wavelength) and structured light, a camera system with a recorder adapted to capture an image of the object in a digital form; and data fusion that processes the data representation to produce useful information to the system managers and wherein palm prints are extracted from scanners which scanned the palm in which vein pattern biometrics records subcutaneous infra red absorption patterns to produce unique and private identification templates of a user.

Description

APPARATUS AND METHOD FOR VOLUMETRIC MULTI-MODAL HAND

BIOMETRIC IDENTIFICATION

FIELD OF THE INVENTION

The present invention relates to an apparatus and a method for volumetric multi modal biometric identification.

BACKGROUND OF THE INVENTION

There are many biometric authentication and identification systems that individually scans either the fingerprint, palm-print or vein patterns on the palm. These are basically based on two-dimension mappings or images of that particular feature such as fingerprint, palm-print or vein patterns. With the decreasing cost of digital cameras as well as improving resolution of such cameras, digital scanner systems will become cost effective and competitive. Coupled with good pattern recognition software, volumetric imaging biometric systems can give better true positive recognition rates while minimizing the false positives readings than their 2D counterparts. Moreover, volumetric imaging techniques can give multi-modal biometric readings, which understandably can provide stronger and more robust security.

SUMMARY OF THE INVENTION

The present invention relates to an apparatus and method for volumetric multi- modal hand biometric identification characterized in that wherein the present invention comprises of an illumination source for emitting near infra-red (0.75-1.4 μm in wavelength) and structured light, a camera system with a recorder adapted to capture an image of the object in a digital form; and data fusion that processes the data representation to produce useful information to the system managers. The camera system scans the palm that is bathed in the light from the illumination source. In addition to the distinctive palm print-print of the user, the scanner will also capture and record the subcutaneous infra red absorption patterns of the veins in the palm of the hand to produce unique and private identification templates of the user. Furthermore, the distinctive vein patterns on the back of the hand are also be used to identify or authenticate humans as a biometric feature.

The present invention further includes an intelligent multi-modal biometrics unit to identify and authenticate users which relies on the stable features of a clear palm print, a clear set of vein patterns on the back of the hand and a unique volumetric image of the hand. The new contactless technology alleviates concerns about hygiene associated with having many people touching the same sensor device, making it ideal for widespread use and wherein the rotating energy source captures more information of the hand for identification purposes, resulting in significantly better levels of authentication accuracy as well as lower levels of misclassification.

The feature fusion method, which combines vein pattern features and palm print features will result in a smaller template matching while indirectly contributing to a smaller database. Furthermore, this will promote faster matching as well as efficient transmission of biometric template through communication networks.

The present invention provides a new approach to capture hand biometric information through a single rotating sensor in an adequately and appropriately illuminated environment in which an individual's three-dimensional vein pattern image is captured by radiating his/her hand with near-infrared rays and structured light. The reflection method illuminates the palm using the appropriate energy source and captures the light given off by the region after diffusion through the palm. The deoxidized hemoglobin in the vein vessels will absorb the infrared ray, thereby reducing the reflection rate and causing the veins to appear as a black pattern. This vein pattern is then combined with features extracted from palm itself in the form of a volumetric palm print. A feature-based fusion method combines features from different sensors with dimensionality reduction to reduce the feature dimension in which such combination would eliminate redundant features and spurious signals after performing dimensionality reduction, produce smaller feature vectors and produce a cancelable biometric which will secure an enhanced privacy for the user since the actual vein pattern and palm print can never used anywhere else as the features are transformed using a unique dimensionality reduction method.

Before the apparatus is able to process the identification request, each user is required to enroll with the system. The user's identity and his/her volumetric multimodal biometric templates are entered to be stored in the storage means, via the input device and wherein from the key input device, the appropriate volumetric key features such as the hand geometry length, width, number of fingers, length of each finger, major unique marks on the hand, etc., the palm print, vein patterns on the back of the hand, general physical characteristics of the hand, etc. are captured. These features are extracted by the extraction means and wherein a fusion of the plurality of live volumetric features of the user is obtained. A template information is then stored away in the database for future reference, to be used during the authentication/identification/verification process.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1 shows a human palm print with line patterns.

Figure 2 shows some conventional palm print scanners.

Figure 3 shows an example of extracted vein patterns from a human palm that has been illuminated under an appropriate infra-red light source.

Figure 4 shows some current commercially available vein pattern biometric systems.

Figure 5 shows a plurality of vein patterns on the back of a hand. Figure 6 shows a volumetric multi-modal hand biometrics system.

Figure 7 shows a flow-chart of the proposed volumetric multi-modal hand biometrics for enrollment.

Figure 8 shows a flow-chart of the proposed volumetric multi-modal hand biometrics during identification.

DETAILED DSCRIPTION OF THE PRESENT INVENTION

The present invention would now be described in detail with reference made to the accompanied drawings but not limited thereto.

The present invention comprises of an illumination source which is an appropriate energy source that is capable of emitting near infra-red (0.75-1.4 μm in wavelength) and structured light, a camera system with a recorder adapted to capture an image of the object in a digital form; and data fusion that processes the data representation to produce useful information to the system managers.

Palmprint biometrics is an emerging technology which recognizes a person based on the principle lines, wrinkles and ridges on the surface of the palm. These line structures are stable and remain unchanged throughout the life of an individual. More importantly, no two palmprints from different individuals are the same, and normally people do not feel uneasy to have their palmprint images taken for testing. Therefore palmprint recognition offers a promising future for medium-security access control systems. Figure I illustrates a typical image of a palmprint taken from a hand. Some of the unique line patterns on the hand can be clearly seen.

It is common that the palm prints are extracted from scanners which scanned the palm, as illustrated in Figure 2. Vein pattern biometrics basically records subcutaneous infra red absorption patterns to produce unique and private identification templates for a user. The technology is a vascular "bar" code reader for humans. Veins and other subcutaneous features present large, robust, stable and largely hidden patterns. Subcutaneous features can be conveniently imaged within the wrist, palm, and dorsal surfaces of the hand. Figure 3 shows an example of extracted vein patterns from the hand that has been illuminated under infra-red light source.

Japan's banking industry is already placing its bets on biometrics as a solution to the growing problem of counterfeit ATM cards and ID theft. First introduced in October 2004 through the Bank of Tokyo-Mitsubishi UFJ biometric bank cards based on the vein patterns in the palm of the hand, 15 other financial institutions (two mega-banks, five regional banks, and eight credit unions) have announced in December 2005 that they will be adopting a similar technology. In addition, more than 50 other institutions have announced plans to introduce the cards over the next several years. These biometric bank cards may be based on either palm vein authentication, Figure 4a or a rival technology based on vein patterns in the finger, Figure 4b.

Nevertheless, in addition to the vein patterns on the palm, distinctive vein patterns also exist on the back of the hand, as illustrated in Figure 5. These may also be used to identify or authenticate humans as a biometric feature.

The present invention is a volumetric imaging biometric system for hand biometric identification. Figure 6 illustrates one possible arrangement for the proposed volumetric multi-modal hand biometrics scanner. The invention which will comprise of such an apparatus together with the intelligent multi-modal biometrics unit to identify and authenticate users relies on the stable features of,

i) a clear palm print, ii) a clear set of vein patterns on the back of the hand. iii) a unique volumetric image of the hand Henceforth, the system comprises of (a) a rotating energy source, with the capability of providing a near-infrared emitting source (0.75-1.4 μm in wavelength) and structured light, (b) a detection system to acquire received energy and (c) pattern recognition software to recognize three-dimensional (3D) vein pattern and palm print of individual's hand.

In contrast to the contact-based biometric technology, this new contactless technology alleviates concerns about hygiene associated with having many people touching the same sensor device, making it ideal for widespread use. The rotating energy source captures more information of the hand for identification purposes, resulting in a significantly better levels of authentication accuracy as well as lower levels of misclassification (lower false-acceptance-rate - FAR and false-rejection-rate- FRR). Feature fusion which combines vein pattern features and palm print features will result in a smaller template matching while indirectly contributing to smaller database, promoting faster matching as well as efficient transmission of biometric template through communication networks.

Basically an ideal biometric system is one with zero FAR (false acceptance rate) and FRR {false rejection rate). Existing products for palm vein pattern recognition is accurate but may be improved further by incorporating more features during recognition. Nevertheless, current systems can provide a false rejection rate of 1% and a false acceptance rate of 0.5%, with an Equal Error Rate of 0.8%. Furthermore, most multimodal biometric system has large databases to handle a large number of "signatures" for authentication. Hence, a unimodal or biometric system based on a singular feature may not be able to correctly identify or authenticate the person in the database.

The invention provides a new approach to capture hand biometric information through a single rotating sensor in an adequately and appropriately illuminated environment (chamber). An individual's three-dimensional vein pattern image is captured by radiating his/her hand with near-infirared rays. The reflection method illuminates the palm using an appropriate energy source (which provides near infrared ray as well as structured light) and captures the light given off by the region after diffusion through the palm. The deoxidized hemoglobin in the vein vessels will absorb the infrared ray, thereby reducing the reflection rate and causing the veins to appear as a black pattern. This vein pattern is then combined (fused) with features extracted from the palm itself in the form of a volumetric palm print.

The feature-based fusion method combines features from different sensors with dimensionality reduction to reduce the feature dimension. The advantages of such an approach is expected to:

- eliminate redundant features and spurious signals (noise) after performing dimensionality reduction

- produce smaller feature vectors,

- produce a cancelable biometric which will secure an enhanced privacy for the user since the actual vein pattern and palm print can never used anywhere else as the features are transformed using dimensionality reduction method.

Before the apparatus is able to process the identification request, each user is required to enroll with the system. The user identity and his/her volumetric multi-modal biometric templates are entered to be stored in the storage means via the input device as illustrated earlier. The flow charts during enrollment and identification are shown in

Figures 7 and 8 respectively.

Figure 7 is the block diagram representation of the enrollment process for the present invention. It generally combines, the multi-modal biometric features related to the human hand that are captured from an appropriate input device, i.e. the volumetric hand input scanner.

From the key input device, the appropriate volumetric key features are captured. This may involve two or more of any of the following measurements, i.e. the hand geometry (1) (length, width, number of fingers, length of each finger, major unique marks on the hand, etc.), the palm print (2), vein patterns (3) (on the back of the hand), general physical characteristics of the hand, etc. These features are to be extracted by the extraction means (4) through any of the effective methods known in the art. From the extraction means, a fusion of the plurality of live volumetric features (4) of the user is obtained. This may involve any of the aforementioned measurements as described earlier. The template information is then stored away in the database (6) for future reference, to be used during the authentication/identification/verification process (5).

Figure 8 is the block diagram representation of the identification/verification process for the present invention. The verification/identification process is to be individually performed on the plurality of biometric data captured from the input device (1 - 4) and compared (5) against that of the stored biometric templates (6) using any of the known effective method in the art to produce the desired identification result - No Match (7) or a Match (8).

Claims

1. Apparatus and method for volumetric multi-modal hand biometric identification characterized in that wherein the present invention comprises of an illumination source for emitting near infra-red (0.75-1.4 μm in wavelength) and structured light, a camera system with a recorder adapted to capture an image of the object in a digital form; and data fusion that processes the data representation to produce useful information to the system managers and wherein palm prints are extracted from scanners which scanned the palm in which vein pattern biometrics records subcutaneous infra red absorption patterns to produce unique and private identification templates for a user.

2. Apparatus and method for volumetric multi-modal hand biometric identification as claimed in Claim 1 wherein in addition to the vein patterns on the palm, distinctive vein patterns on the back of the hand are also be used to identify or authenticate humans as a biometric feature.

3. Apparatus and method for volumetric multi -modal hand biometric identification as claimed in Claim 1 wherein the present invention further includes an intelligent multi-modal biometrics unit to identify and authenticate users relies on the stable features of a clear palm print, a clear set of vein patterns on the back of the hand and a unique volumetric image of the hand.

4. Apparatus and method for volumetric multi-modal hand biometric identification as claimed in Claim 1 wherein the new contactless technology alleviates concerns about hygiene associated with having many people touching the same sensor device, making it ideal for widespread use and wherein the rotating energy source captures more information of the hand for identification purposes, resulting in a significantly better levels of authentication accuracy as well as lower levels of misclassification.

5. Apparatus and method for volumetric multi-modal hand biometric identification as claimed in Claim 1 wherein the feature fusion combines vein pattern features and palm print features results in a smaller template matching indirectly contributing to smaller database while promoting faster matching as well as efficient transmission of biometric template through communication networks.

6. Apparatus and method for volumetric multi-modal hand biometric identification as claimed in Claim 1 wherein the present invention provides a new approach to capture hand biometric information through a single rotating sensor in an adequately and appropriately illuminated environment in which an individual's three-dimensional vein pattern image is captured by radiating his/her hand with near-infrared rays and wherein the reflection method illuminates the palm using an appropriate energy source and captures the light given off by the region after diffusion through the palm and wherein the deoxidized hemoglobin in the vein vessels will absorb the infrared ray, thereby reducing the reflection rate and causing the veins to appear as a black pattern and wherein this vein pattern is then combined with features extracted from palm itself in the form of a volumetric palm print.

7. Apparatus and method for volumetric multi-modal hand biometric identification as claimed in Claim 1 wherein a feature-based fusion method combines features from different sensors with dimensionality reduction to reduce the feature dimension in which such a combination would eliminate redundant features and spurious signals after performing dimensionality reduction, produce smaller feature vectors and produce a cancelable biometric which will secure an enhanced privacy for the user since the actual vein pattern and palm print can never used anywhere else as the features are transformed using dimensionality reduction method.

8. Apparatus and method for volumetric multi-modal hand biometric identification as claimed in Claim 1 wherein before the apparatus is able to process the identification request, each user is required to enroll with the system and wherein the user identity and his/her volumetric multi-modal biometric templates are entered to be stored in the storage means via the input device and wherein from the key input device, the appropriate volumetric key features are captured such as the hand geometry length, width, number of fingers, length of each finger, major unique marks on the hand, etc., the palm print, vein patterns on the back of the hand, general physical characteristics of the hand, etc and wherein these features are extracted by the extraction means and wherein a fusion of the plurality of live volumetric features of the user is obtained and wherein a template information is then stored away in the database for future reference, to be used during the authentication/identification/verificatioπ process.