Certain Investigations on Transform Based Techniques in Palmprint and Finger Knuckle-Print Biometrics for Personal Authentification

TABLE OF CONTENTS

1 INTRODUCTION TO BIOMETRICS
1.1 INTRODUCTION
1.1.1 Biometric Systems
1.2 PALMPRINT BIOMETRICS
1.2.1 Preprocessing and ROI Extraction for Palmprint Biometrics
1.3 FINGER KNUCKLE- PRINT BIOMETRICS
1.3.1 Finger Knuckle-Print Anatomy
1.3.2 Preprocessing and ROI Extraction for Finger Knuckle-Print Biometrics
1.4 PROS OF FINGER KNUCKLE-PRINT AND PALMPRINT
1.5 LOCAL AND GLOBAL FEATURES
1.6 PROBLEM STATEMENT
1.7 MOTIVATION
1.8 OBJECTIVES
1.9 BIOMETRIC DATASETS
1.9.1 College of Engineering – Pune (COEP) Palmprint Datasets
1.9.2 The PolyU Palmprint Datasets
1.9.3 Indian Institute of Technology (IIT Delhi) Touchless Palmprint Datasets
1.9.4 The PolyU Finger Knuckle-print Datasets
1.10 PERFORMANCE METRICS
1.10.1 False Acceptance Rate and False Rejection Rate
1.10.2 Speed
1.10.3 Equal Error Rate (EER)
1.10.4 Correct Classification Rate (CCR)
1.10.5 Data Presentation Curves
1.10.5.1 Receiver Operating Characteristic (ROC) Curve

2 LOCAL AND GLOBAL FEATURE EXTRACTION USING WINDOW WIDTH OPTIMIZED STOCKWELL TRANSFORM IN PALMPRINT BIOMETRIC SYSTEM
2.1 OVERVIEW OF WINDOW WIDTH OPTIMIZED S-TRANSFORM
2.1.1 Algorithm for Determining the Time Invariant p
2.1.2 Algorithm for Determining p(t)
2.1.3 Inverse of the WWOST
2.2 LOCAL - GLOBAL FEATURE EXTRACTION AND MATCHING CHAPTER NO. TITLE PAGE NO
2.2.1 Local Feature
2.2.2 Global Feature
2.2.2.1 Phase-only correlation
2.2.2.2 Band-limited phase-only correlation
2.3 LOCAL GLOBAL FEATURE FUSION FOR PALMPRINT RECOGNITION
2.4 EXPERIMENTAL RESULTS AND DISCUSSION
2.5 SUMMARY

3 CONCLUSIONS
3.1 SUMMARY AND CONCLUSIONS

REFERENCES

LIST OF FIGURES

Chapter 1 Introduction to Biometrics

This chapter emphasize the significance of palmprint biometrics, Finger knuckle-print biometrics and their performance measures. Also the characteristics of local and global features are presented in this chapter.

1.1 Introduction

Biometrics refers to technologies for measuring and analyzing a person’s physiological or behavioural characteristics (Wayman 2001). These characteristics are unique to individuals and it can be used to verify or identify a person. The applications of biometrics are considerably increased in the last years and it is expected in the near future. Depending on the deployment of biometrics, the applications are categorized in the following five main groups: forensic, government, commercial, health-care and travelling-immigration. However, some applications are common to these groups such as physical access, personal computer/network access, time and attendance, etc.

Biometrics has increasing attention in the e-world. Different types of biometrics were used in different applications. There are very few best biometric systems available in the market. The three different types of authentication are used in security system. The first type of authentication is password system and Postal Index Number (PIN) system. The second type of authentication is a card key, smart card or token system. The third type of authentication is a biometric technology. Out of these types of authentications in security system, biometric is the best secure and expedient authentication tool.

Biometrics cannot be easily borrowed, stolen or forgotten compared to the traditional security systems. The forgery of the biometric system is practically impossible. It refers to the person’s unique physical or behavioural characteristics to distinguish or authenticate their own identity. The various physical biometrics are fingerprints (Belguechi et al 2013) hand or palm geometry (Matos et al 2012), retina (Hussain et al 2013), iris technique considers as a resemblance measure in certain biometrics systems (Miyazawa et al 2008), face (Yuchun et al 2002), palmprint (Sun et al 2005) hand vein (Huang et al 2013), palm vein (Venkat Narayana & Preethi 2010), finger knuckle-print (Nanni & Lumini 2009) or ear (Middendorff 2011). The behavioural biometrics is signature (Bertolini et al 2010), voice (Hollein 2002), keystroke pattern (Pin et al 2013) and gait (Hoang et al 2013).

1.1.1 Biometric Systems

The biometric trait can be acquired from an individual and then the feature set is extracted from the acquired data. Finally, this feature set is compared with the template set in the database. Therefore biometric system is also referred as a pattern recognition system. Biometric system may operate either in verification mode or identification mode based on the application it is used in the security system. In the verification mode, an individual’s identity is authenticated in the security system by comparing the captured biometric trait with the own biometric template(s) stored in the system database. An individual may recognize one’s identity with the help of PIN, a user name, or a smart card. Here the biometric system performs a one to one matching to determine whether person’s individuality is correct or not. Identity verification is mainly used for positive recognition. The objective of the individuality verification is to avert several persons from consuming the similar uniqueness. The system recognizes an individual by searching in the verification templates of all the users in the database for a match in the identification approach. Therefore, the system performs a one-to-many matching to establish an individual’s identity (or fails if the subject is not enrolled in the system database) without the subject having to claim an identity.

Abbildung in dieser Leseprobe nicht enthalten

Figure 1.1 Working principle of biometric system (Simon & Mark 2001)

The various steps involved in Figure 1.1 (Simon & Mark 2001) are given below:

Step 1: Capture the chosen biometric;

Step 2: Process the biometric and extract and enroll the biometric templates;

Step 3: Store the template in a local repository, or a portable token such as a smart code;

Step 4: Live scan the chosen biometric;

Step 5: Process the biometric and extract the biometric template;

Step 6: Match the scanned biometric template against stored templates;

Step 7: Provide a matching score to business application;

Step 8: Record a secure audit trail with respect to system use.

The biometric system is divided into four main modules.

1. Sensor Module: It captures the biometric trait of a person.
2. Feature Extraction Module: The biometric trait obtained from the sensor module is processed to extract a set of or salient or discriminatory features.
3. Pattern Matching Module: The information mined in the feature extraction module is matched with the templates to generate the matching scores.
4. System database module: It is used in the biometric system to store the biometric templates of the enrolled users. The enrolment module is responsible for enrolling individuals in the biometric system database. The biometric reader is used to scan the biometric traits of an individual to produce the digital representation or feature values of the biometric characteristics during the enrolment phase. The data captured during the enrolment process may or may not be supervised by human depending on the application. An eminence testing is usually achieved to ensure that the acquired sample is relatively processed by successive stages. The feature extractor is used to process the digital representation for facilitate the matching to generate compact but expensive representation is called as template. The template is stored in the central database of a biometric system depending on the application. The templates are also recorded on a smart card issued to the individual. Usually, different templates of an individual are stored to account for variations observed in the biometric trait and the templates in the database may be updated over time.

The two different techniques to measure the biometric accuracy are the False Acceptance Rate (FAR) and False Rejection Rate (FRR). The limited entry is allowed to authorize the users by two methods focussed on the system’s ability. The sensitivity of the mechanism is adjusted whatever matches to the biometrics. Based on that sensitivity, the biometric measures can vary significantly.

1.2 Palmprint Biometrics

Palmprint verification is implemented in different way compared to the fingerprint technology. The optical readers used in fingerprint technology are used in palmprint scanning. The size of the palmprint scanner is bigger. It has a limiting factor when used in workstations or mobile devices. The palms of the human hands contain pattern of ridges and valleys much like the fingerprints. The region of the palm is greatly higher than the region of a finger. Therefore palmprints are more distinctive than the fingerprints. The palmprint scanner is used to capture the large area of the palm. The low resolution scanner is used to capture the additional distinctive features such as principal lines and wrinkles in the palmprint. It is very cheap. Finally, it is used to capture all the features of the palmprint such as hand geometry, ridge and valley features (e.g., minutiae and singular points such as deltas), principal lines, and wrinkles.

Palmprint recognition inherently implements many of the same matching characteristics that have allowed fingerprint recognition to be one of the most well-known and best publicized biometrics. Both palm and finger biometrics is represented by the information presented in a friction ridge imprint. The palms and fingerprints are used as a trusted form of identification for more than a century. The image captured from the palm region of the hand refers to the palmprint. The image captured from a scanner or Charge Coupled Device (CCD) is known as online image. The image taken with the help of ink and paper are known as offline image. The palm itself consists of principal lines, wrinkles (secondary lines) and epidermal ridges. The palmprint features are different from fingerprint features. The palmprint also contains other features such as indents and marks. These features are used to compare one palm with another palm. Palmprints are used for illegal, pathological, or profitable applications.

The palmprint pattern cannot duplicate with the other people, even in monozygotic twins. Hence palmprint is used as high reliable human identifier. The details of the palmprint ridges are stable. The information remains unchanged from that time on throughout life, excluding for scope. After the demise, disintegration of the skin occurred lastly in the area of the palmprint. Matched with the other physical biometric features, palmprint verification has several benefits:

- Low-resolution imaging
- Low-intrusiveness
- Stable line features
- Low-cost capturing device

Palmprint covers wider area than fingerprint. It contains useful information for recognition. The dominant lines on the palmprint are known as three principal lines. The weaker and more irregular lines on the palm are known as wrinkles. The palmprint biometric system do not require very high resolution acquisition device. The principle lines and wrinkles are also acquired using low resolution acquisition device like 100 dpi (dots per inch) or lesser.

1.2.1 Preprocessing and ROI Extraction for Palmprint Biometrics

A palmprint region is extracted by pre-processing the acquired hand image. The square area inside the palm region of the hand image is considered as the palmprint or Region of Interest (ROI). Due to the regular and controlled uniform illumination conditions during image acquire, the attained hand image and its background contrast in colour.The sample input image is shown in the Figure 1.2(a) (Zhang et al 2003), 1.3(a) (Rohit Khokher et al 2014) and 1.4(a) (Badrinath & Gupta 2010) for PolyU Palmprint database, COEP Palmprint database, IIT - Delhi Palmprint database. The hand from the background is extracted by applying the Global Thresolding. Opening and Closing morphological operations is used to eliminate any isolated small blobs or holes. The contour of the hand image from the palmprint is acquired by applying the Contour-tracing algorithm. The ROI extraction for PolyU Palmprint database, COEP Palmprint database and IIT Delhi Palmprint database are the same as of Figure 1.2(b) (Zhang et al 2003), 1.3(b) (Rohit Khokher et al 2014) and 1.4(b), (Badrinath & Gupta 2010) are shown respectively.

Abbildung in dieser Leseprobe nicht enthalten

Figure 1.2 (a) Sample input image (Zhang et al 2003) , (b) ROI image for PolyU palmprint database (Zhang et al 2003)

Abbildung in dieser Leseprobe nicht enthalten

Figure 1.3 (a) Sample input image (Rohit Khokher et al 2014), (b) ROI image for COEP palmprint database (Rohit Khokher et al 2014)

Abbildung in dieser Leseprobe nicht enthalten

Figure 1.4 (a) Sample input image (Badrinath & Gupta 2010), (b) ROI image for IIT Delhi palmprint database (Badrinath & Gupta 2010)

1.3 Finger knuckle-print biometrics

The biometric analysts have less specialize in the Finger Knuckle-print (FKP) which is shown in the research field. The FKP biometric system provides a high level of security to human identifier. The image pattern of the skin on the rear surface of the finger is known as FKP. It is not a popular biometric recognition system compared to fingerprint biometrics system. The matching of finger knuckle patterns helps to spot the suspects and find out validating scientific proof from the images. Once if there is no information about fingerprints or face, then it is better among the market images. Choosing the biometrics is a challenging task for researcher. As it is contactless, there is chance for less proof of physical presence i.e. antispoofing. Finger knuckle-print has high textured region. Many samples are available per hand and independent to any behavioural aspect.

1.3.1 Finger Knuckle-print Anatomy

Each finger has three joints. The proximal phalanges, the centre phalanges and the distal phalanges are the three bones in each finger. The proximal phalanx is the first join where the finger gets join the hand. The Proximal Interphalangeal Joint (PIP) is the second joint. The Distal Interphalangeal Joint (DIP) is the last joint of the finger as shown in Figure 1.5 (Kulkarni & Rout 2012). The image pattern of the skin on the back surface of the finger is known as the finger knuckle-print. The Finger knuckle is also known as dorsum of the hand. The inherent skin pattern of the outer surface around the phalange joint of one’s finger has high capability to discriminate completely different people. Such image pattern of finger knuckle-print is unique and might be getting on-line, offline for authentication. Extraction of knuckle features for the identification is completely depends upon the user.

Abbildung in dieser Leseprobe nicht enthalten

Figure 1.5 Finger knuckle-print anatomy (Sivaranjani et al 2014)

Several investigators of science mined the options for authentication as shown in Figure 1.6 (Kulkarni & Rout 2012). A pair of features in FKP is centre of phalange joint, U formed line round the middle phalanx, number of lines, length and spacing between lines. Knuckle crease patterns and stray marks are considered as a method of photographic identification. Such features are unique and it is used for an identification process.

Abbildung in dieser Leseprobe nicht enthalten

Figure 1.6 Finger knuckle print features (Kulkarni & Rout 2012)

The contact free imaging of the finger back surface is highly convenient to users. The images can also be acquired in online using either scanner or CCD camera. The acquired images are used to extract scale, translation and rotational invariant knuckle features for user identification (Kumar & Zhou 2009). It is reported that the skin pattern on the FKP is highly rich in texture due to skin wrinkles and creases and henceforth, it is measured as a biometric identifier. Further, advantages of using FKP include rich in texture information, simply available, contact-less image acquisition, invariant to emotions and other behavioural aspects such as fatigue, stable information and adequacy in the society. Despite of these characteristics and advantages of using FKP as biometric identifier, limited work are reported in the literature. The usage of finger knuckle for personal identification is shown in the ensuring results and generated a keen interest in biometrics. However, the research efforts to investigate the utility of finger knuckle patterns for personal identification were much restricted. As a result; there is no recognized use of knuckle pattern in commercial or civilian applications. The user acceptance for employing finger knuckle in human identification is expected to be very high (Kumar & Zhou 2009).

1.3.2 Preprocessing and ROI Extraction for Finger Knuckle-Print Biometrics

FKP images collected from different fingers are extremely assorted. The spatial locality is different for various FKP images. Therefore each FKP image is aligned by constructing the local coordinate system. Figure 1.7 (a) (Zhnag et al 2011) shows the FKP image sensor device and Figure 1.7 (b) (Zhnag et al 2011) shows a sample finger knuckle-print image. Figure 1.8 (c) (Zhnag et al 2011) and Figure 1.8 (d) (Zhnag et al 2011) shows ROI extraction technique and extracted image respectively.

Abbildung in dieser Leseprobe nicht enthalten

Figure 1.7 FKP (a) Image sensor device (Zhnag et al 2011), (b) Sample image (Zhnag et al 2011)

Abbildung in dieser Leseprobe nicht enthalten

Figure 1.8 (c) ROI extraction technique (Zhnag et al 2011) (d) ROI image (Zhnag et al 2011)

1.4 Pros of finger knuckle-print and palmprint

- No expression, pose and ageing.
- No occlusion, less cooperation, inexpensive sensors.

[...]