New Directions in Behavioral Biometrics

Khalid Saeed

• Publisher: CRC Press
• Year: 2016
• eBook ISBN: 9781315349312
• Edition: 1
• Topic: Computer Science
• Subtopic: Computer Science General
• Index: Computer Science

Automatic biometrics recognition techniques are increasingly important in corporate and public security systems and have increased in methods due to rapid field development. This book discusses classic behavioral biometrics as well as collects the latest advances in techniques, theoretical approaches, and dynamic applications. This future-looking book is an important reference tool for researchers, practitioners, academicians, and technologists. While there are existing books that focus on physiological biometrics or algorithmic approaches deployed in biometrics, this book addresses a gap in the existing literature for a text that is solely dedicated to the topic of behavioral biometrics.

1

INTRODUCTION TO BEHAVIORAL BIOMETRICS

Biometrics refers to the study of biological characteristics. It comes from Greek words “bios” (implies life) and “metricos” (implies measuring/to measure). Biometrics can be considered as use of physiological or behavioral characteristics in an automated way to determine identity. The identity verification is performed through the measurement of physiological or behavioral characteristics of an individual. Researchers have proposed a number of biometric techniques for human identification and authentication based on fingerprint, palm print, hand geometry, face, ear, iris, retina, voice, signature, body odor, and so on.

Biometric traits are almost statistical in nature. The system is likely to be unique and reliable as much as data are available from sample. It can work on various modalities based on the measurements of individual’s body and features, and behavioral patterns. The modalities are classified according to individual’s biological traits. The biometric modalities mainly fall under following types:

• Physiological

• Behavioral

• Hybrid

Physiological category includes the features we are born with. This modality is based on the shape and size of the body. Examples are [1–3] as follows:

• Fingerprint recognition

• Hand geometry recognition

• Facial recognition

• Iris recognition

• Retinal scanning

• DNA recognition

Behavioral category deals with the features we learn in our life as a result of our interaction with the environment and the nature. This modality is related to change in human behavior over time. Examples of this category are

• The way we walk (gait).

• The way we write (signature).

• The way we speak or say a word (voice).

• The way we type on a machine (keystroke dynamics).

• Many other ways of our response to the natural events around us and the way we react to or respond.

Hybrid modality includes both traits, where the traits are depending upon physical as well as behavioral changes. As for example, voice recognition may be considered as hybrid modality as it depends on size and shape of vocal cord, nasal cavities, mouth cavity, shape of lips, and so on, and the emotional status, age, illness (behavior) of a person.

Hybrid modality is also considered as a type of multimodality (more than one mode involved). In this book, the concept of behavioral biometrics is presented briefly on the basis of some selected features like signature, keystroke dynamics, gait, and voice. A brief overview of behavioral biometrics is presented in this chapter.

1.1 Behaviometrics

The word “behaviometrics” derives from the terms “behavioral” and “biometrics” [4]. Behavioral refers to the way an individual behaves. Behaviometrics, or behavioral biometrics, is a measurable behavior used to recognize or verify the identity of a person. Behaviometrics focuses on behavioral patterns rather than physical attributes.

1.1.1 How It Works

Each person has a unique pattern—how they interact with computing devices by using keyboard, mouse, and graphical user interface (GUI). The study of the user’s unique nature in this regard is known as behaviometrics.

A human behavioral pattern consists of a variety of different unique behaviors—all are mixed together into a larger unique profile. Since unique behaviometric pattern of every person is formed not only by biometric features, but is also influenced by social and psychological means, it is just impossible to copy somebody else’s behavior.

The behavioral pattern of the person is compared with the stored pattern. Matching scores of similarities for those users are recognized and the software calculates the possibility of accurate identification of users.

The key features of behavioral biometrics are given below.

• Security of applications like user authentication and intrusion detection may be enhanced by behavioral biometrics with very low impact on the users.

• Behavioral biometrics is highly sensitive to the means of implementation, for example, keystroke dynamics depend on the type of used keyboard.

• Behavioral biometrics is most useful in multimodal systems (where more than one type of biometrics is used at the same time) compared to unimodal systems (where only one type of biometrics is used at a time).

• It may be vulnerable to several spoofing attacks [5].

A comprehensive review of different biometric technologies including theory and applications can be found in [1]. A survey of different techniques on behavioral biometrics is summarized in [6]. Behavioral biometrics is popularly used in information security context to identify individuals by using unique features of activities they perform either consciously or unconsciously. In recent times, it has been observed that behavioral biometric data are used for a number of interesting applications. Researchers have proposed methodologies for speaker recognition by tracking movements of lips [7], biometric verification using motions of fingers [8], and extracting biometric features of voice [9] for person identification. A promising application of biometrics is artimetrics, where biometric traits are used for authenticating artificial entities like industrial robots, intelligent software agents, and virtual-world avatars [9]. Biometric data are also used for enhancing the security of cryptographic systems. New algorithms are developed by researchers for filtering biometrics. Performance evaluation of systems is very important for the following reasons.

• Quality of system must be precisely quantified to be used in real context. To determine whether it fulfills the requirements of a specific application based on logical or physical access, context of use, efficiency, and robustness of the logic must be defined.

• Comparison of different biometric modalities is essential to analyze their relative merits and demerits.

• Performance evaluation is also necessary in order to facilitate research in this field.

Evaluation techniques are used to quantify the performance of behavioral biometric systems. A reliable evaluation method is needed in order to analyze advantage of the system.

1.1.2 Major Benefits

Behaviometrics can provide information security solutions by using the nature of individual. It is extremely hard to replicate, which makes it the ultimate solution against identity theft. It is not possible that any unauthorized user could access a computer with confidential information, either by hacking the password or logging in with stolen credentials or accessing a logged on computer. As a result, intrusion can be prevented.

As for example, it is possible to recognize and confirm the identity of a person by analyzing how the user works with the keyboard (typing rhythm), mouse movements (acceleration time, click frequencies), and graphical interface interaction (using programs).

• While many popular security solutions require the user to perform additional tasks, behaviometrics does not interfere with the normal work flow. Simple use of computing device in the everyday work makes the software increasingly more efficient and the confidential information more secure.

• Behaviometrics will allow workstations to be secure even after the user has logged on to the system. Even if the user leaves the workstation and forgets to sign out, computing device stays protected.

• Existing token-based products (such as passwords and smart cards) can be duplicated or stolen, whereas user’s behavior is unique and very difficult to copy.

1.2 What Is Special about Behavioral Biometrics Data Acquisition?

The special aspects of behavioral biometrics data acquisition are presented below.

• It offers increased convenience in data acquisition, because there is no requirement for dedicated or special hardware. As a result, it is also considered as cost-effective.

• Most of the data are acquired through machine-based interactions.

• These traits need to be easily verifiable and identifiable.

• Input data depend on the permanence and distinctiveness metrics.

• It does not introduce delays in operations and are implemented silently. It is mostly used in online platforms. Their acceptance level in the society is high.

1.3 Behavioral Biometrics Features

A biometric system may include different phases. Two phases are mainly considered during the use of a biometric system. A working model of data acquisition is defined in enrollment phase of an individual. Verification phase uses this model to make a decision about an individual. The performance evaluation of a biometric system generally considers the quality of the input data and the output result.

In accomplishing their everyday tasks [1], people employ different strategies, use different styles and apply unique skills and knowledge. One of the defining characteristics of a behavioral biometric is the incorporation of time dimension as a part of the behavioral signature. The measured behavior has a beginning, duration, and an end. Researchers attempt to quantify behavioral traits exhibited by users and use resulting features to verify identity efficiently.

Behavioral biometrics provides a number of advantages. They can be collected without the knowledge of the user. It becomes very cost-effective as collection of behavioral data often does not require any special hardware. Data acquisition devices include computer, keyboard, mouse, stylus, touch screen, microphone, camera, credit card, and scanner to capture most frequently used behavioral biometrics. Although most of the behavioral biometric traits may not be unique enough to provide reliable human identification, it is observed that they can provide sufficiently high accuracy in identity verification.

Behavioral biometric systems are requirement specific. Many characteristics make them difficult to analyze their performance [10]:

• Biometric template generally contains temporal information.

• This type of template can change with time. It means that the biometric template can be quite different compared to the one obtained after the enrollment phase.

• The behavior of biometric char...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Dedication
• Table of Contents
• FOREWORD
• PREFACE
• ACKNOWLEDGMENTS
• CHAPTER 1 INTRODUCTION TO BEHAVIORAL BIOMETRICS
• CHAPTER 2 SIGNATURE RECOGNITION
• CHAPTER 3 KEYSTROKE DYNAMICS
• CHAPTER 4 GAIT ANALYSIS
• CHAPTER 5 VOICE RECOGNITION
• INDEX