Schweitzer Fachinformationen
Wenn es um professionelles Wissen geht, ist Schweitzer Fachinformationen wegweisend. Kunden aus Recht und Beratung sowie Unternehmen, öffentliche Verwaltungen und Bibliotheken erhalten komplette Lösungen zum Beschaffen, Verwalten und Nutzen von digitalen und gedruckten Medien.
With an increasing demand for biometric systems in various industries, this book on multimodal biometric systems, answers the call for increased resources to help researchers, developers, and practitioners.
Multimodal biometric and machine learning technologies have revolutionized the field of security and authentication. These technologies utilize multiple sources of information, such as facial recognition, voice recognition, and fingerprint scanning, to verify an individual's identity. The need for enhanced security and authentication has become increasingly important, and with the rise of digital technologies, cyber-attacks and identity theft have increased exponentially. Traditional authentication methods, such as passwords and PINs, have become less secure as hackers devise new ways to bypass them. In this context, multimodal biometric and machine learning technologies offer a more secure and reliable approach to authentication.
This book provides relevant information on multimodal biometric and machine learning technologies and focuses on how humans and computers interact to ever-increasing levels of complexity and simplicity. The book provides content on the theory of multimodal biometric design, evaluation, and user diversity, and explains the underlying causes of the social and organizational problems that are typically devoted to descriptions of rehabilitation methods for specific processes. Furthermore, the book describes new algorithms for modeling accessible to scientists of all varieties.
Audience
Researchers in computer science and biometrics, developers who are designing and implementing biometric systems, and practitioners who are using biometric systems in their work, such as law enforcement personnel or healthcare professionals.
Sandeep Kumar, PhD, is a professor in Computer Science & Engineering, Koneru Lakshmaiah Educational Foundation, India, He has published more than 150 journal articles and conference papers, 20 patents, and authored 13 books.
Deepika Ghai, PhD, is an assistant professor at Lovely Professional University, India. She has published more than 35 research papers in refereed journals and conferences. She received the Dr. C.B. Gupta Award in 2021 at Lovely Professional University.
Arpit Jain, PhD, is a professor at the Koneru Lakshmamai University Education Foundation, Vijayawada, A.P., India. He has published more than 40 research papers in international journals, filed 25+ patents as well as authored/edited five books.
Suman Lata Tripathi, PhD, is a professor at Lovely Professional University with more than 21 years of experience in academics. She has published more than 105 research papers in refereed journals and conferences. She has published three books and currently has multiple volumes scheduled for publication from Wiley-Scrivener.
Shilpa Rani, PhD, is an associate professor at the Neil Gogte Institute of Technology, Hyderabad, India, and specializes in computer science & engineering. She has authored seven books, more than 50 journal articles conference papers, as well as 14 patents.
Preface xiii
1 Multimodal Biometric in Computer Vision 1Sunayana Kundan Shivthare, Yogesh Kumar Sharma and Ranjit D. Patil
1.1 Introduction 2
1.2 Importance of Artificial Intelligence, Machine Learning and Deep Learning in Biometric System 2
1.3 Machine Learning 4
1.4 Deep Learning 6
1.5.1 Discussions 11
1.6 Biometric System 11
1.7 Need for Multimodal Biometric 15
1.8 Databases Used by Biometric System 17
1.9 Impact of DL in the Current Scenario 19
1.10 Conclusion 24
2 A Vaccine Slot Tracker Model Using Fuzzy Logic for Providing Quality of Service 31Mohammad Faiz, Nausheen Fatima and Ramandeep Sandhu
2.1 Introduction 32
2.2 Related Research 33
2.3 Novelty of the Proposed Work 37
2.4 Proposed Model 38
2.5 Proposed Fuzzy-Based Vaccine Slot Tracker Model 42
2.6 Simulation 44
2.7 Conclusion 47
2.8 Future Work 50
3 Enhanced Text Mining Approach for Better Ranking System of Customer Reviews 53Ramandeep Sandhu, Amritpal Singh, Mohammad Faiz, Harpreet Kaur and Sunny Thukral
3.1 Introduction 53
3.2 Techniques of Text Mining 55
3.3 Related Research 58
3.4 Research Methodology 63
3.5 Conclusion 67
4 Spatial Analysis of Carbon Sequestration Mapping Using Remote Sensing and Satellite Image Processing 71Prashantkumar B. Sathvara, J. Anuradha, R. Sanjeevi, Sandeep Tripathi and Ankitkumar B. Rathod
4.1 Introduction 72
4.2 Materials and Methods 75
4.3 Results 77
4.4 Conclusion 79
5 Applications of Multimodal Biometric Technology 85Shivalika Goyal and Amit Laddi
5.1 Introduction 85
5.2 Components of MBS 87
5.3 Biometrics Modalities 89
5.4 Applications of Multimodal Biometric Systems 89
5.5 Conclusion 97
6 A Study of Multimodal Colearning, Application in Biometrics and Authentication 103Sandhya Avasthi, Tanushree Sanwal, Ayushi Prakash and Suman Lata Tripathi
6.1 Introduction 104
6.2 Multimodal Deep Learning Methods and Applications 108
6.3 MMDL Application in Biometric Monitoring 113
6.4 Fusion Levels in Multimodal Biometrics 116
6.5 Authentication in Mobile Devices Using Multimodal Biometrics 119
6.6 Challenges and Open Research Problems 122
6.7 Conclusion 123
7 A Structured Review on Virtual Reality Technology Application in the Field of Sports 129Harmanpreet Kaur, Arpit Kulshreshtha and Deepika Ghai
7.1 Introduction 130
7.2 Related Work 132
7.3 Conclusion 142
8 A Systematic and Structured Review of Fuzzy Logic-Based Evaluation in Sports 145Harmanpreet Kaur, Sourabh Chhatiye and Jimmy Singla
8.1 Introduction 146
8.2 Related Works 148
8.3 Conclusion 159
9 Machine Learning and Deep Learning for Multimodal Biometrics 163Danvir Mandal and Shyam Sundar Pattnaik
9.1 Introduction 163
9.2 Machine Learning Using Multimodal Biometrics 165
9.3 Deep Learning Using Multimodal Biometrics 167
9.4 Conclusion 169
10 Machine Learning and Deep Learning: Classification and Regression Problems, Recurrent Neural Networks, Convolutional Neural Networks 173R. K. Jeyachitra and Manochandar, S.
10.1 Introduction 174
10.2 Classification of Machine Learning 174
10.3 Supervised Learning 175
10.4 Unsupervised Learning 201
10.5 Reinforcement Learning 203
10.6 Hybrid Approach 204
10.7 Other Common Approaches 205
10.8 DL Techniques 210
10.9 Conclusion 219
11 Handwriting and Speech-Based Secured Multimodal Biometrics Identification Technique 227Swathi Gowroju, V. Swathi and Ankita Tiwari
11.1 Introduction 228
11.2 Literature Survey 230
11.3 Proposed Method 231
11.4 Results and Discussion 237
11.5 Conclusion 248
12 Convolutional Neural Network Approach for Multimodal Biometric Recognition System for Banking Sector on Fusion of Face and Finger 251Sandeep Kumar, Shilpa Choudhary, Swathi Gowroju and Abhishek Bhola
12.1 Introduction 252
12.2 Literature Work 253
12.3 Proposed Work 256
12.4 Results and Discussion 260
12.5 Conclusion 265
13 Secured Automated Certificate Creation Based on Multimodal Biometric Verification 269Shilpa Choudhary, Sandeep Kumar, Monali Gulhane and Munish Kumar
13.1 Introduction 270
13.2 Literature Work 274
13.3 Proposed Work 276
13.4 Experiment Result 278
13.5 Conclusion and Future Scope 279
14 Face and Iris-Based Secured Authorization Model Using CNN 283Munish Kumar, Abhishek Bhola, Ankita Tiwari and Monali Gulhane
14.1 Introduction 284
14.2 Related Work 285
14.3 Proposed Methodology 287
14.4 Results and Discussion 291
14.5 Conclusion and Future Scope 296
References 297
Index 301
Sunayana Kundan Shivthare1*, Yogesh Kumar Sharma2 and Ranjit D. Patil3
1MAEER's MIT Arts, Commerce and Science College, Alandi, Pune, Maharashtra, India
2Department of Computer Science and Engineering, Koneru Lakshmaiah Education Foundation, Vaddeswaram, Guntur, AP, India
3Dr D.Y. Patil ACS College, Pimpri, Pune, Maharashtra, India
In conjunction with the growing requirement for security regulations and information security worldwide, biometric technology is more prevalent daily than ever. Multimodal biometrics technology has gained popularity due to overcoming several significant drawbacks of unimodal biometric systems. Using numerous biometric markers by personal identification systems to identify individuals is multimodal biometrics. Unlike unimodal biometrics, which uses only one biometric feature, such as a fingerprint, face, palm print, or iris, multimodal authentication is more secure different biometrics systems aid in confirming that only authentic users are using the services. Using cutting-edge approaches like ML, computer vision, object detection and recognition, image analysis pattern recognition, and CNN is the general idea behind biometric identification methods. Machine learning and deep learning are widespread fields in today's digital era. While surfing the Internet, algorithms for machine learning and deep learning are used in every aspect of the online world. This shows that these fields have become an inseparable part of our lives. Abundance data produced through online mediums are classified through these techniques. In computer vision, these algorithms have prominently left their footprints. Deep learning is a subset of machine learning that studies and applies artificial neural networks (ANNs). Deep learning is at the heart of modern artificial intelligence, and its applications rapidly spread across industries and domains. In this chapter, the authors have tried to illuminate applications of machine learning and deep learning concepts and algorithms in connection with multimodal biometrics.
Keywords: Machine learning, deep learning, computer vision, biometric systems, multimodal biometrics, authentication
Biometrics is the scientific method of identifying individuals based on their qualities or characteristics [1]. It is also used to locate people in groups that are being watched. Unique, quantifiable qualities called biometric identifiers identify and specifically define people. Biometric systems are crucial to finding a person and increasing worldwide security. Many biometrics, including height, DNA, handwriting, and others, could be used, but computer vision-based biometrics have become increasingly significant in human identification [2, 3]. The ability to identify a person's face, fingerprints, iris, and other biometrics using computer vision is used to build effective authentication systems.
Personal biometric authentication has developed into a necessary and in-demand technology with the development of intelligent artificial systems and e-technologies today. Artificial neural networks (ANN) with numerous hidden layers have been designed to extract low-level to abstract-level characteristics for deep learning (DL), a new subcategory of machine learning. DL approaches include distributed and parallel data processing, adaptive feature learning, dependable fault tolerance, and resilient resilience properties [4]. It is extensively employed in buildings, airports, mobile phones, identity cards, etc. Robust recognition systems must be learned using biometric data. A person can be recognized by various physical characteristics (hand geometry, fingerprint, face, palm print, iris, and ear) and behavioral factors (such as gait, signature, and voice). These qualities can be utilized to separate one person from another and will not be forgotten or lost with time [5]. Combining two or more of these traits helps to increase security, demonstrate excellent performance, and address the shortcomings and limitations of unimodal biometric systems.
Machine learning techniques have been used for recognition by several biometrics researchers. Before classifying the raw biometric data, machine learning algorithms must transform it into a suitable format and extract its characteristics. Before feature extraction, machine learning techniques call for some preprocessing operations [6].
Deep learning has recently had a significant impression and achieved outstanding achievements in biometrics systems. A lot of the drawbacks of traditional machine learning techniques, particularly those related to feature extraction methods, have been solved by deep learning algorithms. Deep learning techniques can handle biometric image changes, take raw data, and extract features [7-9].
Language is the primary means of inter-human communication. Humans are superior to all living things because they communicate using language. Humans can communicate via language because they have the senses of sight and sound. The idea for the intelligent thinking machine was closely related to the inspiration for the computer's invention. The five senses of sight, hearing, smell, taste, and touch enable us to observe, comprehend, appreciate, and engage with our environment [10, 11]. The two senses most contributing to a human's intelligence are sight and sound. The human brain receives information about the company and sound through the eyes and hearing, processes it, and then executes the required actions [12].
Artificial intelligence (AI) should be able to process human language and auditory and visual input. Artificial intelligence was also being developed concurrently with creating generic software consisting of wholly programmed instructions and logic. Making software that simulates the human brain was one of the goals of researchers and programmers in artificial intelligence. Artificial intelligence underwent a revolution with the creation of Deep Neural Networks and the necessary sophisticated technology to process enormous amounts of data. When computers first entered the world in the early 1900s, they were utilized to solve complicated equations [13, 14]. Later, when other technology emerged, people began to view computers as more than just calculators. One of the leading technologies that are replacing human labour is artificial intelligence. Deep learning is a subdomain of artificial intelligence subset of machine learning and was first introduced in 1943 [1, 15, 16].
These deep neural networks contributed to the ability of computers to process speech, images, videos, and other types of natural language. Deep learning is the name of the branch of computer science that deals with these deep neural networks. This chapter aims to explore many facets of machine and deep learning. Deep learning aims to use mathematical algorithms to learn how human brain networks work. Deep learning was created with the core goal of simulating the complicated cognitive process of the human brain to empower machines with independent thought and decision-making [3, 17, 18]. Deep learning is a method for using neural networks to handle massive amounts of data. Several issues in natural language processing, image recognition, and speech recognition can be solved best by this stage. One of the key benefits of adopting deep learning over different machine learning algorithms is that it can create new features from a small number of existing characteristics in the training data set. As a result, deep learning algorithms can solve current problems by creating new tasks [4, 19, 20].
Artificial neural networks, a type of algorithm used in deep learning, are inspired by the structure and operation of the brain. An input layer, a hidden layer, and an output layer make up artificial neural networks. Deep neural networks, which feature numerous hidden layers, are the more sophisticated iterations of artificial neural networks. Deep learning mimics how human brains work, in other words. The nervous system's structure, where every neuron is connected to the others and advances different input types, is precisely how the deep learning algorithm works [21]. The layer system in deep learning is its best asset. Between machine learning and deep learning, there are significant differences. While deep learning models tend to perform exceptionally well with massive data collections and continuously improve, machine learning models reach a saturation point where they cease improving [22]. The feature extraction zones are the other distinction in the future. Machine learning requires people to extract features every single time manually, but deep learning models can learn on their own without human intervention [7]. More significance is placed on computation power in the deep learning process [23, 24]. It depends on our layers; the necessary GPU and CPU quantity are required if the coating is practical. Otherwise, it may be challenging to obtain the outcome after a day, a month, or even a year [8, 9].
The system is given input in traditional programming, producing output based on the logic. In machine learning, input and output are provided to the system, and models are built using machine learning algorithms. That model makes predictions and solves...
Dateiformat: ePUBKopierschutz: Adobe-DRM (Digital Rights Management)
Systemvoraussetzungen:
Das Dateiformat ePUB ist sehr gut für Romane und Sachbücher geeignet – also für „fließenden” Text ohne komplexes Layout. Bei E-Readern oder Smartphones passt sich der Zeilen- und Seitenumbruch automatisch den kleinen Displays an. Mit Adobe-DRM wird hier ein „harter” Kopierschutz verwendet. Wenn die notwendigen Voraussetzungen nicht vorliegen, können Sie das E-Book leider nicht öffnen. Daher müssen Sie bereits vor dem Download Ihre Lese-Hardware vorbereiten.Bitte beachten Sie: Wir empfehlen Ihnen unbedingt nach Installation der Lese-Software diese mit Ihrer persönlichen Adobe-ID zu autorisieren!
Weitere Informationen finden Sie in unserer E-Book Hilfe.
