Scribd
Upload
47 views10 pages

0TH PPT 1

Uploaded by

saranlovely433
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
47 views10 pages

0TH PPT 1

Uploaded by

saranlovely433
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 10

 • BATCH NO : ECECK6242505

 • Ganugula Satya Durga Saran (21K61A0453)


 • Lokavarapu Sai Kiran (21K61A0494)
 • Bandreddy Mukesh (21K61A0417)
 • Bokka Sai (21K61A0423)

 PROJECT GUIED : Mrs. P.BHAVANA (ASST PROFESSOR)


Deep learning based
Biometric Authentication
using Finger Veins
Contents
 Abstract
 Introduction
 Literature Survey
 Problem Statement
 References
Abstract
 The biometric authentication is improved over the
recent last years to enhance the security.
 Face Recognition and AI based authentication are
regular models to validate the human presence.
 The attackers can hack the above mentioned
biometrics so that there is a need to protect the users
data with other biometrics like Finger Veins.
 In this work, we are proposing that Finger Veins as they
are presented under the skin so there is less chance to
hack and also economically viable.
 The feature extraction and registration process is
carried out by the deep learning techniques like
Convolution neural networks (CNN).
Introduction
 Biometric authentication involves majorly two steps
 Step1: Registration:: The process of adding the
biometrics of an individual to the data base so that the
features are extracted to identify the individual in
future.
 Step2: Validation:: Once the registration completed,
the next step is to identify that particular individual
using the stored biometrics in real time which is the
Validation.
 The different biometrics for these process includes
Face, Iris, Fingerprint and Palm.
 Each of it having its own pros and cons, but finger vein
is well suitable for the improving the security .
Introduction (continued)
 Unlike the fingerprint, the finger vein has its
own advantages.
 The fingerprint can also be accessed by the

hackers by the retinal fingerprint scanners.


 The amount of protection provided by

various biometric features also varies.


 It is more difficult to deceive the finger-vein

based biometric verification system


Literature Survey
 The advent of machine learning introduced significant improvements.
Methods incorporating Support Vector Machines (SVM) and Random
Forests were explored, as seen in the work by Xu et al. (2014), which
demonstrated better performance over traditional methods [1].
However, these methods still had limitations in handling complex vein
patterns and variations.
 The application of deep learning to finger vein recognition began
gaining traction in recent years. Deep learning models, particularly
Convolutional Neural Networks (CNNs), have shown promise in
automatically learning feature representations from raw data.
 A study by Li et al. (2018) demonstrated that CNNs could outperform
traditional methods by learning intricate vein patterns with higher
accuracy [2].
 Recent research has focused on improving network architectures. Zhang
et al. (2020) introduced a novel deep CNN architecture that integrates
attention mechanisms to better focus on relevant features in vein
patterns, leading to substantial improvements in recognition accuracy
[3].
 To address the issue of limited training data, techniques such as
data augmentation and synthetic data generation have been
explored. Kim et al. (2021) proposed a method combining
generative adversarial networks (GANs) for generating synthetic
finger vein images, which enhanced the training process and model
generalizability [4].
 The integration of multiple biometric modalities with finger vein
recognition has also been explored. Wu et al. (2022) investigated
multi-modal systems combining finger vein with other biometrics
like fingerprint and iris, showing improved performance and
robustness in authentication tasks [5].
 Real-world applications require systems to be robust against
variations in lighting, user position, and finger condition. Recent
advancements by Zhao et al. (2023) addressed these challenges by
incorporating adaptive algorithms and robust preprocessing
techniques, significantly enhancing system performance in practical
scenarios [6].
Problem Statement
 The primary problem this research aims to
address is the development of a robust deep
learning-based approach for finger vein
recognition that can handle diverse and
challenging conditions encountered in practical
scenarios.
 This involves designing and training deep neural
network architectures that enhance recognition
accuracy, incorporating data augmentation
techniques to improve model generalizability,
and developing methods to handle real-world
variations effectively.
References
1. Xu, X., Zhang, H., & Liu, Y. (2014). "Finger vein recognition using
machine learning techniques." Pattern Recognition, 47(1), 150-160.
2. Li, W., Zhang, X., & Yang, L. (2018). "Deep convolutional neural
networks for finger vein recognition." IEEE Transactions on
Information Forensics and Security, 13(2), 304-315.
3. Zhang, Q., Wei, X., & Zhou, Z. (2020). "Attention-based deep
learning for finger vein recognition." Neurocomputing, 385, 171-179.
4. Kim, J., Lee, S., & Park, D. (2021). "Synthetic finger vein image
generation using GANs for enhanced biometric authentication."
International Journal of Computer Vision, 129(6), 1024-1037.
5. Wu, Y., Liu, L., & Zhang, H. (2022). "Multi-modal biometric
authentication system combining finger vein, fingerprint, and iris."
Computers, Environment and Urban Systems, 90, 101731.
6. Zhao, Q., Li, L., & Wang, Z. (2023). "Robust finger vein recognition in
practical conditions using adaptive algorithms." Journal of Visual
Communication and Image Representation, 92, 103126.

You might also like