Deep Representation Based Feature

Extraction And Recovering For

Finger-vein Verification
 OBJECTIVE

• To extract and recover vein features using limited a priori knowledge.

• To automatically discard the ambiguous region and to label the pixels of the
clear region as foreground or background.

• To recover missing finger-vein patterns in the segmented image.

 INTRODUCTION

• The tremendous growth in the demand for secured systems, automatic personal verification using
biometrics has drawn increasing attention in today’s world.

• In this paper, Finger-vein biometrics has been extensively investigated for personal verification to
proposes a deep learning model based on a combination of known state of the art handcrafted finger-
vein image segmentation techniques, we automatically identify two regions: a clear region with high
separability between finger-vein patterns and background, and an ambiguous region with low
separability between them.
 EXISTING SYSTEM

• In this approaches extract vein patterns by assuming that they generate distributions such
as valleys and line segments. They can be broadly classified into two categories:

1) Finger-vein extraction approaches based on detecting valley features

2) Finger-vein extraction approaches based on detecting line-like features

• This system which operates in two stages that is minutia extraction and minutia
matching.
 DRAWBACKS
• It completely depend on domain knowledge.

• Lack of robustness to extract finger-vein features from raw images.

• Verification accuracy is low.

• Imbalanced and poor local illumination problem.

• Privacy and security problem occur.

• Not always effective to detect the finger vein patterns.

• Not easily observed in visible light.

 LITERATURE REVIEW
Title Aim Merits Demerits
Face recognition In this paper, the detection and It provides the It do not necessarily
using Eigen faces identification of human faces is presented, ability to learn to correspond to isolated
M. A. Turk and A. P. and a working, near-real-time face recognize new features such as eyes,
Pentland recognition system which tracks a subject's faces in an ears, and noses.
Year-2009 head and then recognizes the person by unsupervised
comparing characteristics of the face to manner.
those of known individuals is described.
On-line fingerprint To describes the design and implementation It is one of the This system doesn’t meets
verification of an online fingerprint verification system most reliable the response time
A.Jain, L. Hong, and which operates in two stages: minutia personal requirements of online
R. Bolle extraction and minutia matching. identification verification with high
Year-1997 methods. accuracy.
How iris recognition To explains the iris recognition algorithms The minutia It produce combinatorial
works and presents results of 9.1 million extraction complexity of the phase
J. Daugman comparisons among eye images from trials algorithm is much information across
Year-2009 in Britain, the USA, Japan, and Korea. faster and more different persons spans.
reliable.
 LITERATURE REVIEW
Title Aim Merits Demerits
Human identification using To develop and investigate two new This system This system
finger images score-level combinations, i.e., holistic is low-cost, near simultaneously
A.Kumar and Y. Zhou and nonlinear fusion, and comparatively infrared, acquires the low-
Year-2012 evaluate them with more popular score- contactless resolution fingerprint
level fusion approaches to ascertain their imaging. images.
effectiveness.
Personal authentication using To presents a new approach to It is fully The user
hand vein triangulation and authenticate individuals using automated and identification is not
knuckle shape triangulation of hand vein images and suggest more user reliable and not a
A. Kumar and K. V. Prathyusha simultaneous extraction of knuckle friendly alternative effective access
Year-2009 shape information. for user control.
identification.
Human identification using To evaluate the performance It improve the This system
palm-vein images improvement in verification and performance of has lower security
Y. Zhou and A. Kumar recognition scenarios and analyze the palm-vein-based and privacy for the
Year-2011 influence of enrollment size on the identification user.
performance. systems.
 PROPOSED SYSTEM
• In this paper, we propose a deep learning model for finger-vein verification.

• In this, a Convolutional Neural Network (CNN) is trained on the resulting dataset to

predict the probability of each pixel of being foreground given a patch centered on it.

• The CNN learns what a finger-vein pattern is by learning the difference between vein
patterns and background ones.

• We propose another new and original contribution by developing and investigating a Fully
Convolutional Network (FCN) to recover missing finger-vein patterns in the segmented
image.
ARCHITECTURE
 EXPLANATION

• We propose an automatic scheme to label pixels in vein regions and background regions, given
very limited human knowledge.

• We employ several baselines approaches to extract (segment) the vein network from an image
and use their combined output automatically to assign a label for each pixel.

• Such a scheme avoids the heavy manual labeling and may also reduce label errors, especially
for ambiguous pixels.

• As a finger-vein consists of clear regions and ambiguous regions, several baselines are
employed to automatically label pixels as vein or background in the image clear regions, thus
avoiding the tedious and prone-to-error manual labeling.
 CONT….
• A CNN-based scheme is employed to automatically learn features from raw pixels for finger-vein

verification.

• First, a dataset is constructed based on patches centered on the labeled pixels, and we take the

patches as input for CNN training.

• Secondly, in the test phase, the patch of each pixel is input into CNN the output of which is taken

as the probability of the pixel to belong to a vein pattern.

• Then, the vein patterns are segmented using a probability threshold of 0.5. Compared to existing

approaches, our CNN automatically learns robust attributes for finger-vein representation.
 CONT….
• This paper investigates a new approach for recovering vein patterns in the extracted finger-vein

image.

• As finger-vein patterns may be missing by corruption during the imaging stage and the

inaccurate estimation of parameters during the preprocessing stage (i.e. alignment and feature

extraction), we develop a robust finger-vein feature recovering scheme based on a Fully

Convolutional Network (FCN).

• In this context, we perform a rigorous experimental analysis that shows that our scheme does

succeed in recovering missing patterns which further improves the verification performance.
 CONCLUSION

• This paper proposed a deep learning based segmentation model for finger-vein
verification.

• Experimental results on large public datasets show that the proposed model is able to
extract the vein patterns from raw images in a robust way, which leads to a significant
improvement in finger-vein verification accuracy.

• This approach achieves promising performance for finger-vein verification.

 REFERENCES
• M. A. Turk and A. P. Pentland, “Face recognition using Eigen faces,” CVPR, pp. 586–591, 1991.

• A. Jain, L. Hong, and R. Bolle, “On-line fingerprint verification,” IEEE Transactions on Pattern Analysis
and Machine Intelligence, vol. 19, no. 4, pp. 302–314, 1997.

• J. Daugman, “How iris recognition works,” IEEE Transactions on Circuits and Systems for Video
Technology, vol. 14, no. 1, pp. 21–30, 2004.

• A. Kumar and Y. Zhou, “Human identification using finger images,” IEEE Transactions on Image
Processing, vol. 21, no. 4, pp. 2228–2244, 2012.

• A. Kumar and K. V. Prathyusha, “Personal authentication using hand vein triangulation and knuckle
shape,” IEEE Transactions on Image Processing, vol. 18, no. 9, pp. 2127–2136, 2009.
THANK YOU……