Engineering and Scientific International Journal (ESIJ) ISSN 2394-7187(Online)

Volume 5, Issue 2, April – June 2018 ISSN 2394 - 7179 (Print)

Fingerprint Enhancement using Gabor Filter Algorithm

Selvi.M*1,Sridhar.S*2
1
Student, Master of Computer Applications, S.A. Engineering college, Chennai-77.
[email protected]
2
Asso Prof., Department of Computer Applications, S.A. Engineering college, Chennai-77.

Abstract — The most important thing in fingerprint surfaces of a finger trip. Finally, the ridge orientation
enhancement method is the segmentation of region and the represents the orientation of each pixel in each block [6].
interest. Many automatic systems for fingerprint
enhancement techniques are based on minutiae matching
concepts. Minutiae matching concept pointed to the core
point, ridge ending and bifurcations of the ridges that
identifies the fingerprint image pattern. The proposed
enhancement method is an essential pre-processing of the
fingerprint enhancement applications. Enhanced fingerprint
is a common and complex in fingerprint identification
method as it gives the base for the performance of
computing system. The ridge structure in a fingerprint can
be viewed as an orientation patterns, having a spatial
frequency and orientation. The important step in fingerprint Fig.1: Ridge ending, core point and ridge bifurcation
image recognition is the segmentation of the region of
interest. Results define the robustness of the proposed 2. Existing System
system.
Fingerprint enhancement has to represent the
Keywords — Fingerprint Reconstruction; Gabor Filter performance of database size which should be large enough,
Algorithm; Filtering; Fingerprint Matching; Core Point. but the number of large fingerprint images and the duration
of the image tests are very long. Many papers discuss the
1. Introduction image verification techniques which are adaptive, curved
and two dimensional (2D) methods to match the different
Fingerprint image enhancement is one of the important images by using Gaussian methods. But these methods can
biometric security which have drawn a substantial amount take more time to identify the input images. Calculating the
of attention. The proposed method is more reliable standard deviation of the input image and using a value as
compare to other biometric method. A fingerprint image argument of the bi-dimensional Gaussian width did not
can make the impression of the friction ridges, from the present good results. The proposed method uses the
surface of a finger tip. Orientations, frequency and angles segmentation and Gabor filters algorithms, whose spatial
are calculated from the images and it can be defined by frequencies correspond to the average inter-ridges spacing
ridges and valleys [1]. in fingerprints, which is used to capture the ridge strength
The fingerprint image enhanced method using at equally spaced orientations. A circular tessellation of
segmentation divides the image into small blocks and filtered image is used to construct the ridge feature
calculate the ridge orientation of each block [2] [3]. This mapping.
ridge orientation represents the orientation of each pixel in
each block. This block captures the local information and it 2.1 Disadvantages
can be used to remove the unwanted noise efficiently. The
structure of fingerprint images is enhanced by Gabor filter Reconnects broken ridges (dryness or scars and wetness
using arguments according to the region of the interest. The image patterns) are difficult to be identified and enhanced.
fingerprint matching methods are contained ease of Sometimes fingerprint get more noisy and not accepted
acquisition and high level image matching accuracy rates because bad quality of fingerprint images are very difficult
[4] [5]. to identify and match from the database. Cannot pass entire
All technologies of fingerprint image matching, harmonic excepted signal with particular frequency. A
identification, verification, minutiae extraction and spectral tolerance box is located around borders of the fingerprint
features require different treatments and methods. A region of interest to discard minutiaes captured in
fingerprint is an impression of the friction ridges from the orientation area, as they are not stable for recognition [7].

12
Engineering and Scientific International Journal (ESIJ) ISSN 2394-7187(Online)
Volume 5, Issue 2, April – June 2018 ISSN 2394 - 7179 (Print)

3. Proposed System

The segmentation method is used to split the finger print

area from the background. This is useful to avoid
subsequent extraction of fingerprint features in the
background, which is a noisy area. The performance of
Gabor filter algorithm has been evaluated in terms of
decision error trade-off curves. Pre-processing of finger
print images is used to remove noise to give perfect quality.
They may be degraded and corrupted with obstacles due to The most local ridge structures of fingerprints come
many factors including variations in skin, impression and with well-defined local frequency and orientation. By using
their conditions. Most Important step in segmentation part a too big f, it can create noises in the filtered image; if f is
is to exclude non-ridge and unrecoverable regions. very small, it can interface to the ridges. Being an
argument, K variable in the automatic process permits
3.1 Advantages verification of fingerprint image from a unique finger with
an accurate matching process, and the image from different
 Conversion of bad quality images into good quality fingers.
Filtering the fingerprint using Gabor filters can enhance
images.
the ridges and valleys structures. The developer effectively
 Time variance reduction. removes undesired noise, preserve perfect ridge and valley
 Identify and match exact finger print image. structures. This isolates feature information contained in a
specific image orientation and frequencies.
4. Methodology
5. Implementation
The image frequency is due to inter ridge-spacing
present in a fingerprint and the orientation is due to the Gabor filter-based techniques can be divided by three
flow pattern exhibited by ridges. Fingerprint patterns are major methods, namely: (i) Overlapping Blocks, (ii) Ridge
characterized by ridges and valleys. The ridge and valley Frequency Computation and (iii) Heuristic Constraints.
structures are well defined and not always visible in a (i) Overlapping blocks
scanned fingerprint image, even though information about Overlapping concepts uses the block sized by windowing
direction of ridges and distance among ridges are preserved. concepts with an overlapping of W/2 pixels.
(ii) Ridge frequency computation
The ridge frequency computation method is used to
estimate the local frequency field of the fingerprints and to
use estimated values.
(iii) Heuristic constraints
The normalization technique is used for minutia
extraction stage to remove the noise from the image; the
Morphological thinning operation is finally bid out with
high efficiency and pretty good thinning quality. Masking
image can select ridge bifurcation points and core point as
minutia and minutia has identified with and without using
the image mask.
Finally, intersection method has been used on minutia
points which gives the effective minutia points in quality
areas of images in the perfect manner. This has been done
in order to find valid minutia points in order to get
accuracy in matching procedures. Fingerprint enhancement
techniques are more efficient and more reliable to produce
the excellent results.

Fig.2: Fingerprint enhancement algorithm 5.1 Fingerprint Matching System

Gabor filter algorithm using general format in the spatial In finger print matching, the complete sequence
domain is follows, for fingerprint image enhancement is follows, (i)

13
Engineering and Scientific International Journal (ESIJ) ISSN 2394-7187(Online)
Volume 5, Issue 2, April – June 2018 ISSN 2394 - 7179 (Print)

normalization, (ii) region of interest extraction, (iii) ridge makes it possible to enhance such images and improves the
extraction and (iv) masking. decision step.

Fig.4: Correct and incorrect scores for test fingerprint

Fig. 3: Results of fingerprint enhancement
5. Conclusion
5.2 Image Enhancement
Enhanced Gabor filter-based segmentation method
Fingerprint image enhancement techniques are used to includes block overlapping concepts and ridge frequency
provide high quality images. Different kinds of images are computation techniques. This method solves the problem
sequenced; after it can be connected. The component of loss in valid regions within the foreground area. The
analysis is applied to carry out the reconstruction process in overlap-based approach takes into account of segmentation.
which unnecessary ridge segments and best segments are The matching implements the order to check the accuracy
kept and rejoining to get the valid results from the minutia and performance of the enhancement algorithm.
points. After this, image template has been stored in a Enhancement of image is a series like techniques for
folder which is used in next session for fingerprint fingerprint image enhancement to addresses the image
matching process. The ridge orientations are located in the extraction of minutiae point.
image and reliable orientations are selected as a tool for Fingerprint image enhancement is a complex and more
Gabor filter algorithm. Ridge frequency and orientation has challenging problem and there are still a lot of works
done by dividing the image into small blocks. Median needed. Enhanced fingerprint image has received the
values have been used in each and every block which strong attention from researchers in biometrics, pattern
enhances the performance and effectiveness of the recognition, image processing, verification and cognitive
algorithm. Gabor filter algorithm usually catches both psychology groups.
orientation and frequency data from a fingerprint image.
Each sample image test is matched opposite to the 6. Future Enhancement
remaining samples of that finger to compute the False
Rejection Rate (FRR).If the matching g against h is The following areas of future work are also being
performed, the symmetric one is not executed to avoid studied and implemented:
correlation. All the result scores from such matches are  New matching techniques for matching the ridge
composed into a series of correct and approximate score feature maps of two images.
which is not 100%.  Constructing the ridge feature maps using adaptive
techniques for optimal selection of Gabor filters.
6. Experimental Results It can capture the local information and can use them to
efficiently remove the undesired noise from images. The
The database of fingerprint images contains more than algorithm will be modified to get more minutia points as in
180 images. There are 8 different kinds of impressions per this algorithm we used only terminations and bifurcations.
finger. The result of connected component labeling and
reconstruction process is shown bellow.
Reference
The fingerprint ridge timing process is necessary to
catch the minutia points in the image. The output performs [1] A. K. Jain, A. Ross, and S. Prabhakar, An introduction to biometric
the minutiae based approach over wider range sample tests recognition,‖ IEEE Transactions on Circuits and Systems for Video
that indicate the use of appropriate filter parameters. This Technology, vol. 14,pp. 4–20, Jan. 2004.

14
Engineering and Scientific International Journal (ESIJ) ISSN 2394-7187(Online)
Volume 5, Issue 2, April – June 2018 ISSN 2394 - 7179 (Print)

[2] D. Maltoni, D. Maio, A. Jain, and S. Prabhakar, Hand book of [7] D. Simon-Zorita, J. Ortega-Garcia, J.Fierrez-Aguilar, and J.
Fingerprint Recognition. Springer, New York,2003. Gonzalez-Rodriguez, ―Image quality and position variability
[3] A. K. Jain, L. Hong, S. Pankanti, and R. Bolle, An identity assessment in minutiae-based fingerprint verification,‖ IEE
authentication system using fingerprints, Proceedings of IEEE, Vol. Proceedings - Vis. Image SignalProcess., Vol. 150, pp. 402–408,
85, pp. 1365–1388, Sept. 1997. Dec. 2003.
[4] Dario Maio and Davide Maltoni, ―Direct gray-scale minutiae [8] Lin Lin Shen, Alex Kot, and Wai Mun Koo, Quality measures of
detection in fingerprints,‖ IEEE Trans. On Pattern Analysis and fingerprint images, Proceedings of 3rd Audio and Video-Based
Machine Intelligence, vol. 19, pp.27–40, Jan. 1997. Person Authentication, AVBPA 2001,pp. 266–271.
[5] A. Bazen and S. Gerez, ―Segmentation of fingerprint images,‖ in [9] A. Martin, G. Doddington, T. Kamm, M. Ordowski and M.
Proc. Workshop on Circuits Systems and Signal Processing, Przybocki, the DET curve in assessment of decision task
Proceedings of RISC 2001, pp. 276–280. performance, Proceedings of ESCA European Conference on
[6] B. Mehtre, Fingerprint image analysis for automatic identification, Speech Communication and Technology, EuroSpeech1997, pp.
Machine Vision and Applications, Vol.6, pp. 124–139, 1993. 1895–1898.

15