Fpga implementation of fusion technique for fingerprint application

Proceedings of the 2nd International Conference on Current Trends in Engineering and Management ICCTEM -2014
INTERNATIONAL JOURNAL OF ELECTRONICS AND
17 – 19, July 2014, Mysore, Karnataka, India
COMMUNICATION ENGINEERING & TECHNOLOGY (IJECET)
ISSN 0976 – 6464(Print)
ISSN 0976 – 6472(Online)
Volume 5, Issue 8, August (2014), pp. 171-177
© IAEME: http://www.iaeme.com/IJECET.asp
Journal Impact Factor (2014): 7.2836 (Calculated by GISI)
www.jifactor.com
IJECET
© I A E M E
FPGA IMPLEMENTATION OF FUSION TECHNIQUE FOR FINGERPRINT
APPLICATION
S. Gayathri1, V Sridhar2
1Dept of E & C, SJCE , Mysore, Karnataka, India
2Dept of E&C, PESCE, Mandya, Karnataka, India
171
ABSTRACT
Image Fusion is a process of combining relevant information from a set of images, into a
single image, wherein the resultant fused image will be more informative and complete than any of
the input images. This paper discusses Laplacian Pyramid (LP) based image fusion techniques for
fingerprint application. The technique is implemented in MatLab and evaluation parameters Mean
Square Error (MSE), Peak Signal to Noise Ratio (PSNR) and Matching score are discussed. As well
the same implemented on Virtex-5 FPGA development board using Verilog HDL. LP based
technique provides better results for image fusion than other techniques.
Keywords: Fusion, Algorithm, Parameters, Implementation, Matching.
1. INTRODUCTION
Fingerprints are the oldest and most widely used form of biometric for personal
identification. Despite the widespread use of fingerprints, there is little statistical theory on the
uniqueness of fingerprint minutiae.
A critical step in studying the statistics of fingerprint minutiae is to reliably extract minutiae
from the fingerprint images. However, fingerprint images are rarely of good quality. They may be
degraded and corrupted due to variations in skin and impression conditions. Thus, image
enhancement techniques are employed prior to minutiae extraction to obtain a more reliable
estimation of minutiae locations [1].
Image fusion is a branch of data fusion where data appear in the form of arrays of numbers
representing brightness, colour, temperature, distance, and other scene properties. Such data can be
two-dimensional (still images), three-dimensional (volumetric images or video sequences in the form
of spatio-temporal volumes), or of higher dimensions [2].
Image fusion is the process of combining information from two or more images of a scene
into a single composite image that is more informative and is more suitable for visual perception or

172
computer processing.
The objective of image fusion is to convey uncertainty and minimize redundancy in the
output while maximizing relevant information particular to an application. Given the same set of
input images, different fused images may be created depending on the specific application and
relevant information. There are several benefits in using image fusion: wider spatial and temporal
coverage, decreased uncertainty, improved reliability, and increased robustness of system
performance [3].
PCA & DCT image fusion technique can be used for applications, which does not require
high quality & precision. Whereas DWT based fusion techniques provide us good quality fused
images than PCA & DCT based techniques [4].
Hierarchical multi-scale and multi-resolution image processing techniques, pyramid
decomposition are the basis for the majority of image fusion algorithms. Principal component
analysis (PCA) is a well-known scheme for feature extraction and dimension reduction and is used
for image fusion. Proposed image fusion using hierarchical PCA is better for the fusion of
multimodal images [5].
2. LAPLACIAN PYRAMID FUSION TECHNIQUE
It is a straight forward, intuitive band-pass decomposition which is simple to implement and
computationally efficient. To implement Laplacian pyramid decomposition, one must first define
two elementary scaling operations, usually referred to as shrink and expand. The shrink operation
applies a low-pass filter to the image and down samples it by a factor of two.
The expand operation employs a predefined interpolation method and up samples the image
by a factor of two. An image pyramid consists of a set of low pass or band pass copies of an image,
each copy representing pattern information of a different scale.
At every level of fusion using pyramid transform, the pyramid would be half the size of the
pyramid in the preceding level and the higher levels will concentrate upon the lower spatial
frequencies. The basic idea is to construct the pyramid transform of the fused image from the
pyramid transforms of the source images and then the fused image is obtained by taking inverse
pyramid transform.
Given these two operations, the Laplacian pyramid is obtained via the following two-step
process:
1. Generate the Gaussian pyramid of the image. The Gaussian pyramid is basically a series of copies
{G1,G2,….,GK} of the original image I at different scales. It is obtained by setting G1 = I, and
iteratively applying Gi+1 = shrink (Gi).
2. Generate the Laplacian pyramid of the image. The Laplacian pyramid {L1,L2,……LK} is
obtained by backward-processing the Gaussian pyramid, setting Lk = Gk and iteratively applying
Li = Gi-expand (Gi+1).
The inverse transform, for recovering an image from its Laplacian pyramid, is computed by
setting Gk = Lk, and iteratively computing Gi-1 = Li-1+expand (Gi). The image is then given by
I = G1.
Assuming the low-pass filter in the shrink operation roughly eliminates the higher half of the
image frequencies, the Gaussian pyramid is simply a series of scaled-down versions of the original
image, each Gi+1 representing the lower half of the frequencies of its predecessor Gi.
For the expand operation, we assume a simple interpolation method (such as linear or cubic
interpolation), such that the upscaling process roughly preserves the frequency composition of the
image, and introduces minimal artificial high frequencies.
Consequently, applying Li = Gi-expand (Gi+1) essentially removes the lower half of the
frequency spectrum from Gi, retaining only the higher frequency band. The resulting pyramid
{L1,L2,……LK} is therefore a form of band-pass decomposition of the image, where L1 represents

the lowest part of the spectrum, and each Li+1 represents a higher frequency band than its
predecessor. L1 is often referred to as the approximation level in the pyramid, and the remaining Lis
are the detail levels [1].
The Laplacian pyramid transform is specifically designed for capturing image details over
multiple scales. It is obviously an over-complete transformation, and as opposed to the wavelet
decomposition, for instance, each band-pass level is sampled at precisely its Nyquist frequency
making it less sensitive to noise.
Also, one may easily verify that the Laplacian pyramid transform is invariant under affine
transformations. All these properties make the Laplacian pyramid transform a well-suited
representation for the task at hand.
3. PERFORMANCE EVALUATION PARAMETERS
The Laplacian pyramid fusion technique is evaluated by considering the parameters like
mean square error, peak signal to noise ration and matching Score.
173
3.1. Mean Square Error
Begin with a discussion of the MSE as a signal fidelity measure. The goal of a signal fidelity
measure is to compare two signals by providing a quantitative score that describes the degree of
similarity/ fidelity or, conversely, the level of error/distortion between them. Usually, it is assumed
that one of the signals is a pristine original, while the other is distorted or contaminated by errors.
Suppose that x = {xi|i = 1, 2, · · · , N} and y = {yi|i =1, 2, · · · , N} are two finite-length,
discrete signals (e.g., visual images), where N is the number of signal samples (pixels, if the signals
are images) and xi and yi are the values of the ith samples in x and y, respectively. The MSE
between the signals is
..
…………………….......... (1)
3.2. Peak Signal to Noise Ratio
The PSNR block computes the peak signal-to-noise ratio, in decibels, between two images.
This ratio is often used as a quality measurement between the original and a compressed image.
The Mean Square Error (MSE) and the Peak Signal to Noise Ratio (PSNR) are the two error
metrics used to compare image compression quality.
The MSE represents the cumulative squared error between the compressed and the original
image, whereas PSNR represents a measure of the peak error. Lower the value of MSE, lower will be
the error.
To compute the PSNR, the block first calculates the mean-squared error using the following
equation:
………………………….(2)
In the previous equation, M and N are the number of rows and columns in the input images,
respectively. Then the block computes the PSNR using the following equation

174
…..………..………..… (3)
In the previous equation, R is the maximum fluctuation in the input image data type.
3.3 Matching
In this process two fingerprint images are fused and the matching is performed between fused
image and one of the two fingerprint image and the performance evaluation parameters are tabulated.
The matching score will be indicated in percentage which will be helpful for comparison purpose.
4. RESULTS AND DISCURSION
The LP fusion technique is best among the different fusion technique and that algorithm is
coded using Xilinx ISE Design Suite 13.1 and the corresponding simulator results are depicted in
Figure 1.
Figure 2 shows the RTL schematic of Laplacian pyramid fusion algorithm which is obtained
after executing the corresponding Verilog code in Xilinx ISE design suite 13.1.
Figure1: Simulation Diagram
Figure 2: RTL schematic

The synthesis report of Laplacian pyramid fusion technique which is implemented on Virtex-
5 FPGA development board depicts the device utilization summary is shown in the Table 1.
Table 1: Synthesis
175
Report
The evaluation parameters such as mean square error, peak signal to noise error and the
matching score for 10 sets of images are tabulated in Table 2.
The database consists of fingerprint images which are generated using NFD scanner. Each set
consists of fingerprint images of Thumb and index fingers of a person and the tabulation are made
for 10 different set of images of different persons for evaluation purpose. Minimum the Mean square
error and maximum the peak signal to noise ratio better the performance of the proposed fusion
technique. Mean square error will be minimum and peak signal to noise ratio will be maximum for
good quality of images acquired through the optical scanner. The Laplacian pyramid fusion
techniques perform well for medium quality of images.
Table 2: Evaluation parameters
Samples MSE PSNR(db) Matching score (%)
Set 1 174.18 25.92 78.40
Set 2 134.56 27.96 71.81
Set 3 164.98 26.15 72.10
Set 4 165.95 25.14 75.67
Set 5 157.71 28.03 74.5
Set 6 117.19 26.17 76.33
Set 7 146.75 25.94 73.5
Set 8 174.52 26.40 72.3
Set 9 166.72 26.12 73.9
Set 10 173.02 25.61 76.8

Comparison of different fusion techniques for the same 10 set of images with respect to
matching scores are tabulated as in Table 3. It is clear from the values tabulated as in Table 3 that
Laplacian pyramid fusion technique provides consistency in the Matching (%) score compared with
the other three fusion techniques.
Table 3: Comparison of matching score
Samples DCT Sum LP PCA
Set 1 52.39 88.06 78.40 62.9
Set 2 60.77 63.2 71.81 64.8
Set 3 62.73 66.09 72.10 64.4
Set 4 61.64 59.22 75.67 55.05
Set 5 76.87 61.33 74.5 78.13
Set 6 76.02 62.59 76.33 73.2
Set 7 61.63 41.95 73.5 54.2
Set 8 64.57 42.48 72.3 53.4
Set 9 73.04 58.22 73.9 74.5
Set 10 66.62 56.42 76.8 79.9
Laplacian pyramid fusion technique not only provides better matching score but also retrieves
any loss in the ridges in either of the input images. Figure 3 shows the fingerprint image1 and
fingerprint image 2 and fused image of LP Fusion technique. In both Fingerprint image 1 and
Fingerprint image 2 there is some loss in the ridges. After fusion these losses are eliminated,
providing good fingerprint image.
Image1 Image2 Fused Image
Figure 3: Laplacian Pyramid fusion technique
176
5. CONCLUSION
From the above discussion it can be concluded that the Laplacian pyramid fusion technique
for fingerprint application is carried out using MatLab for determining the evaluation parameters
such as mean square error, peak signal to noise ration and the matching score. The obtained

evaluation parameters are satisfactory. The matching scores obtained from Laplacian pyramid fusion
technique is better than that of DCT, Sum rule and PCA fusion techniques.
The proposed fusion technique is coded using Verilog HDL and implemented on Vertex-5
FPGA development board. Further the Mean square error and peak signal to noise ratio can be used
to compare the different fusion techniques such as sum rule, PCA and DCT.
177
REFERENCES
Journal papers
[I] V.P.S. Naidu and J.R. Rao, Pixel-level Image Fusion using wavelets and Principal
Component Analysis , Defense Science Journal, VoI. 58, No. 3, May 2008, pp 338-352.
[2] VPS Naidu, Discrete Cosine Transform based Image Fusion techniques, Journal of
Communication, Navigation and Signal Processing, January 2012, VoI. I, No. I, pp 35-45.
[3] Nirosha Joshitha J, R. Medona Selin, lmage Fusion using PCA in Multifeature Based
Palmprint Recognition, lnternational Journal of Soft Computing and Engineering (IJSCE)
May 20 I 2, ISSN: 2231-2307, Volume-2, Issue-2.
Proceedings papers
[4] Mr. Rajenda Pandit Desale and Prof Sarita V Verma, Study and Analysis of PCA, DCT and
DWT based Image Fusion Techniques, proceedings of International Conference on Signal
Processing, Image Processing and Pattern Recognition (ICSIPRl), 2013, pp 1-4.
[5] Ujwala Patil and Uma Mudengudi, Image fusion using hierarchical PCA , proceedings of the
2011 International Conference on Image Information Processing (ICIIP 2011), 2011 IEEE,
pp 1-6.

Fpga implementation of fusion technique for fingerprint application

Recommended

More Related Content

What's hot (15)

Viewers also liked (20)

Similar to Fpga implementation of fusion technique for fingerprint application (20)

More from IAEME Publication (20)

Recently uploaded (20)

Fpga implementation of fusion technique for fingerprint application