International Journal of Engineering Research in Computer Science and

Engineering (IJERCSE) Vol 3, Issue 3, March 2016

Fingerprint Compression Based On Representation

Techniques
[1]
N Sai Ramya, [2] M Pallavi, [3] Prasad B
[1]
II/IV, [2][3]Associate Professor
[1][2][3]
Department of CSE, Marri Laxman Reddy Institute of Technology and Management (MLRITM)
Hyderabad
[1]
[email protected] [2], [email protected] [3][email protected]

Abstract The primary aim of this project is to implement techniques for fingerprint image enhancement and minutiae extraction.
Recognition of people by means of their biometric characteristics very popular among the society. But a fingerprint image consists
of enormous amount of data. For a given whole fingerprint, divide it into small blocks called patches. Obtaining an over complete
dictionary from a set of fingerprint patches allows us to represent them as a sparse linear combination of dictionary atoms. In the
algorithm, we first construct a dictionary for predefined fingerprint image patches. Large volume of fingerprint is collected and
stored everyday in a wide range of applications. The experiments demonstrate that this is efficient compared with several
competing compression techniques especially at high compression ratios. There are many image compression techniques available.
Fingerprint images are rarely of perfect quality. There are many image compression techniques available. JPEG, JPEG 2000,
Wavelet Scalar Quantization (WSQ) are the existing image compression techniques. The JPEG, JPEG 2000 methods are for
general image compression. Fingerprint identification methods are widely used by police agencies and customhouse to identify
criminals or transit passengers since the late nineteenth century. ISO standardized the characteristics of the fingerprint in 2004.
After the image enhancement construct a base matrix whose columns represent features of the fingerprint images, referring the
matrix dictionary whose columns are called atoms.

Keywords: Fingerprint images, Wavelet Scalar Quantization, Compression techniques, Matrix dictionary, Atoms.

I. INTRODUCTION are often employed in the output coefficients of lossy

compression.
Recognition of persons by means of biometric
characteristics is an important technology in the society, Lossy compression technologies usually transform
because biometric identifiers can’t be shared and they the an image into another domain, quantize and encode its
individual’s bodily identity. Among many biometric coefficients. During the last three decades, transform based
recognition technologies, fingerprint recognition is very image compression technologies have been extensively
popular for personal identification due to the uniqueness, researched and some standards have appeared. Two most
universality, collectability and invariance .Large volumes of common options of transformation are the Discrete Cosine
fingerprint are collected and stored every day in a wide Transform (DCT) and the Discrete Wavelet Transform
range of applications, including forensics and access (DWT).
control. In 1995, the size of the FBI fingerprint card archive
contained over 200 million items and archive size was The DCT-based encoder can be thought of as
increasing at the rate of 30000 to 50 000 new cards perday. compression of a stream of 8 × 8 small block of images.
Large volume of data consumes the amount of memory. This transform has been adopted in JPEG. The JPEG
Fingerprint image compression is a key technique to solve compression scheme has many advantages such as
the problem. Generally, compression technologies can be simplicity, universality and availability. However, it has a
classed into lossless and lossy. bad performance at low bit-rate mainly because of the
underlying block-based DCT scheme. For this reason, as
Lossless compression allows the exact original early as 1995, the JPEG-committee began to develop a new
images to be reconstructed from the compressed data. wavelet-based compression standard for still images,
Lossless compression technologies are used in cases where namely JPEG 2000. In this paper, a novel approach based
it is important that the original and the decompressed data on sparse representation is given. The proposed method has
are identical. Avoiding distortion limits their compression the ability by updating the dictionary. The specific process
efficiency. When used in image compression where slight is as follows: construct a base matrix whose columns
distortion is acceptable, lossless compression technologies represent features of the fingerprint images, referring the
matrix dictionary whose columns are called atoms; for a

All Rights Reserved © 2016 IJERCSE 443

 International Journal of Engineering Research in Computer Science and
Engineering (IJERCSE) Vol 3, Issue 3, March 2016

given whole fingerprint, divide it into small blocks called personnel in providing input, Methods for preparing input
patches whose number of pixels are equal to the dimension validations and steps to follow when error occur.
of the atoms; use the method of sparse representation to The input form of an information system should accomplish
obtain the coefficients; then, quantize the coefficients; last, one or more of the following objectives.
encode the coefficients and other related information using
lossless coding methods. In most instances, the evaluation of Input Design is the process of converting a user-
compression performance of the algorithms is restricted to oriented description of the input into a computer-based
Peak Signal to Noise Ratio (PSNR) computation. The system. This design is important to avoid errors in the data
effects on actual fingerprint matching or recognition are not input process and show the correct direction to the
investigated. In this paper, we will take it into consideration. management for getting correct information from the
In most Automatic Fingerprint identification System computerized system.
(AFIS), the main feature used to match two fingerprint It is achieved by creating user-friendly screens for
images are minutiae (ridges endings and the data entry to handle large volume of data. The goal of
bifurcations).Therefore, the difference of the minutiae designing input is to make data entry easier and to be free
between pre- and post-compression. from errors. The data entry screen is designed in such a way
that all the data manipulates can be performed. It also
provides record viewing facilities.
II. SYSTEM ANALYSIS
When the data is entered it will check for its
validity. Data can be entered with the help of screens.
Motivation: Appropriate messages are provided as when needed so that
Accurate automatic personal identification is the user will not be in maize of instant. Thus the objective of
critical in wide range of application domains such as input design is to create an input layout that is easy to
national ID cards, electronic commerce and automatic follow.
banking. Biometrics, which refers to automatic
identification of a person based on his or her personal
Output Design Objectives:
physiological or behavioral characteristics, is inherently A quality output is one, which meets the
more reliable and more capable in differentiating between a requirements of the end user and presents the information
reliable person and a fraudulent impostor than traditional clearly. In any system results of processing are
methods such as PIN and passwords. Automatic fingerprint communicated to the users and to other system through
identification is one of the most reliable biometric outputs. In output design it is determined how the
technology among the different major biometric information is to be displaced for immediate need and also
technologies which are either currently available or under the hard copy output. It is the most important and direct
investigation. The objective of our project is to implement source information to the user. Efficient and intelligent
the image enhancement and minutiae extraction algorithm output design improves the system’s relationship to help
which is capable of doing the matching between different user decision-making.
digitized fingerprints of standard image file formats namely; Designing computer output should proceed in an
BMP, JPEG with high level of accuracy and confidence.
organized, well thought out manner; the right output must be
Input Design Objectives: developed while ensuring that each output element is
The input design is the link between the designed so that people will find the system can use easily
information system and the user. It comprises the and effectively. When analysis design computer output, they
developing specification and procedures for data preparation should Identify the specific output that is needed to meet the
and those steps are necessary to put transaction data in to a requirements. Select methods for presenting information.
usable form for processing can be achieved by inspecting
Create document, report, or other formats that contain
the computer to read data from a written or printed
document or it can occur by having people keying the data information produced by the system.
directly into the system. The design of input focuses on
controlling the amount of input required, controlling the The output form of an information system should
errors, avoiding delay, avoiding extra steps and keeping the accomplish one or more of the following objectives.Convey
process simple. The input is designed in such a way so that information about past activities, current status or
it provides security and ease of use with retaining the projections of the Future, Signal important events,
privacy. Input Design considered the following things: What opportunities, problems, or warnings, Trigger an action and
data should be given as input?, How the data should be Confirm an action.
arranged or coded?, The dialog to guide the operating
Existing System:

All Rights Reserved © 2016 IJERCSE 444

 International Journal of Engineering Research in Computer Science and
Engineering (IJERCSE) Vol 3, Issue 3, March 2016

In existing finger print compression algorithms and

to test the robustness of our algorithm to extract minutiae. A. Normalization
Since existing fingerprint quality assessment algorithms are An input fingerprint image is normalized by cropping a
designed to examine if an image contains sufficient rectangular region of the fingerprint, which is located at the
information (say, minutiae) for matching, they have limited center of the fingerprint and aligned along the longitudinal
capability in determining if an image is a natural fingerprint direction of the finger, using the NISTBiometric Image
or an altered fingerprint. Obliterated fingerprints can evade Software (NBIS). This step ensures that the features
fingerprint quality control software, depending on the area extracted in the subsequent steps are invariant with respect
of the damage. If the affected finger area is small, the to translation and rotation of finger.
existing fingerprint quality assessment software may fail to
detect it as a fingerprint. B. Orientation Field Estimation
The orientation field of the fingerprint is computed
Proposed System: using the gradient-based method. The initial orientation field
In proposed system, construct a base matrix whose
is smoothed averaging filter, followed by averaging the
columns represent features of the fingerprint images,
orientations in pixel blocks. A foreground mask is obtained
referring the matrix dictionary whose columns are called
by measuring the dynamic range of gray values of the
atoms; for a given whole fingerprint, divide it into small
fingerprint image in local blocks and morphological process
blocks called patches whose number of pixels are equal to
for filling holes and removing isolated blocks is performed.
the dimension of the atoms; use the method of sparse
representation to obtain the coefficients; then, quantize the
C. Orientation Field Approximation
coefficients; last, encode the coefficients and other related
information using lossless coding methods. In most The orientation field is approximated by a
instances, the evaluation of compression performance of the polynomial model to obtain.
algorithms is restricted to Peak Signal to Noise Ratio
(PSNR) computation. The effects on actual fingerprint D. Feature Extraction
matching or recognition are not investigated. In this paper, The error map is computed as the absolute
we will take it into consideration. In most Automatic difference between and used to construct the feature vector.
Fingerprint identification System (AFIS), the main feature
used to match two fingerprint images are minutiae (ridges 2. Analysis of Minutiae Distribution
endings and bifurcations). Therefore, the difference of the
minutiae betweenpre- and post-compression is considered in In this module, a minutia in the fingerprint
the paper. . indicates ridge characteristics such as ridge ending or ridge
bifurcation. Almost all fingerprint recognition systems use
minutiae for matching. In addition to the abnormality
observed in orientation field, we also noted that minutiae
III. SYSTEM DESIGN distribution of altered fingerprints often differs from that of
natural fingerprints. Based on the minutiae extracted from a
Implementation is the stage of the project when the fingerprint by the open source minutiae extractor in NBIS, a
theoretical design is turned out into a working system. Thus minutiae density map is constructed by using the Parzen
it can be considered to be the most critical stage in window method with uniform kernel function.
achieving a successful new system and in giving the user,
confidence that the new system will work and be effective. 3. Wavelet Scalar Quantization (WSQ) Compression
The implementation stage involves careful planning, In this module, it is used for gray-scale fingerprint
investigation of the existing system and it’s constraints on images. It is based on wavelet theory and has become a
implementation, designing of methods to achieve standard for the exchange and storage of fingerprint images.
changeover and evaluation of changeover methods. This compression method is preferred over standard
compression algorithms like JPEG because at the same
The main modules involved in the system are compression ratios WSQ doesn't present the "blocking
Detection of Fingerprints (Normalization, Orientation field artifacts" and loss of fine-scale features that are not
estimation, Orientation field approximation and Feature acceptable for identification. After processing the
extraction), Analysis of Minutiae Distribution, Wavelet appropriate files for the encoder and decoder, a certification
Scalar Quantization (WSQ) Compression and Compression request containing the test report, test results, and all
Based on Sparse Representation generated compressed and reconstructed files (cmp000xx
with extensions '.wsq' and '.pgm') are forwarded to the FBI
1. Detection of Fingerprints for review and evaluation.

All Rights Reserved © 2016 IJERCSE 445

 International Journal of Engineering Research in Computer Science and
Engineering (IJERCSE) Vol 3, Issue 3, March 2016

4. Compression Based on Sparse Representation

In this module, for a given fingerprint, slice into small
patches. For each patch its mean is calculated and subtracted
from the patch. For each patch solve the minimization
problem by those coefficients whose absolute value are less
than a given threshold are treated as zero. Record the
remaining coefficients and their locations. Encode the atom
number of each patch.

Activity Diagram: Fig 2: Registration page

Fig 3: New registration form

Fig 1: Activity Diagram

Activity diagram describes the flow of control in a

system. So it consists of activities and links. The flow can Fig 4 : Message send to server
be sequential, concurrent or branched. Activities are nothing
but the functions of a system. Numbers of activity diagrams
are prepared to capture the entire flow in a system. Activity
diagrams are used to visualize the flow of controls in a
system. This is prepared to have an idea of how the system
will work when executed.

IV. SYSTEM IMPLEMENTATION

Fig 5: Login Form

All Rights Reserved © 2016 IJERCSE 446

 International Journal of Engineering Research in Computer Science and
Engineering (IJERCSE) Vol 3, Issue 3, March 2016

developing compression algorithms for fingerprints resides

in the need for preserving the minutiae which are used in the
identification. The experiments show that our algorithm can
hold most of the minutiae robustly during the compression
and reconstruction. There are many intriguing questions that
future work should consider. First, the features and the
methods for constructing dictionaries should be thought
over. Secondly, the training samples should include
fingerprints with different quality (“good”, “bad”, “ugly”).
Thirdly, the optimization algorithms for solving the sparse
representation need to be investigated. Fourthly, optimize
Fig 6 : Message From Client the code to reduce complexity of our proposed method.

REFERENCES

[1] D. Maltoni, D. Miao, A. K. Jain, and S. Prabhakar,

Handbook of Fingerprint Recognition, 2nd ed. London,
U.K.: Springer-Verlag, 2009.

[2] N. Ahmed, T. Natarajan, and K. R. Rao, “Discrete

cosine transform,” IEEE Trans. Comput., vol. C-23, no. 1,
pp. 90–93, Jan. 1974.

[3] C. S. Burrus, R. A. Gopinath, and H. Guo, Introduction

to Wavelets and Wavelet Transforms: A Primer. Upper
Saddle River, NJ, USA: Prentice-Hall, 1998.
Fig 7 : FVC Experiment form
[4] W. Pennebaker and J. Mitchell, JPEG—Still Image
Compression Standard.New York, NY, USA: Van Nostrand
Reinhold, 1993.

[5] M. W. Marcellin, M. J. Gormish, A. Bilgin, and M. P.

Boliek,“An overview of JPEG-2000,” in Proc. IEEE Data
Compress. Conf.,Mar. 2000, pp. 523–541.

[6] A. Skodras, C. Christopoulos, and T. Ebrahimi, “The

JPEG 2000 stillimage compression standard,” IEEE Signal
Process. Mag., vol. 11, no. 5,pp. 36–58, Sep. 2001.
Fig 8 : Remote server
[7] T. Hopper, C. Brislawn, and J. Bradley, “WSQ gray-
V. CONCLUSION scale fingerprintimage compression specification,” Federal
Bureau of Investigation,Criminal Justice Information
A new compression algorithm adapted to Services, Washington, DC, USA, Tech.Rep. IAFIS-IC-
fingerprint images is introduced. Despite the simplicity of 0110-V2, Feb. 1993.
our proposed algorithms, they compare favorably with
existing more sophisticated algorithms, especially at high [8] C. M. Brislawn, J. N. Bradley, R. J. Onyshczak, and T.
compression ratios. Due to the block-by-block processing Hopper, “FBI compression standard for digitized fingerprint
mechanism, however, the algorithm has higher images,” Proc. SPIE,vol. 2847, pp. 344–355, Aug. 1996.
complexities.The experiments show that the block effect of
our algoritm is less serious than that of JPEG. We consider [9] A. Said and W. A. Pearlman, “A new, fast, and efficient
the effect ofthree different dictionaries on fingerprint image codecbased on set partitioning in hierarchical trees,”
compression. The experiments reflect that the dictionary IEEE Trans. CircuitsSyst. Video Technol., vol. 6, no. 3, pp.
obtained by the K-SVD algorithm works best. Moreover, the 243–250, Jun. 1996.
larger the number of the training set is, the better the
compression result is.One of the main difficulties in

All Rights Reserved © 2016 IJERCSE 447

 International Journal of Engineering Research in Computer Science and
Engineering (IJERCSE) Vol 3, Issue 3, March 2016

[10] R. Sudhakar, R. Karthiga, and S. Jayaraman,

“Fingerprint compressionusing contourlet transform with
modified SPIHT algorithm,” IJECEIranian J. Electr.
Comput. Eng., vol. 5, no. 1, pp. 3–10, 2005.

[11] S. G. Mallat and Z. Zhang, “Matching pursuits with

timefrequencydictionaries,” IEEE Trans. Signal Process.,
vol. 41, no. 12,pp. 3397–3415, Dec. 1993.

[12] S. S. Chen, D. Donoho, and M. Saunders, “Atomic

decomposition by basis pursuit,” SIAM Rev., vol. 43, no. 1,
pp. 129–159, 2001.

[13] J. Wright, A. Y. Yang, A. Ganesh, S. S. Sastry, and Y.

Ma, “Robust facerecognition via sparse representation,”
IEEE Trans. Pattern Anal. Mach. Intell., vol. 31, no. 2, pp.
210–227, Feb. 2009.

[14] M. Elad and M. Aharon, “Image denoising via sparse

and redundant representation over learned dictionaries,”
IEEE Trans. Image Process., vol. 15, no. 12, pp. 3736–
3745, Dec. 2006.

[15] S. Agarwal and D. Roth, “Learning a sparse

representation for object detection,” in Proc. Eur. Conf.
Comput. Vis., 2002, pp. 113–127.

All Rights Reserved © 2016 IJERCSE 448