TM SECURITY USING FACE AND FINGERPRINT RECOGNITION

International Journal of Research in Engineering, Technology and Science, Volume

VII, Special Issue, Feb 2017

www.ijrets.com, [email protected], ISSN 2454-1915

Chaitali Bhosale, Pooja Dere, Chaitali Jadhav 1

TM SECURITY USING FACE AND FINGERPRINT RECOGNITION

Chaitali Bhosale, Pooja Dere, Chaitali Jadhav

Department of Electronics And Telecommunications

Bharati Vidyapeeth’s College Of Engineering For Women, Katraj Dhankawadi,

Pune-411043 Savitribai Phule Pune University

[email protected], [email protected], [email protected]

ABSTRACT

Due to rapid development in science and The purpose of this project is provide security technology,

upcoming innovations are being built-up to the conventional ATM model. This paper posited with

strong security. But on the other hand threats are a new concept that enhances the overall experience,

also being posed to destroy this security level. Though usability and convenience of transaction at the

enhancement in automation has made a positive impact ATM. Features like face recognition,

fingerprint overall, but various financial institutions like banks and recognition and one time password

(OTP) are used applications like ATM are still subjected to thefts and for the enhancement of security

of accounts the frauds. The existing ATM model uses a card and PIN privacy users. Face as key. This

completely which gives rise to increase in attacks in the form of eliminates the chances of fraud due

to theft and stolen cards, or due to statically assigned PINs, duplicity of the ATM cards. Fingerprint

also helps duplicity of cards and various other threats. To machine to identify user because of their

uniqueness overcome, hybrid model consist of conventional and consistency over time. Moreover, the

randomly features along with additional features like face generated OTP frees the user from

remembering recognition, Fingerprint recognition and one time PINs as it itself acts as a PIN. There

is no worry of password (OTP) is used. Database holds information losing ATM card and no need to

carry ATM card in about a user's account details, images of his/her face your pocket. and mobile

number which will improve security to a large extent.

Keywords: PCA, OTP, Face recognition, Curvelet

[1] INTRODUCTION

Face recognition finds its variety of such as homeland transform, Fingerprint recognition. security,

criminal identification, human-computer interaction, privacy security etc. Face recognition has been

attracting intense research efforts due to its importance both as one of the main building blocks of natural

human computer interfaces and as a biometric trait. Face recognition has the advantage of everywhere

and of being universal over other major biometrics, in that everyone has a face and everyone readily

http://www.ijrets.com/

mailto:[email protected]



displays the face. Fingerprint is made up of a series of ridges and furrows on the surface of the finger

has been used extensively for personal identification. Fingerprint technology is most widely accepted

because fingerprints from before the birth and except for cut resulting in permanent scars, remain

unchanged till death and mature biometric method and is the easiest to deploy and for a higher level of

security at your fingertips.

First, user will enter user id and password. After that a live image is captured automatically though a

webcam install on the ATM. After that fingerprint recognition done with curvelet transform. which is

compared with the images stored in the database. If it matches, then the person is authorized. And then

an OTP will be send to the corresponding registered mobile number. This randomly generated code has

to be entered by the user in the text box. If the user correctly enter the OTP, the transaction can proceed.

Therefore, the combination of face recognition, fingerprint recognition and an OTP drastically reduces

the chances of frauds plus frees a user from an extra burden of remembering a complex password.

[2] METHODOLOGY

In ATM security system, firstly user will enter user id and password which is provided by bank. After

that a live image is captured automatically though a webcam install on the ATM, at this stage a user

simply needs to look at camera installed on ATM, the user is authorized When this image matches with

the image stored in database. When a customer create an account he/she needs to provide image, this

can be done by capturing his/her image from webcam in the bank. The accountant capture some images

and stores them to the database which has label as account number associates to each of them.

After verification of that, the persons fingerprint recognition done with curvelet transform. It is done by

finding the Euclidean distance between the two corresponding finger codes, if the Euclidean distance

between two feature vectors is less than minimum value, then the two images come from the same finger

otherwise the two images come from the different fingers. The test finger code is compared with the

entire finger codes in the database. If it matches, an OTP will be send to the corresponding registered

mobile number. The idea to use mobile phones is preferred over email because the people in rural area

have simple phones which can receive a text messages but have no internet connections and e-mail

facilities. Since mobile phones are ubiquitous, we intend to use mobile phones so that everyone can take

the benefit of the new proposed system. Once OTP is received user has to enter the code, User gets three

chances to enter the code.

Initially we store the fingerprint and face of user and that will be verified with the face and fingerprint

that we are giving, when the time of authentication. If both the face and fingerprint as well as OTP are

matched then account will open . As it is based on the fingerprint and face authentication there is no

chance of disclosing of password or pin to the third parties.





[3] BLOCK DIGRAM

Fig 3.1. Block diagram of ATM security system using fingerprint and face recognition system

3.1. BLOCK DIAGRAM DESCRIPTION

3.1.1.Web Camera:

Webcam is used for taking the live image of person face. we are using Web Cam after observing its

specification. it capture the image and sent it to PC through USB data cable. It require +5v , 0.35 amp

(max) voltage and current for its operation.

Fig.3.2 Webcam

Specifications:

1.Camera :1.3 mp..

2.Video Resolution : 640 x 480 sensor resolution.

3.USB certification : USB 2.0 high speed certified.





3.1.2.Fingerprint Sensor(R305):-

Fig 3.3 Fingerprint sensor

Specification:

• Power : DC 3.6 V-6.0 V

• Working current: Typical :100mA Peak:150mA

• Image acquiring time: <0.5 s

• Storage capacity :256

3.1.3. Serial interface:

RS232 is the most commonly used serial port in transmitting the data in communication and interface.

Even though serial port is harder to program than the parallel port, this is the most effective method in

which the data transmission requires less wires that yields to the less cost. The RS232 is the

communication line which authorize the data transmission by only using three wire links. The three links

provides ‘transmit’, ‘receive’ and common ground.

[4] FACE DETECTION AND RECOGNITION

Fig.4.1 Block diagram of face image processing





4.1.1.Pre-processing:-

1. Face Image Reading In this step, the face image is loaded by using the Matlab built-in function

imread. 2. RGB to Gray Image Raw data of face image obtained is in the RGB format. The face image

in the RGB format is then changed into gray scale image so that image processing can be executed on

the image. Matlab builtin function, rgb2gray is used to convert RGB image to gray scale image.

I=rgb2gray( RGB);

4.1.2. Histogram Equalization:

Histogram equalization is a method of improving the global contrast of an image by moving slightly the

intensity distribution on a histogram. This permits areas of lower local contrast to gain a higher contrast

without affecting the global contrast. Histogram equalization fulfill this by effectively spreading out the

most frequent intensity values.

Histogram :It is discrete function h(rk)=nk, where rk is kth grey level in the range of [0,L-1]

and nk is number of pixels having grey level rk.

4.1.3.Feature Extraction:-

For feature Extraction PCA algorithm is used.

Purpose of using PCA algorithm:

• It Demands less storage space for storing data set.

• Reduced dimensions increase the efficiency of process.

• Uses to build eigenfaces, good data is required for component matching.

• Time taken for computation is very less as it considers only essential components from images.

Steps for PCA algorithm:





Let an image Γ(x, y) of face be a two dimensional M by N array of intensity values. In this

statement, we used a set of image by 200 × 149 pixels. An image may also be considered as a vector of

dimension M × N, so that a typical image of size 200 × 149 becomes a vector of dimension 29,800.

Step1: prepare the training faces.

Obtain training faces I1, I2, I3, I4 , . . . . . . IM. The face images must be centered and of the same size.

Step 2: Prepare the data set.

Each image of face Ii in the database is transformed into a vector and placed into a training set S.

𝑆 = {𝛤1, 𝛤2, 𝛤3, … . . 𝛤𝑀} .......(1)

For example M = 34. Each image is converted into a vector of size MN × 1 and placed into the set. For

simplicity, the face images are assumed to be of size N

× N resulting in a point in 𝑁2 dimensional space. A group of images, then, maps to a collection of points

in this huge space.

Step 3: compute the average face vector

The average face vector (Ψ) has to be calculated by using the following formula:

Step 6: Calculate the eigenvectors and eigenvalues of the covariance matrix.

The covariance matrix C in step 5 has dimensionality of N2𝑋 N2 , so one would have N2eigenface and

eigenvalues. For a 256 × 256 image that means that on must compute a 65, 536 × 65, 536 matrix and

calculate 65,536 eigenfaces. Arithmatically, this is not very efficient as most of those eigenfaces are not

useful for our task. In general, PCA is used to describe a large dimensional space with a relative small

set of vectors .

Compute the eigenvectors µ𝑖of AAT

The matrix 𝐴𝐴𝑇 is very large ...> not practical!!!

Step 6.1: consider the matrix

L = ATA (M × M matrix)

𝑀

Step 4: Subtract the average face vector The average face vector Ψ is subtracted from the

original faces 𝛤𝑖 and the result stored in the variable , 𝛷𝑖 = 𝛤𝑖 − ψ ...(3)

Step 5: Calculate the covariance matrix.

We obtain the covariance matrix C in the following manner,

𝜓𝑀𝑛𝛤𝑛 .......(2)





C M nT

= 𝐴𝐴𝑇 ...(4)

Where (N2𝑋 N2 𝑚𝑎𝑡𝑟𝑖𝑥)

A=[𝛷1, 𝛷2,𝛷3,𝛷4 …. . 𝛷𝑀]

(N2𝑋 𝑀 𝑚𝑎𝑡𝑟𝑖𝑥)

Step 6.2: compute eigenvectors 𝑣𝑖 of L = ATA

ATA𝜗𝑖 = µ𝑖ϑi

What is the relationship between µ𝑖 and 𝑣𝑖 ?

ATA𝜗𝑖 = µ𝑖ϑi

AATA𝜗𝑖 = µ𝑖𝐴ϑi

CA𝜗𝑖 = µ𝑖𝐴ϑi [ since ]

Cµ𝑖=µ

𝑖𝐴ϑi where, µ𝑖 = 𝐴ϑi

Thus, C = ATA and L = ATA have the same eigenvalues and their eigenvectors are related as follows:

µ𝑖 = 𝐴ϑi Note 1:

𝐶 = 𝐴𝐴𝑇 can have upto eigenvalues and eigenvectors.

Note 2: L = ATA can have upto M eigenvalues and eigenvectors.





𝑢𝑖 𝑣𝑖 𝜑𝑖 ...(5)

Where, µ𝑖 are the Eigenvectors i.e. Eigenfaces.

Step 7: keep only K eigenvectors (corresponding to the K largest eigenvalues) Eigenfaces with low

eigenvalues can be omitted, as they explain only a small part of

Characteristic features of the faces.

[4] FINGERPRINT DETECTION AND RECOGNITION:

For fingerprint detection, user simply needs to look into the camera installed on ATM. If the user is

recognized, then the OTP is sent to user’s mobile phone.

Three basic patterns of fingerprint ridges are

.

Note 3: The M eigenvalues of 𝐶 = 𝐴𝐴𝑇 (along with their corresponding eigenvectors) correspond

to the M largest eigenvalues of L = ATA (along with their corresponding eigenvectors)

Where 𝑣𝑖 is an eigenvector of L = ATA From this simple proof we can see that 𝐴ϑi is an

eigenvector of 𝐶 = 𝐴𝐴𝑇

The M eigenvectors of L = ATA are used to find the

M eigenvectors µ𝑖 of C that form our eigenface basis: Fig.4.1.Block diagram of

fingerprint image processing.





4.1.1.Pre-processing:-

1.Fingerprint Image Reading In this step, the fingerprint image is loaded by using the Matlab built-i

function imread. 2.RGB to Gray Image Rdata of fingerprint image obtained is in the RGB format. The

fingerprint image in the RGB format is then transformed into gray scale image so that image processing

can be executed on the image. Matlab built-in function, rgb2gray is used to convert RGB image

to gray scale image. I=rgb2gray( RGB);

4.2.2.Image Enhancement:

The quality of image is not good and thus the method of enhancement is to process the image obtained

so as to make it clearer by improving perception and hence the accuracy of matching will be increased.

By enhancing the image, the quality of image can be improved and thus the contrast between ridges and

valleys can be increased. Enhancement of image can make the following process easier. This process is

very important in keeping the performance of the fingerprint analysis at high accuracy.

4.2.3.Core point detection :





The core point is used to join between the fingerprints. The fingerprint authentication system faster than

the conventional method. Fingerprint may have multiple cores or no cores.

4.2.4.Image Segmentation :-

After enhancement of image the next step is fingerprint image segmentation. In general, only a Region

of Interest (ROI) is useful to be recognized for each fingerprint image. The image area without effective

ridges and furrows is first dispose of , since it holds only the background information. Then the hurdle

of the remaining successful area is sketched out since the minutiae in the hurdle region are confusing

with those spurious minutiae that are generated when the ridges are out of the sensor. To remove the

region of interest, two steps are followed: Block direction estimation and ROI extraction by

Morphological methods.

4.2.5ROI:-

ROI can be removed by Morphological operations ROI extraction is done using two Morphological

operations called OPEN and CLOSE. The OPEN operation can expand images and remove peaks

introduced by background noise. The ‘CLOSE’ operation can reduces images and eliminate small cavities

.

4.2.6.Feature Extraction:

For fingerprint recognition curvelet transform is used.

Purpose for using curvelet transform:

• multiscale and multidirectional transform.

• Used to represent edges.

• Reduced the dimentinality of the fingerprint image.

• Improve the identification rate.

Steps for curvelet transform:

We work all over in two dimensions, i.e., 𝑅2, with spatial variable x, with ω a frequency domain

variable, and with r and θ polar coordinates in the frequencydomain. We start with a pair of windows W(r)

and V





(t), which we will call the “radial window” and

“angular window,” respectively. These are both smooth, positive and real valued, with W taking positive

real arguments and supported on r (1/2,2) and

V taking real arguments and supported on t [−1,1].

These windows will always obey the admissibility conditions:

, 𝑟 ∈ (3⁄4 , 3⁄2); ...(1)

, 𝑡 ∈ (−1⁄2 , 1⁄2); ...(2)

Now, for each j ≥ j0, we introduce the frequency window Uj defined in the Fourier domain by

2𝜋

where [j/2]is the integer part of[ j/2]. Thus the support of Uj is a polar “segment” defined by the

support of W and V , the radial and angular windows, applied with scale-dependent window widths in

each direction. To obtain real-valued curvelets, we work with the symmetrized version of (2.3), namely,

Uj(r,θ)+

Uj(r,θ + π).

Define the waveform ∅𝑗(𝑥) by means of its Fourier transform ∅𝑗(𝜔) = 𝑈𝑗(𝜔). We may think of ϕj

as a “mother” curvelet in the sense that all curvelets at scale 2−𝑗 are obtained by rotations and

translations of ϕj. Introduce

• the equispaced sequence of rotation angles

𝑗 𝜃𝑙 . 𝑙, 𝑤𝑖𝑡ℎ 𝑙 = 0,1, … .. such that 0≤ 𝜃𝑙`< 2π (note that the spacing between consecutive angles

is scale-dependent),

•the sequence of translation parameters

K=(k1,k2) 𝑧2 With these notations, we define curvelets (as function of x = (x1,x2)) at scale 2−𝑗

𝑗 orientation `𝜃𝑙 & position 𝑥𝑘(𝑗,𝑙) = 𝑅𝜃𝑙

−1(𝑘1. 2−𝑗, 𝑘2. 2−2 ) by

𝜑𝑗,𝑙,𝑘(𝑥) = 𝜑𝑗 (𝑅𝜃𝑙(𝑥−𝑥𝑘(𝑗𝑙))) ...(4)

where 𝑅𝜃 is the rotation by θ radians and 𝑅𝜃−1 its inverse (also its transpose),

𝑥𝑘(𝑗,𝑙) = 𝑅𝜃 = ( −cossin𝜃𝜃 cossin 𝜃𝜃 ),

−3𝑗 −𝑗𝑟)𝑉(2[2𝑗]𝜃) ...(3) given

by −𝑇 𝑈𝑗(𝑟,𝜃)=2 4 𝑊(2





𝑅𝜃−1=𝑅𝜃

𝑇= 𝑅−𝜃 ...(5)

[5] USING FFT:

The digital coronization suggests Cartesian curvelets of

3𝑗 the for �̃�𝑗,𝑙,𝑘(𝑥) = 2 4 �̃�𝑗 (𝑆𝜃𝑙𝑇 (𝑥 − 𝑆𝜃𝑙

−𝑇𝑏)) where b takes on the discrete values b:=(k1.2−𝑗, 𝑘2.

2−𝑗/2).The goal is to find a digital analog of the coefficients now c(j,l,k)= 𝑒𝑖(𝑠𝜃𝑙 𝑏,𝜔)𝑑𝜔

...(6)

Suppose for simplicity that 𝜃𝑙=0. To numerically evaluate (1) with discrete data, one would just (1)

take the 2D FFT of the object f and obtain �̂�, (2) multiply 𝑓 ̂ with the window �̂�𝑗, and (3) take the inverse

Fourier transform on the appropriate Cartesian grid

b=(k1. 2−𝑗,k2. 2−𝑗/2) The difficulty here is that for 𝜃𝑙 ≠ 0, we are asked to evaluate the inverse discrete

Fourier transform(DFT) on the nonstandard sheared

𝑗 grid 𝑆𝜃𝑙−𝑇(k1. 2−𝑗, k2. 2−

2) and unfortunately, the classical FFT algorithm does not apply. To recover

the convenient rectangular grid, however, one can passthe shearing operation to �̂� and rewrite (1) as

c(j,l,k)=

∫ 𝑓̂(𝑆𝜃𝑙𝜔)𝑈̂𝑗(𝜔)𝑒−𝑖(𝑏,𝜔)𝑑𝜔 …..(7)

Suppose now that we are given a Cartesian array f[t1, t2], 0 <t1, t2 < n and let f[n1, n2] denote its

2D discrete Fourier transform

�̂�[𝑛1, 𝑛2) = ∑𝑛𝑡1−,𝑡12=0 𝑓[𝑡1, 𝑡2]𝑒−𝑖2𝜋(𝑛1𝑡1+𝑛2𝑡2)/𝑛

……….n/2<n1,n2<n/2

which here and below, we shall view as samples^2

�̂�[𝑛1, 𝑛2)=�̂�(2𝞟𝒏𝟏, 2𝞟𝒏𝟐)

from the interpolating trigonometric polynomial, also denoted �̂�, and defined by

�̂�(𝟂𝟏, 𝟂𝟐) = ∑𝟎<𝑡𝟏,𝒕𝟐<𝑛 𝒇[𝒕𝟏, 𝒕𝟐]𝒆−𝒊(𝝎𝟏𝒕𝟏+𝟐𝒕𝟐)/𝒏 ...(8)

Assume next that ˜Uj [n1, n2] is supported on some rectangle of length L1,j and width L2,j

𝑝𝑗 = {(𝑛1, 𝑛2): 𝑛1,0 ≤ 𝑛1, < 𝑛1,0 + 𝐿1, 𝑗, 𝑛2,0 ≤ 𝑛2 <

𝑛2,0 + 𝐿2, 𝑗}, …(9)

(where (n1,0, n2,0) is the index of the pixel at the bottom-left of the rectangle.) Because of the parabolic

scaling, L1,j is about 2j and L2,j is about 2^-j/2. With these notations, the FDCT via USFFT simply

evaluates �̂�[𝑛1, 𝑛2 − 𝑛1 tan 𝜃𝑙]=�̂�(2𝞟𝒏𝟏, 2𝞟(𝒏𝟐 −

𝒏𝟏 tan 𝜃𝑙))) ...(11)

and is therefore faithful to the original mathematical transformation.

This point of view suggests a first implementation we shall refer to as the FDCT via USFFT, andwhose

architecture is then roughly as follows

1. Apply the 2D FFT and obtain Fourier samples f[n1, n2], −n/2 _ n1, n2 < n/2.

∫ 𝑓 ( 𝜔 ) 𝑈 𝑗 ( 𝑆 𝜃𝑙 − 1 𝜔 ) 𝑒 𝑖 ( 𝑠 𝜃𝑙

− 𝑇 𝑏 , 𝜔 ) 𝑑𝜔 =





2. For each scale/angle pair (j, L), resample (or interpolate) ˆ f[n1, n2] to obtain sampled valuesˆ

f[n1, n2 − n1 tan θι] for (n1, n2) 2 Pj .

3. Multiply the interpolated (or sheared) object ˆ f with the parabolic window Uj effectively

localizing f near the parallelogram with orientation θι ,and obtain �̃�𝑗,𝑙[𝑛1, 𝑛2] = �̂�[𝑛1, 𝑛2 − 𝑛1 tan

𝜃𝑙]�̂�𝑗[𝑛1, 𝑛2]

4. Apply the inverse 2D FFT to each ˜ fj,`, hence collecting the discrete coefficients𝐶𝐷(𝑗, 𝑙, 𝑘) Of

all the steps, the interpolation step is the less standard and we shall see that it is possible to design an

algorithm which, for practical purposes, is exact and takes O(n2 log n) flops for computation, and requires

O(n2) storage, where n2 is the number of pixels.

[6] OTP WORKING:

For implementing OTP, we will make use of GSM modem to send SMS (an OTP) to user’s mobile

number. The idea to use mobile phones is preferred over e-mail because the people in rural areas have

simple phones which can receive text messages but have no internet connections and e-mail facilities.

Since mobile phones are ubiquitous, we intend to use mobile phones so that everyone can take the benefit

of the new is of 6-digit. User gets three chances to enter the code. If the code is entered incorrectly in three

consecutive attempts account gets temporarily blocked and notification is sent to registered mobile

number. This feature is added in order to restrict the fraudulent means of attacking the account of a user

by wearing masks or in rare cases, if unauthorized user’s face mistakenly matches authorized user’s face.

Random Number Generation

Generation of sequence of Pseudo-Random Numbers, (Yn):

Y n+1 =(a X Yn+C)mod (m)

Choices of a (multiplier), C (increment) and m (modulus) are important because random numbers

generated will be in sequence if not handled properly.

Proposed Random Number Generation formula The drawback of the above random number generator is

that the sequence has a finite number of integers and the sequence gets repeated over a period of time11.

Therefore, we have modified the formula by applying the same random number generator formula to ‘C’

and this value is substituted in the random number generator’s increment.

So the new random number generator formula will be:

C = ( b X n + d ) mod (m)

X n+1= C

Y n+1 =( a X Y n + C ) mod (m)

The random number(Y n+1) generated will be the OTP.





The value of ‘m’ should be a large prime number in order to distinct unrelated numbers. Though the

overhead is increased due to computation, but the repetition of a sequence is completely eliminated.

Cryptographic hash functions

Various Cryptographic hash functions are used to improve the security level. We have chosen MD5 also

knownas Message Digest because it is widely used hash function. Since, it is the fastest cryptographic

hash function, it is convenient to use MD5 and is mostly accepted by a wide variety of platforms.

stored in a central or local database or even on a smart card. This technique is very useful in future for

avoiding the fraud in ATM system.

[7] CONCLUSION:

Biometrics, in particular face scanning, continues to gain acceptance as a reliable form of securing

access through identification and verification process.

Steps

1. A 6-bit OTP is generated using the random number generator technique.

2. This OTP generated is texted to a user’s mobile phone number.

3. This OTP undergoes MD5 hashing technique thus converting it into encrypted form and is

temporarily stored in the database which will be erased after one minute. 4. The user will have to enter the

OTP within one minute time limit.

5. The user’s entered OTP again undergoes similar hashing technique and is compared with the stored

temporary encrypted OTP value in database 6. If it matches, then the transaction can be proceeded.

7. Steps 1-5 are repeated for every new transaction

[8] FUTURE SCOPE:

A lot of criminals tamper with the ATM terminal card details by illegal means. Once user’s ATM

card is lost and password is stolen, the user’s account is vulnerable to attack. Traditional ATM system

authenticate generally by using card and a password or PIN which no doubt has some defects. Biometrics

authentication technology may solve this problem since a person’s biometric data is undeniably connected

to its owner, is non-transferable and unique for every individual. The system can differntiate scans to record

requires less computation time and less storage space as a trainee images are stored in the form of their

projections on a reduced basis. OTP is used to improve security level.

The system has successfully overcome some of the aspects existing with the present technologies ,by

the use of face and OTP as the authentication technology.

PCA based face recognition is very accurate,





REFERANCE:

1. Elsevier B.V., “Enhanced security for ATM machine with OTP and Facial recognition features”,International

conference on advanced computing technologies and applications . pp.390-396,2015.

2. Ruggero Donida Labati,Angelo

Genovese,vincenzo Piuri,and Fabio Scotti, “Toward unconstraint fingerprint recognition:A fully touchless 3D

system based on two views on the move”,IEEE transactions on systems, man ,and cybernetics

:system,pp.2168-2216,2015.

3. Avinashkumar ojha, “ATM security using

fingerprint recognition”,International journal of advanced research in computer science and software

engineering. volume 5,Issue 6,pp.172-175,June 2015.

4. N.Shanmuga priya,S.Shanthini

Devi,K.Rajesh,R.Swathy,” A Statistical analysis of curvlet based fingerprint identification and authentication”,

international journal of electronics and communication technology.volume3,Issue1, pp.213-216,2012.

5. D.Shekar goud,Ishaq Md ,PJ.Saritha. “A secured approach for authentication system using fingerprint and

iris”,Global journal of advanced engineering technologies ,volume 1,Issue3, pp.170-174-2012.

6. Hugh Wimberly,Lorie M. Liebrock, “Using fingerprint authentication to reduce system security:An empirical

study, IEEE symposium on security and privacy.pp.32-36,2011.

7. Mrs. Sampada Dhole, “Face Recognition

Using Hidden Markov Modelling”,

International Journal of Biometrics and Bioinformatics, Vol 3, No. 10 ,November 2011(Impact Factor=.56)

8. Mrs.Dhole S.A. & Dr. Prof. Patil V.H,

“Review of Multimodal Biometric

Identification Using Hand Feature and Face”, Bulletin of Electrical Engineering and Informatics, IAES

Indonesia section in collaboration with UAD,Vol.1, No.3, pp. 179184, September 2012.

9. Mrs.Dhole S.A. & Dr. Prof. Patil V.H, “Face Authentication Technique Robust to

Orientation Using Pca Features”, International

Journal of Applied Engineering and

Technology, CIB_Tech, Vol. 5 (1), pp.102109, January-March2015.

10. Mrs.Dhole S.A. & Dr. Prof. Patil V.H., “ Performance Comparison of Face Recognition

Using PCA , DWT & Gabor Bank Features”,

3rd International conference on Recent Trends in engineering & technology( ICRTET), Elsevier Publication

,Volume 3, March 2014.

11. Mrs. Sampada A.. Dhole, “A human authentication System using Fingerprint”, International conference on

advances in electronics and communications (ADELCO), National engg. College, Kovilpatti, Tamilnadu, India,

Feb 1-3, 2007.

12. Mrs. Sampada A. .Dhole, “Human Identity Authentication system using fingerprint Biometric Technique”,

International conference on Emerging Techniques in computing Electronics, Embedded system & VLSI

Design(ICEVD), March 20-21,2008



Date post:	10-Dec-2021
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times

TM SECURITY USING FACE AND FINGERPRINT RECOGNITION

Documents