All submissions of the EM system will be redirected to Online Manuscript Submission System. Authors are requested to submit articles directly to Online Manuscript Submission System of respective journal.


C Yuganya1, A V Khirthana2, and Udayakumara pandian3
  1. The Computer Software Department, Bharath University, Chennai, TamilNadu, India
  2. The Computer Software Department, Bharath University, Chennai, TamilNadu, India
  3. The Computer Science Department, Bharath University, Chennai, TamilNadu,India
Corresponding Author: C Yuganya, E-mail: [email protected]
Related article at Pubmed, Scholar Google

Visit for more related articles at Journal of Global Research in Computer Sciences.


The bio-metric is ultra secure and more than one form of biometric identification is required. To use a combination of different biometric recognition we are using multi-modal biometric recognition. In this paper, the multi-level wavelet transform algorithm that combines information from palm, face, iris, and digital signature images and recognition of signature that makes use of biometric traits to recognize individuals. Multimodal biometric systems technology uses more than one biometric identifier to compare the identity of the image..A multi modal biometric system of iris and palm print based on Wavelet Packet Analysis is described. The visible texture of a person's face, palm, iris and signature is encoded into a compact sequence of 2-D wavelet packet coefficients, which generate a "feature vector code". The multi-resolution approach based on Wavelet Packet Transform (WPT) for texture analysis and recognition of face, iris, palm, and signature images. WPT sub images coefficients are quantized into 1, 0 or -1 as multi-resolution. The input of the iris,palm,face and signature is matched and stores as a scrumbled image usin DWT algorithm,then reconstruct the method and then gives the result to get access into the data.

Index terms

Wavelet Packet, Face, Palm, Iris, digital signature, recognition, bio-metric, multi level.


A major motivation for using biometrics is the ability to authenticate the true identity of an individual. Sometimes the problem is associated with noisy data that collapse and loose datas.All the process left to time consumption. To overcome these problems, multi-modal biometrics relies on more than one form of biometric data. The images can be separated into three types they are, 1.RGB (its value is 0-255), 2.Gray Scale (its value is 0-255), 3.Binary (its value is either 0 or 1). Image creates an image graphics object by interpreting each element in a matrix as an index into the figure's color map or directly as RGB values. once the preprocessing and gray conversion of the image is completed it is set to resize the image in short it is said to be as imresize.Later combine the preprocessed image, and to get the fussed image we apply IDWT(Inverse Discrete Wavelet Transform).The data-base process of DWT(Discrete Wavelet Transform) is stored as fused image. At last the recognize process by giving the individuals user name and the password which will reconstruct the image given by the user previously and match them and gives the output. The advantages of the proposed approach are reduction in memory size, increase in recognition accuracy due to the use of multi-modal biometric. The proposed system provides a very accurate outcome and more secured access to the database.


A successful implementation of iris recognition system was proposed in 1993[3].It was published more than 29 years ago still remains valuable since because it provides solutions for each part of the data. It is important that most of the data’s were secure. They are based on Gabor wavelet analysis [1] [2] [3] in order to extract iris, palm, face, and signature images. It consists in convolution of image with complex Gabor filters. The approach of the iris image is pre processed for contrast enhancement [4].After preprocessing; a ring mask is created and moved through the entire image to obtain the iris data. Using the data the iris and pupil are reconstructed from the original picture. Using the iris center and its radius, the iris was cropped out from the reconstructed image. The iris data (iris donut shape) is transformed into a rectangular shape. A scrumbled fusion iris pattern is matched. The method followed by Jie Wang [8] the iris texture extraction is performed by applying wavelet packet transforms (WPT)and DWT,IDWT algorithms were used.
The iris image is decomposed in to sub images by applying WPT and suitable sub images are selected and WPT coefficients are encoded. The method of iris features extraction method is designed [10].The iris imaging should capture a minimum of 70 pixels in iris radius. The iris radius of 100 to 140 pixels has been more typical. CCD monochrome cameras of (480 x 640) have been used because the NIR illumination in the 700nm - 900nm band was required for imaging to be invisible to humans view. A wide-angle camera is used to detect the eyes in faces that acquired higher resolution images of eyes. Many alternative methods for finding the iris image is used. An iris recognition system can be detected briefly such as an iris detector for detection and location of iris image, a feature extractor is to extract the features and a pattern matching module for matching the given input image the iris is extracted from the given image of the entire eye. To perform the iris pattern matching, the iris is localized and extracted from the acquired image. So, to resolve this process are using a wavelet-packet transform algorithm. In that, the standard discrete wavelet transform (DWT) is a powerful tool used successfully to solve various problems in image processing.
The DWT divides an image into four sub-sampled images. The results of the image that has been highly passed in the horizontal and vertical directions (HH) and one that has been low passed in the vertical and high passed in the horizontal (LH),the one that has been highly passed in the vertical and low passed in the horizontal (HL) and last that has been low pass filters in both directions (LL).The H and L are the high pass and low pass filter,instantly.While HH means that the high pass filter is applied to the images of both directions, represent diagonal features of the image, the horizontal size is equal to HL,the vertical size is equal to LH and LL is used for further process. Wavelet Packets Transform (WPT) is a wavelet transform that is used as a Image analysis. With WPT, it is possible to zoom into any desired frequency channels for further extraction. Compared with WT, WPT gives a very fine extraction. A algorithm of progressive texture classification of WPT gives better performance because the iris image frequencies are located in the low and middle frequency channels.
A digital signature is a term used to describe a data string which associates a digital message with an assigned person’s image only [5].The authentication, data integrity, and non-repudiation has large applications in Data security. One of the most significant advances in digital signature technologies is the development of the first practical cryptographic scheme called RSA, while it still remains as one of the most practical and versatile digital signature techniques available today.
A face recognition system is successfully deployed, it is fully automatic [6].A fully automatic system detects and identifies/varies a face in an image without human intervention [7]. Fully automatic face recognition systems generally have two components, recognition and detection. The recognition component identifies or varies the face. Usually, a recognition module requires that the face be in a standard position. A new image of a face is transformed into an image components, then compare this image with our given input image and it is extracted using DWT algorithm and it is extracted scrum bled and stored after the the process gets completed the matching process reconstruct’s and match both the input image and original secure image. The image has its difference with the matrix value. The given input image of an face is a process of fusion, and then the face is extracted and reconstructed [6].The inputs will match to the response of the image of iris, palm, face and digital signature. It can be used in setting the user name and the passwords to the personal data where your face, iris, palm, and digital signature is taken as the password of your personal data. This can be used to the various set of data’s to maintain and secure projects are objectives with their faces, iris, palm, and signature as their data identity. This is helpful to maintain our data’s and is thread safe. Every person has their unique identity of their iris, palm, face and signature as their identity. The basic input of the face recognition is vey important. All the recognition is used in mat lab using its image processing applications and tool box.
The palm print of an image is the feature extraction of the acquisition of raw informed acquarency.The image matching process with the variation and similarity with the given image process. Performance and image verification mode are indentified. Proposed image analysis is the limited advantages to match the process. The palm line feature and the wrinkles are determined using some techniques such as, Datum point determination finds and locate the endpoints of each palm lines. Each line is located inside the palm parallel. The end point are closer to the finger wrist. Digital palm print images are use for clear and accurate matching for data base protection. The wavelet based DWT transform is used. To recognize the palm print image Gabor filter is used [11].The palm print as the feature of Line Edge Map (LEM) ,and for distance matching Hausdorff distance algorithm were used[12].The interlaminar interactions of neural network executes two layers, the one has a fixed weights and other has a layer with adjustable weights, the co-efficient of 2-D Gabor transforms without the related conditions. The image has no complete transforms, in which the coefficients may be interpreted by signifying the presence of a certain features in the image; the network finds optimal coefficients in the sense of minimal mean-squared-error in representing the image. The normal image should have 7.57 in the pixel representation to 2.55 in the complete 2-D Gabor transform. In “wavelet” expansions based on a WPT and DWT of 2-D Gabor wavelet template, image compression is elaborated with ratios up to 20: 1.The image segmentation is demonstrated based on the clustering of coefficients in the complete 2-D Gabor transform.


A image processing recognition system is proposed, to safe a individual data’s that has been stored by a particular user. The first thing is, collect a different type of persons face,iris,palm,and signature images, each person have their identity face,iris,palm and signature images of their’s.When the input is given the process starts by matrix method. There are three steps going to be accomplished .They are preprocessing, data-base process and recognition.


The persons or an individual image that was already given and stored .The given image may be in JPG or any type of image. When the person gives the input of their Face,palm,iris,and signature images the process start to convert the image, first the gray conversion method occurs, In this the given image is converted into gray shaded image and the four input image is subploted,and then image is resized.
The input of the face, palm, iris, and signature images are given. A normal image has three variations and has its value rate such as, RGB (0-255), Grayscale (0-255) and Binary (0 or 1).Using the values we change the image color’s, like that we change the value of the given input image and convert them in to gray. If we give the binary value as 0(then the image will be in black),else we give 1(the image will be in white).After converted the image started to read the image by giving the syntax as “a=imread(“water.jpg)”the image starts to read and show the image by giving the code line as “Figure,imshow(a)”.Now we are converting the RGB image in ti gray so we include as=rgb2gray(a);”when this syntax is given it is converted and shown.
The subplot syntax is given below”h = subplot(m,n,p) or subplot(mnp)”. The subplot divides the current figure into rectangular panes and they are numbered row wise. Every row and column are calculate b matrix method. Subsequent plots are the output to the current pane. The syntax = subplot(m,n,p) or subplot(mnp)” breaks the figure window into an m-by-n matrix of small axes, and selects the path axes object to the subplot, and returns the matrix . The matrix value are counted along the top row of the figure window, then the second row, etc. For example,
subplot(2,1,1), plot(income)
subplot(2,1,2), plot(outgo)
The plots income on the top half of the window and outgo on the bottom half. If the Current matrix is plotted to the entered value of the panel, the panel is used as the parent for the subplot instead of the current image. The new axes object becomes the current axes. Replace syntax “subplot(m,n,p,'replace')”.If entered matrix already matchesexists,delete it and create new axes.
A resize method is used in matrix. By using scale method the image gets resized. A normal value of an image is <600*800*3unit8>, after resized with 0.3we get<120*240*3 unit8>.A image resize syntax is given below:
B = imresize(A, scale)
B = imresize(A, [mrows ncols])
[Y newmap] = imresize(X, map, scale)
The B = imresize(A, scale) returns a image B that is scale times the size of A. The given input matrix image A can be a grayscale image, RGB, or binary image. If scale is between 0 and 1.0, B is smaller than A. If scale is greater than 1.0, B is larger than A. The “B = imresize(A, [mrows ncols])” returns the image B that has the number of rows and columns specified by [mrows ncols]. Either NUMROWS or NUMCOLS may be null, then the imresize computes the number of rows or columns automatically to preserve the image aspect ratio.
The “[Y newmap] = imresize(X, map, scale)’ resizes the indexed image X. The scale can either be a numeric scale specifies the size of the output image” ([numrows numcols])”,A default imresize returns a new, matrix featured image of color map (newmap) with the resized image. To return a color map that is the same as the original color map, to use the 'Colormap' parameter (given below).


In this process the converted image is combined with the IDWT (Inverse Discrete Wavelet Transform) algorithm and gets an fused image and then al the process were reconstructed. The converted image is combined and scrum bled and reconstructed using DWT (Discrete Wavelet Transform) algorithm and then stores as a fused image. The dwt (Computes discrete wavelet transform DWT of input) block computes the discrete wavelet transform (DWT) of each column of the matrix image. The output is a reconstructed fused and scrum bled matrix image with the same given identity as the input. Each column of the output is the DWT of the corresponding input column. The Wavelet Toolbox were installed product for the block to automatically design wavelet-based filters to compute the DWT. Otherwise, you must specify your own low pass and high pass FIR filters by setting the Filter parameter to User defined. For the same input, the DWT block and the Wavelet Toolbox function do not produce the same results. Because the block set is designed for real-time implementation and the toolbox are said to be designed for analysis, the matrix image subplots are the output to the various given data. To make the output of the dwt function and the DWT block match, complete the following steps:
a. For the dwt function, set the boundary condition to zero-padding by typing dwtmode('zpd') at the MATLAB® command prompt.
b. To match the latency of the DWT and IDWT, are implemented using the FIR filters, add zeros to the input of the dwt function. The number of zeros you add must be equal to the half the filter length.


In this process reconstructs the methods and matches with the users input and then recognize. If any one of the images is mismatched it won’t recognize till it match. The idwt alogarithm performs a single-level one-dimensional wavelet reconstruction transform of the image with respect to either a particular wavelet (wname) or particular wavelet reconstruction filters (Lo_R and Hi_R) that were specify.
The syntax were given as:
X = idwt(cA,cD,'wname')
X = idwt(cA,cD,Lo_R,Hi_R)
X = idwt(cA,cD,'wname',L)
X = idwt(cA,cD,Lo_R,Hi_R,L)
X = idwt(...,'mode',MODE)
X = idwt(cA,cD,'wname') returns the single-level reconstructed approximation coefficients vector X based on approximation and detail coefficients vectors cA and cD, and using the wavelet 'wname'. X = idwt(cA,cD,Lo_R,Hi_R) reconstructs as above using filters which is specifed.
* Lo_R is the reconstruction low-pass filter.
* Hi_R is the reconstruction high-pass filter.
Lo_R and Hi_R must be the same length. Let la be the length of cA (which also equals the length of cD) and lf the length of the filters Lo_R and Hi_R; then the length(X) = LX where LX = 2*la if the DWT extension mode is set to per iodization. For the other extension modes LX = 2*la-lf+2.
X = idwt(cA,cD,'wname',L) or X = idwt(cA,cD,Lo_R,Hi_R,L) returns the length-L central portion of the result obtained using idwt(cA,cD,'wname'). L must be less than LX. X = idwt(...,'mode',MODE) computes the wavelet reconstruction using the specified extension mode MODE. X = idwt(cA,[],...) returns the single-level reconstructed approximation coefficients vector X based on approximation coefficients vector cA.X = idwt([],cD,...) returns the single-level reconstructed detail coefficients vector X based on detail coefficients vector cD.




The experimental results clearly demonstrate that the candidate’s bio-metric image processing. By matching the iris, palm, face, and signature of the image, are fused and scrum bled as an distracted image using IDWT algorithm and at last it is reconstructed and matches with the given input and then show the result as either it is matches are mismatched. The overall recognition of the data is improved. On the other hand, the storage for all these data requires very less memory in the data.


The 2D cross correlation has a statistical results were conducted with multi-model biometric images in which the user was looking directly at the imaging device. The purpose for approaching the issue of multi-modal pattern extraction without the assumption that patterns are circular is to allow for the extraction of fused and scrum bled biometric images. This approach is applicable to non-orthogonal biometric images, because the image is rotated away from the normal to the imaging device, current commercial systems develop complications extracting and authenticating the singe biometric pattern. As for it this system multimodal biometric recognition is proposed. It is a 2D wavelet packet transform method.


  1. John Daughman "Complete Discrete 2-D Gabor Transforms by Neural Networks for Image Analysis and Compression", IEEE Transactions on Acoustics, Speech and signal Processing, VOL.36, No. 7, July 1988
  2. John Daughman, "High confidence visual recognition of persons by a atensdt Mofasctahtiinseticinalteilnlidgeepnecned, ence", IEEE Transactions on Pattern Analysis VOL.15, No.ll, November 1993
  3. John Daughman, "How iris recognition works" IEEE Transactions on Circuits and Systems for Video Technology, VOL.14, No. 1, January 2004.
  4. Lye Wil Liam, Ali Chekima, Liau Chung Fan, "Iris recognition using self-organizing neural network", IEEE 2002.
  5. Taekyoung Kwon and Jae-il Lee” Practical Digital Signature Generation usingBiometrics”
  6. P. Jonathon Phillips, R.MichaelMCCabe” Biometric Face Recognition and Image Pocessing”.
  7. Bradford Bonney, Robert Ives, Delores Etter, "Iris pattern extraction using bit planes and standard deviations", iEEE 2004
  8. Lu Chenghong, Lu Zhao yang, "Efficient iris recognition by computing discriminable textons", IEEE 2005.
  9. Jie Wang, Xie Mei, "Iris Feature extraction based on wavelet packet analysis", iEEE 2006
  10. K. Grabowski, W. Sankowski,"lris recognition algorithm optimized for hardware implementation", IEEE 2006.
  11. Ajay Kumar, Helen C. Shen, "Palm print Identification using Palm Codes", Proceedings of the Third International Conference on Image and Graphics, 2004.
  12. Sree Rama Murthy kora,Praveen Verma, Yashwant Kashyap.”Palm print recognition” and the digital extracrion.
  13. Fang Li, Maylor K. H. Leung, Xiaozhou You, "Palmprint Identification Using Hausdorff Distance", 2004 IEEE International Workshop on Biomedical Circuits & Systems, 2004.
  14. www.
Global Tech Summit