SlideShare a Scribd company logo

Paper id 36201507

•
•
I
IJRAT

This document presents a new approach for fingerprint matching called the Minutia Cylindrical Code (MCC) approach. It involves extracting minutia points from fingerprint images, then generating a code for each fingerprint based on the local structure and spatial relationships of minutia points within a cylindrical neighborhood. MCC codes make the fingerprints invariant to scale and rotation. The approach is tested on a database of 200 fingerprints and achieves false acceptance ratios between 6-13% and false rejection ratios below 0.12% depending on the threshold used. The MCC approach performs fingerprint matching efficiently while maintaining accuracy even when fingerprints are rotated or scaled.

Related slideshows

El4301832837
El4301832837El4301832837
El4301832837
 
FACE RECOGNITION USING DIFFERENT LOCAL FEATURES WITH DIFFERENT DISTANCE TECHN...
FACE RECOGNITION USING DIFFERENT LOCAL FEATURES WITH DIFFERENT DISTANCE TECHN...FACE RECOGNITION USING DIFFERENT LOCAL FEATURES WITH DIFFERENT DISTANCE TECHN...
FACE RECOGNITION USING DIFFERENT LOCAL FEATURES WITH DIFFERENT DISTANCE TECHN...
 
SEMANTIC IMAGE RETRIEVAL USING MULTIPLE FEATURES
SEMANTIC IMAGE RETRIEVAL USING MULTIPLE FEATURESSEMANTIC IMAGE RETRIEVAL USING MULTIPLE FEATURES
SEMANTIC IMAGE RETRIEVAL USING MULTIPLE FEATURES
 
1 of 7
Download to read offline
International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 32 Minutia Cylindrical Code Based Approach for Fingerprint Matching Dilip Tamboli1 , Mr.Sandeep B Patil 2 , Dr.G.R.Sinha 3 1 P.G. Scholar, Department of Electronics & Telecommunication Engg. SSGI Bhilai, C.G.India 2 P.G. Scholar, Department of Electronics & Telecommunication Engg. SSGI Bhilai, C.G.India2 3 Assosiate Director & Prof., Department of Electronics & Telecommunication Engg. SSGI Bhilai, C.G.India3 diliptamboli_123@yahoo.com 1 , Patilsandeepb1212@gmail.com 2 ,drgrsinha@ssgi.edu.in3 Abstract- Impression from the fingers and its matching is one of the important task of law enforcing body. Minutia extraction from the fingerprint image decides the accuracy of the matching. Method presented in this paper is also very efficient in matching the algorithm. Minutia cylindrical code(MCC) which is local descriptor of the fingerprint image is used for matching the fingerprint image. MCC, codes the local direction and distance between the minutia and hence invariants to the scale and rotation. False Acceptance ratio (FAR) and False Rejection ration(FRR) is also computed to test the accuracy of the proposed method. Index Terms- Minutia, FFT, MCC, Euclidean distance 1. INTRODUCTION Recognition of fingerprint is one of the complicated pattern recognition problems which is being studied for the last 40 years. Though various efficient algorithm have been designed for fingerprint matching, it cannot be concluded that this problem has solved. Accuracy, interoperability and computational efficient algorithm are still an open issue [1] in fingerprinting matching. Most of the current fingerprint matching algorithms are based on the minutia. Special ridge pattern are called minutia. Ridge ending and ridge bifurcation are some of the minutia. In the past, minutia matching is considered as the two dimensional pattern matching problem for aligning the two minutia pair. This forced the researcher to find all the possible transformation for two minutia matching. Hough transform is one of the solution of this problem[2][3]. High computational cost and lack of robustness are some the problems of the global minutia matching algorithm. In the last few years, these problem is addressed by introducing the local minutia descriptor and its matching. The characteristics of the local minutia descriptor is such that it is invariants to the global transformation and therefore appropriate for matching without the need of global alignment. Since local minutia descriptor based matching is based on the arrangement of the local property therefore it excludes the global feature and give better results. Global matching algorithms have some of the benefits which is not possible in case of local descriptor based matching and hence an hybrid approach can be applied to get the benefit of both the algorithm. In this types of approach, first of all the local structure is used to match the minutia quickly and robustly. In the next step, matching at global level is performed for validation purpose. The evaluation of the minutia matching based on the local structure passed in three stages. Stage one correspond to the earlier approach in which local structure are formed by considering the minutia lies inside some regions. In this approach no global validation was performed [4][5]. The approach adopted by the [6] and [7] comes in the second stage. [6] and [7] were the first to establishe a relationship between minutia and its neighbourhood in term of invariant angle distance and structure. In this stage, global validation was also performed . Third stage comprises of the method proposed by Jiang and Yau[6] and Ratha [7]. These methods are variants of the method proposed by the same author. In this phase they extend the feature set of the fingerpring by incorporating local ridge, local frequency, shape its. [1] contain the exhaustive review in fingerprint matching and recognition. For further reference reader may go through the [8]-[33]. Two types of local minutia structure were proposed. First one is based on the nearest neighbourhood while the other one is based on the fixed radius. In the nearest neighbourhood [6] approach, centre minutia is characterized by the K- closest minutia which are spatially arranged. This gives the fixed length descriptor which can be matched very efficiently. The later approach is given by the [7]. In this approach, neighbours are defined as the all the neighbours which are inside the radius R of the centre minutia. In this approach, the length of the descriptor is not fixed making it difficult to match the fingerprint locally. But this approach shows better tolerant against missing and spurious minutia.
International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 33 2. PROPOSED METHODOLOGY In this paper local descriptor based fingerprint recognition is presented. Minutia cylinder code is used as local descriptor. Block diagram of proposed methodology is shown in figure 2. First of all, with the help of scanner, Fingerprint image of the input finger is registered. Minutia extraction block is used to extract out the minutia from the image. With the help of minutia, a cylindrical code for minutia is formed. Minutia cylindrical code represent minutia local structure of the minutia distribution. Once, minutia cylindrical code is generated for each finger print then a database is created which contain the minutia cylindrical codes of all the fingerprint image. The last step of this method is matching or testing. Next section describe the functioning of last four block of this algorithm. Fig.1 Basic Flow Diagram of Proposed Method 2.1. Minutia Extraction Process In any fingerprint recognition system, Minutia extraction is one of the important phase because the performance of any fingerprint recognition system depends on the how accurately minutia are extracted. In this algorithm following steps have been adopted for minutia extraction. 2.2 FP Image Enhancement First step in any fingerprint recognition system is the enhancement of the fingerprint image. This step is adopted to improve the quality of the fingerprint image so that all the minutia points become more clear and extraction of the minutia points become easier. Generally, Fingerprint image obtained by the fingerprint scanner is noisy and of poor contrast. Ridge and furrows are not visible due to the poor contrast. Image enhancement increase the contrast of ridge and furrows and also joined some of the broken line which occurred in fingerprint image due to sensor deficiency. 2.3 Histogram equalization Histogram equalization basically adjust the intensity distribution of the histogram and with this approach it improve the global contrast of the image. This step helps the lower contrast region to gain a higher contrast while keeping the global contrast intact. This method basically spread out the most frequently occurred intensity value. Figure 3 represent the histogram before and after applying histogram equalization. 2.4 Fast Fourier Transform. This process is accomplished for connecting the ridge broken line . In this process first of all, Fingerprint image is divided in to a block of 32 * 32 pixel and then Fourier transform of this block is computed using following formula. for u = 0, 1, 2, ..., 31 and v = 0, 1, 2, ..., 31. 2.5 Image Binarization This process is performed to convert the 8-bit gary image to 1 bit image. In this, 0 represent the ridge and 1 represent the furrows. This step highlight the ridge and furrows clearly in the image. A local adaptive binarization method is used for this purpose. 2.6 Fingerprint Segmentation This process is accomplished to extract out the region of interest from the fingerprint image. In fingerprint image, some of the area is not useful and carry insignificant ridge and furrows. It is necessary to discard this area . Segmentation process consist of two Finger Finger print Scanner Minutia Extraction Minutia Cylindrical code formation Database Creation Matching Algorithm Fingerprint Registration
International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 34 part i.e. Estimation of block direction, extraction of region of interest by applying morphological operation. 2.7 Estimation of Block Diagram Block direction estimation is performed to separate out background and foreground information. In finger print image, foreground is represented by the ridges and furrows while the background represent the no information. In order to compute the block diagram estimation, a block of 16 *16 pixel is chosen and then block direction is computed by using below mentioned formula tan2ß = here gx= Gradient along x-direction gy= Gradient along y-direction Certainity level for each block is then computed using formula- E = Here W= size of block (i.e. 16 in this case) The value of E is compared with the threshold value to decide whether the block comes under the background( if E<threshold) or foreground (i.e if E>threshold). 2.8 Morphological Operation for ROI extraction In the last step morphological operation “opening” and “closing” is applied to to extract the region of interest. 2.8.1 Ridge Thinning This operation is applied to make the ridge width 1 pixel wide. This is performed by thinning operation. 2.8.2Minutia Marking Next step is applied to mark the minutia by applying a 3*3 window. 2.8.3False Minutia Removal Inter-ridge distance is used for eliminating the false minutia points which may affect the over all accuracy of the fingerprint recognition system. 2.8.4Minutia Marking At the final step each true minutia is marked with three element. i.e. x-coordinate, y-coordinate and the minutia direction or angle. 2.9 Minutia Cylindrical Code Information Once the minutia information is obtained then the next step is to form the minutia cylindrical code. In minutia cylindrical code, a cylinder with 6 different layer is used to represent the relative direction and distance among different minutia. One such cylinder is shown in figure 3. Figure 3 represent the cylinder associated with the minutia m={xm, ym, θm}. Radius of the cylinder is R and xm and ym is the center of this cylinder. As shown in the figure, a cuboid encloses the cylinder whose base is aligned as per the direction of the minutia i.e θm. The cuboid is divided in to a cell where The size and height of each cell is × and respectively. Where Fig.2 Minutia Cylinder Each cell of the cylinder can be accessed by three coordinate i.e. i, j, k which represent the position of the cell inside the cylinder. The value of i and j is between 1 and Ns while the k takes the value between
International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 35 1 to Nd. The height k of the cylinder represent the angle which is represented by the If Represent the centre of the cell having indices i,j. In MCC, the contribution of each minutia for each cell is computed and is given by Which is computed by the spatial contribution and directional contribution Where spatial contribution is given by And directional contribution is given by And Is the neighbourhood of the . Here is the neighbourhood radius and ds (p, m) represent the Euclidean distance between minutia m and point p as shown in the figure given below- Fig.3 Neighbourhood Radius Once the MCC for all the fingerprint image is obtained then a database of MCC is created. 2.10 Fingerprint Matching In order to match the two fingerprint MCC, Euclidean distance is computed between two fingerprint MCC as given in the below Where = MCC of the testing fingerprint image = MCC of the database. MCC for each fingerprint image is rotation and scale invariants therefore it gives better result in even if the fingerprint image is rotated or scaled. The database MCC for which the value of Ed is less than the certain threshold is considered as the matching one and corresponding fingerprint image is considered as the matching image 3. EXPERIMENTAL RESULTS A database of fingerprint images comprises of the 200 different fingerprint images are created for testing the performance of the proposed method. A simulation program is designed with the help of MATLAB ver 2009B. The simulation program is tested in a core2duo processor computer with 2GB RAM.
International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 36 Fig.4 Original Fingerprint Image Figure 5 display the original fingerprint image while figure 6 display the extracted minutia points of the fingerprint image. Fig.5 Extracted Minutia Points in fingerprint image In order to evaluate the performance of this method, statistical measure like FAR(False acceptance Ratio) and FRR(False rejection ratio) is also computed and tabulated in table 1. Fr AR is computed using following formula Table 1 FAR and FRR for different value of Threshold (Without Rotation) Threshold Value FAR(in %) FRR(in %) 6 6.9 0.07 7 7.4 0.05 8 9.2 0.05 9 11.9 0.03 10 13.1 0.02 Table2 FAR and FRR of rotated fingerprint image for different value of Threshold Threshold Value FAR(in %) FRR(in %) 6 6.3 0.12 7 7.6 0.08 8 10.1 0.07 9 12.1 0.03 10 13.9 0.03 Fig.6 Minutia Cylindrical Code of the fingerprint image shown in figure 5 4. CONCLUSION Since from its inception, finger-print matching area has been in the search of some robust technique for matching the fingerprint accurately. The requirement of rotation , scale invariants re the need of the hour. In this paper a MCC based fingerprint matching algorithm is implemented and presented. The performance of the proposed system is satisfactory as clear from the table 1. Even after rotating and scaling the fingerprint, proposed system is able to achieve good accuracy.
International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 37 REFERENCES [1] D. Maltoni, D. Maio, A.K. Jain, and S. Prabhakar, Handbook of Fingerprint Recognition. second ed. Springer-Verlag, 2009. [2] N.K. Ratha, K. Karu, S. Chen, and A.K. Jain, “A Real-TimeMatching System for Large Fingerprint Databases,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 18, no. 8, pp. 799-813, Aug. 1996. [3] S.H. Chang, F.H. Cheng, W.H. Hsu, and G.Z. Wu, “Fast Algorithmfor Point Pattern Matching: Invariant to Translations, Rotations and Scale Changes,” Pattern Recognition, vol. 30, pp. 311- 320, 1997. [4] A.K. Hrechak and J.A. McHugh, “Automated Fingerprint Recognition Using Structural Matching,” Pattern Recognition, vol. 23, pp. 893-904, 1990. [5] A.J. Willis and L. Myers, “A Cost-Effective Fingerprint Recognition System for Use with Low-Quality Prints and Damaged Fingertips,” Pattern Recognition, vol. 34, pp. 255-270, 2001. [6] X. Jiang and W.Y. Yau, “Fingerprint Minutiae Matching Based on the Local and Global Structures,” Proc. Int’l Conf. Pattern Recognition, vol. 2, pp. 6038-6041, 2000. [7] N.K. Ratha, V.D. Pandit, R.M. Bolle, and V. Vaish, “Robust Fingerprint Authentication Using Local Structural Similarity,” Proc. IEEE Workshop Applications of Computer Vision, pp. 29-34, 2000. [8] T.Y. Jea and V. Govindaraju, “A Minutia-Based Partial Fingerprint Recognition System,” Pattern Recognition, vol. 38, pp. 1672-1684, 2005. [9] S. Chikkerur, A.N. Cartwright, and V. Govindaraju, “K-plet and Coupled BFS: A Graph Based Fingerprint Representation and Matching Algorithm,” Lecture Notes in Computer Science, vol. 3832, pp. 309-315, Springer, 2006. [10] D. Kwon, I.D. Yun, D.H. Kim, and S.U. Lee, “Fingerprint Matching Method Using Minutiae Clustering and Warping,” Proc. Int’l Conf. Pattern Recognition, vol. 4, 2006. [11] H. Chen, J. Tian, and X. Yang, “Fingerprint Matching with Registration Pattern Inspection,” Lecture Notes in Computer Science, pp. 327- 334, Springer, 2003. [12] J. Feng, “Combining Minutiae Descriptors for Fingerprint Matching,” Pattern Recognition, vol. 41, pp. 342-352, 2008. [13] X. Tan and B. Bhanu, “A Robust Two Step Approach for Fingerprint Identification,” Pattern Recognition Letters, vol. 24, pp. 2127-2134, 2003. [14] G. Parziale and A. Niel, “A Fingerprint Matching Using Minutiae Triangulation,” Lecture Notes in Computer Science, pp. 241-248, Springer, 2004. [15] X. Chen, J. Tian, J. Yang, and Y. Zhang, “An Algorithm for Distorted Fingerprint Matching Based on Local Triangle Feature Set,” IEEE Trans. Information Forensics and Security, vol. 1, no. 2, pp. 169-177, June 2006. [16] D.Q. Zhao, F. Su, and A. Cai, “Fingerprint Registration Using Minutia Clusters and Centroid Structure,” Proc. Int’l Conf. Pattern Recognition, pp. 413-416, 2006. [17] W. Xu, X. Chen, and J. Feng, “A Robust Fingerprint Matching Approach: Growing and Fusing of Local Structures,” Lecture Notes in Computer Science, vol. 4642, p. 134, Springer, 2007. [18] X. Linag, A. Bishnu, and T. Asano, “A Robust Fingerprint Indexing Scheme Using Minutia Neighborhood Structure and Low-Order Delaunay Triangles,” IEEE Trans. Information Forensics and Security, vol. 2, no. 4, pp. 721- 733, Dec. 2007. [19] M. Tico and P. Kuosmanen, “Fingerprint Matching Using an Orientation-Based Minutia Descriptor,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 25, no. 8, pp. 1009- 1014, Aug. 2003. [20] J. Qi and Y. Wang, “A Robust Fingerprint Matching Method,” Pattern Recognition, vol. 38, pp. 1665-1671, 2005. [21] E. Zhu, J. Yin, and G. Zhang, “Fingerprint Matching Based on Global Alignment of Multiple Reference Minutiae,” Pattern Recognition, vol. 38, pp. 1685-1694, 2005. [22] X. Wang, J. Li, and Y. Niu, “Fingerprint Matching Using Orientationcodes and PolyLines,” Pattern Recognition, vol. 40, pp. 3164-3177, 2007. [23] Y. He, J. Tian, L. Li, H. Chen, and X. Yang, “Fingerprint Matching Based on Global Comprehensive Similarity,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 28, no. 6, pp. 850-862, June 2006. [24] X. He, J. Tian, L. Li, Y. He, and X. Yang, “Modeling and Analysis of Local Comprehensive Minutia Relation for Fingerprint Matching,” IEEE Trans. Systems, Man, and Cybernetics, Part B, vol. 37, no. 5, pp. 1204- 1211, Oct. 2007. [25] H. Wei, M. Guo, and Z. Ou, “Fingerprint Verification Based on Multistage Minutiae Matching,” Proc. Int’l Conf. Pattern Recognition, pp. 1058-1061, 2006. [26] G.S. Ng, X. Tong, X. Tang, and D. Shi, “Adjacent Orientation Vector Based Fingerprint Minutiae Matching System,” Proc. Int’l Conf. Pattern Recognition, pp. 528-531, 2004. [27] X. Tong, J. Huang, X. Tang, and D. Shi, “Fingerprint Minutiae Matching Using the Adjacent Feature Vector,” Pattern Recognition Letters, vol. 26, pp. 1337-1345, 2005.
International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 38 [28] L. Sha, F. Zhao, and X. Tang, “Minutiae-Based Fingerprint Matching Using Subset Combination,” Proc. Int’l Conf. Pattern Recognition, pp. 566-569, 2006. [29] J. Feng, Z. Ouyang, and A. Cai, “Fingerprint Matching Using Ridges,” Pattern Recognition, vol. 39, pp. 2131-2140, 2006. [30] Y. Zhang, X. Yang, Q. Su, and J. Tian, “Fingerprint Recognition Based on Combined Features,” Lecture Notes in Computer Science, vol. 4642, p. 281, Springer, 2007. [31] D. Lee, K. Choi, and J. Kim, “A Robust Fingerprint Matching Algorithm Using Local Alignment,” Proc. Int’l Conf. Pattern Recognition, vol. 16, pp. 803-806, 2002. [32] L. Sha and X. Tang, “Orientation-Improved Minutiae for Fingerprint Matching,” Proc. Int’l Conf. Pattern Recognition, vol. 4, pp. 432-435, 2004. [33] Y. Feng, J. Feng, X. Chen, and Z. Song, “A Novel Fingerprint Matching Scheme Based on Local Structure Compatibility,” Proc. Int’l Conf. Pattern Recognition, pp. 374-377, 2006.

More Related Content

Paper id 36201507

  • 1. International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 32 Minutia Cylindrical Code Based Approach for Fingerprint Matching Dilip Tamboli1 , Mr.Sandeep B Patil 2 , Dr.G.R.Sinha 3 1 P.G. Scholar, Department of Electronics & Telecommunication Engg. SSGI Bhilai, C.G.India 2 P.G. Scholar, Department of Electronics & Telecommunication Engg. SSGI Bhilai, C.G.India2 3 Assosiate Director & Prof., Department of Electronics & Telecommunication Engg. SSGI Bhilai, C.G.India3 diliptamboli_123@yahoo.com 1 , Patilsandeepb1212@gmail.com 2 ,drgrsinha@ssgi.edu.in3 Abstract- Impression from the fingers and its matching is one of the important task of law enforcing body. Minutia extraction from the fingerprint image decides the accuracy of the matching. Method presented in this paper is also very efficient in matching the algorithm. Minutia cylindrical code(MCC) which is local descriptor of the fingerprint image is used for matching the fingerprint image. MCC, codes the local direction and distance between the minutia and hence invariants to the scale and rotation. False Acceptance ratio (FAR) and False Rejection ration(FRR) is also computed to test the accuracy of the proposed method. Index Terms- Minutia, FFT, MCC, Euclidean distance 1. INTRODUCTION Recognition of fingerprint is one of the complicated pattern recognition problems which is being studied for the last 40 years. Though various efficient algorithm have been designed for fingerprint matching, it cannot be concluded that this problem has solved. Accuracy, interoperability and computational efficient algorithm are still an open issue [1] in fingerprinting matching. Most of the current fingerprint matching algorithms are based on the minutia. Special ridge pattern are called minutia. Ridge ending and ridge bifurcation are some of the minutia. In the past, minutia matching is considered as the two dimensional pattern matching problem for aligning the two minutia pair. This forced the researcher to find all the possible transformation for two minutia matching. Hough transform is one of the solution of this problem[2][3]. High computational cost and lack of robustness are some the problems of the global minutia matching algorithm. In the last few years, these problem is addressed by introducing the local minutia descriptor and its matching. The characteristics of the local minutia descriptor is such that it is invariants to the global transformation and therefore appropriate for matching without the need of global alignment. Since local minutia descriptor based matching is based on the arrangement of the local property therefore it excludes the global feature and give better results. Global matching algorithms have some of the benefits which is not possible in case of local descriptor based matching and hence an hybrid approach can be applied to get the benefit of both the algorithm. In this types of approach, first of all the local structure is used to match the minutia quickly and robustly. In the next step, matching at global level is performed for validation purpose. The evaluation of the minutia matching based on the local structure passed in three stages. Stage one correspond to the earlier approach in which local structure are formed by considering the minutia lies inside some regions. In this approach no global validation was performed [4][5]. The approach adopted by the [6] and [7] comes in the second stage. [6] and [7] were the first to establishe a relationship between minutia and its neighbourhood in term of invariant angle distance and structure. In this stage, global validation was also performed . Third stage comprises of the method proposed by Jiang and Yau[6] and Ratha [7]. These methods are variants of the method proposed by the same author. In this phase they extend the feature set of the fingerpring by incorporating local ridge, local frequency, shape its. [1] contain the exhaustive review in fingerprint matching and recognition. For further reference reader may go through the [8]-[33]. Two types of local minutia structure were proposed. First one is based on the nearest neighbourhood while the other one is based on the fixed radius. In the nearest neighbourhood [6] approach, centre minutia is characterized by the K- closest minutia which are spatially arranged. This gives the fixed length descriptor which can be matched very efficiently. The later approach is given by the [7]. In this approach, neighbours are defined as the all the neighbours which are inside the radius R of the centre minutia. In this approach, the length of the descriptor is not fixed making it difficult to match the fingerprint locally. But this approach shows better tolerant against missing and spurious minutia.
  • 2. International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 33 2. PROPOSED METHODOLOGY In this paper local descriptor based fingerprint recognition is presented. Minutia cylinder code is used as local descriptor. Block diagram of proposed methodology is shown in figure 2. First of all, with the help of scanner, Fingerprint image of the input finger is registered. Minutia extraction block is used to extract out the minutia from the image. With the help of minutia, a cylindrical code for minutia is formed. Minutia cylindrical code represent minutia local structure of the minutia distribution. Once, minutia cylindrical code is generated for each finger print then a database is created which contain the minutia cylindrical codes of all the fingerprint image. The last step of this method is matching or testing. Next section describe the functioning of last four block of this algorithm. Fig.1 Basic Flow Diagram of Proposed Method 2.1. Minutia Extraction Process In any fingerprint recognition system, Minutia extraction is one of the important phase because the performance of any fingerprint recognition system depends on the how accurately minutia are extracted. In this algorithm following steps have been adopted for minutia extraction. 2.2 FP Image Enhancement First step in any fingerprint recognition system is the enhancement of the fingerprint image. This step is adopted to improve the quality of the fingerprint image so that all the minutia points become more clear and extraction of the minutia points become easier. Generally, Fingerprint image obtained by the fingerprint scanner is noisy and of poor contrast. Ridge and furrows are not visible due to the poor contrast. Image enhancement increase the contrast of ridge and furrows and also joined some of the broken line which occurred in fingerprint image due to sensor deficiency. 2.3 Histogram equalization Histogram equalization basically adjust the intensity distribution of the histogram and with this approach it improve the global contrast of the image. This step helps the lower contrast region to gain a higher contrast while keeping the global contrast intact. This method basically spread out the most frequently occurred intensity value. Figure 3 represent the histogram before and after applying histogram equalization. 2.4 Fast Fourier Transform. This process is accomplished for connecting the ridge broken line . In this process first of all, Fingerprint image is divided in to a block of 32 * 32 pixel and then Fourier transform of this block is computed using following formula. for u = 0, 1, 2, ..., 31 and v = 0, 1, 2, ..., 31. 2.5 Image Binarization This process is performed to convert the 8-bit gary image to 1 bit image. In this, 0 represent the ridge and 1 represent the furrows. This step highlight the ridge and furrows clearly in the image. A local adaptive binarization method is used for this purpose. 2.6 Fingerprint Segmentation This process is accomplished to extract out the region of interest from the fingerprint image. In fingerprint image, some of the area is not useful and carry insignificant ridge and furrows. It is necessary to discard this area . Segmentation process consist of two Finger Finger print Scanner Minutia Extraction Minutia Cylindrical code formation Database Creation Matching Algorithm Fingerprint Registration
  • 3. International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 34 part i.e. Estimation of block direction, extraction of region of interest by applying morphological operation. 2.7 Estimation of Block Diagram Block direction estimation is performed to separate out background and foreground information. In finger print image, foreground is represented by the ridges and furrows while the background represent the no information. In order to compute the block diagram estimation, a block of 16 *16 pixel is chosen and then block direction is computed by using below mentioned formula tan2ß = here gx= Gradient along x-direction gy= Gradient along y-direction Certainity level for each block is then computed using formula- E = Here W= size of block (i.e. 16 in this case) The value of E is compared with the threshold value to decide whether the block comes under the background( if E<threshold) or foreground (i.e if E>threshold). 2.8 Morphological Operation for ROI extraction In the last step morphological operation “opening” and “closing” is applied to to extract the region of interest. 2.8.1 Ridge Thinning This operation is applied to make the ridge width 1 pixel wide. This is performed by thinning operation. 2.8.2Minutia Marking Next step is applied to mark the minutia by applying a 3*3 window. 2.8.3False Minutia Removal Inter-ridge distance is used for eliminating the false minutia points which may affect the over all accuracy of the fingerprint recognition system. 2.8.4Minutia Marking At the final step each true minutia is marked with three element. i.e. x-coordinate, y-coordinate and the minutia direction or angle. 2.9 Minutia Cylindrical Code Information Once the minutia information is obtained then the next step is to form the minutia cylindrical code. In minutia cylindrical code, a cylinder with 6 different layer is used to represent the relative direction and distance among different minutia. One such cylinder is shown in figure 3. Figure 3 represent the cylinder associated with the minutia m={xm, ym, θm}. Radius of the cylinder is R and xm and ym is the center of this cylinder. As shown in the figure, a cuboid encloses the cylinder whose base is aligned as per the direction of the minutia i.e θm. The cuboid is divided in to a cell where The size and height of each cell is × and respectively. Where Fig.2 Minutia Cylinder Each cell of the cylinder can be accessed by three coordinate i.e. i, j, k which represent the position of the cell inside the cylinder. The value of i and j is between 1 and Ns while the k takes the value between
  • 4. International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 35 1 to Nd. The height k of the cylinder represent the angle which is represented by the If Represent the centre of the cell having indices i,j. In MCC, the contribution of each minutia for each cell is computed and is given by Which is computed by the spatial contribution and directional contribution Where spatial contribution is given by And directional contribution is given by And Is the neighbourhood of the . Here is the neighbourhood radius and ds (p, m) represent the Euclidean distance between minutia m and point p as shown in the figure given below- Fig.3 Neighbourhood Radius Once the MCC for all the fingerprint image is obtained then a database of MCC is created. 2.10 Fingerprint Matching In order to match the two fingerprint MCC, Euclidean distance is computed between two fingerprint MCC as given in the below Where = MCC of the testing fingerprint image = MCC of the database. MCC for each fingerprint image is rotation and scale invariants therefore it gives better result in even if the fingerprint image is rotated or scaled. The database MCC for which the value of Ed is less than the certain threshold is considered as the matching one and corresponding fingerprint image is considered as the matching image 3. EXPERIMENTAL RESULTS A database of fingerprint images comprises of the 200 different fingerprint images are created for testing the performance of the proposed method. A simulation program is designed with the help of MATLAB ver 2009B. The simulation program is tested in a core2duo processor computer with 2GB RAM.
  • 5. International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 36 Fig.4 Original Fingerprint Image Figure 5 display the original fingerprint image while figure 6 display the extracted minutia points of the fingerprint image. Fig.5 Extracted Minutia Points in fingerprint image In order to evaluate the performance of this method, statistical measure like FAR(False acceptance Ratio) and FRR(False rejection ratio) is also computed and tabulated in table 1. Fr AR is computed using following formula Table 1 FAR and FRR for different value of Threshold (Without Rotation) Threshold Value FAR(in %) FRR(in %) 6 6.9 0.07 7 7.4 0.05 8 9.2 0.05 9 11.9 0.03 10 13.1 0.02 Table2 FAR and FRR of rotated fingerprint image for different value of Threshold Threshold Value FAR(in %) FRR(in %) 6 6.3 0.12 7 7.6 0.08 8 10.1 0.07 9 12.1 0.03 10 13.9 0.03 Fig.6 Minutia Cylindrical Code of the fingerprint image shown in figure 5 4. CONCLUSION Since from its inception, finger-print matching area has been in the search of some robust technique for matching the fingerprint accurately. The requirement of rotation , scale invariants re the need of the hour. In this paper a MCC based fingerprint matching algorithm is implemented and presented. The performance of the proposed system is satisfactory as clear from the table 1. Even after rotating and scaling the fingerprint, proposed system is able to achieve good accuracy.
  • 6. International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 37 REFERENCES [1] D. Maltoni, D. Maio, A.K. Jain, and S. Prabhakar, Handbook of Fingerprint Recognition. second ed. Springer-Verlag, 2009. [2] N.K. Ratha, K. Karu, S. Chen, and A.K. Jain, “A Real-TimeMatching System for Large Fingerprint Databases,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 18, no. 8, pp. 799-813, Aug. 1996. [3] S.H. Chang, F.H. Cheng, W.H. Hsu, and G.Z. Wu, “Fast Algorithmfor Point Pattern Matching: Invariant to Translations, Rotations and Scale Changes,” Pattern Recognition, vol. 30, pp. 311- 320, 1997. [4] A.K. Hrechak and J.A. McHugh, “Automated Fingerprint Recognition Using Structural Matching,” Pattern Recognition, vol. 23, pp. 893-904, 1990. [5] A.J. Willis and L. Myers, “A Cost-Effective Fingerprint Recognition System for Use with Low-Quality Prints and Damaged Fingertips,” Pattern Recognition, vol. 34, pp. 255-270, 2001. [6] X. Jiang and W.Y. Yau, “Fingerprint Minutiae Matching Based on the Local and Global Structures,” Proc. Int’l Conf. Pattern Recognition, vol. 2, pp. 6038-6041, 2000. [7] N.K. Ratha, V.D. Pandit, R.M. Bolle, and V. Vaish, “Robust Fingerprint Authentication Using Local Structural Similarity,” Proc. IEEE Workshop Applications of Computer Vision, pp. 29-34, 2000. [8] T.Y. Jea and V. Govindaraju, “A Minutia-Based Partial Fingerprint Recognition System,” Pattern Recognition, vol. 38, pp. 1672-1684, 2005. [9] S. Chikkerur, A.N. Cartwright, and V. Govindaraju, “K-plet and Coupled BFS: A Graph Based Fingerprint Representation and Matching Algorithm,” Lecture Notes in Computer Science, vol. 3832, pp. 309-315, Springer, 2006. [10] D. Kwon, I.D. Yun, D.H. Kim, and S.U. Lee, “Fingerprint Matching Method Using Minutiae Clustering and Warping,” Proc. Int’l Conf. Pattern Recognition, vol. 4, 2006. [11] H. Chen, J. Tian, and X. Yang, “Fingerprint Matching with Registration Pattern Inspection,” Lecture Notes in Computer Science, pp. 327- 334, Springer, 2003. [12] J. Feng, “Combining Minutiae Descriptors for Fingerprint Matching,” Pattern Recognition, vol. 41, pp. 342-352, 2008. [13] X. Tan and B. Bhanu, “A Robust Two Step Approach for Fingerprint Identification,” Pattern Recognition Letters, vol. 24, pp. 2127-2134, 2003. [14] G. Parziale and A. Niel, “A Fingerprint Matching Using Minutiae Triangulation,” Lecture Notes in Computer Science, pp. 241-248, Springer, 2004. [15] X. Chen, J. Tian, J. Yang, and Y. Zhang, “An Algorithm for Distorted Fingerprint Matching Based on Local Triangle Feature Set,” IEEE Trans. Information Forensics and Security, vol. 1, no. 2, pp. 169-177, June 2006. [16] D.Q. Zhao, F. Su, and A. Cai, “Fingerprint Registration Using Minutia Clusters and Centroid Structure,” Proc. Int’l Conf. Pattern Recognition, pp. 413-416, 2006. [17] W. Xu, X. Chen, and J. Feng, “A Robust Fingerprint Matching Approach: Growing and Fusing of Local Structures,” Lecture Notes in Computer Science, vol. 4642, p. 134, Springer, 2007. [18] X. Linag, A. Bishnu, and T. Asano, “A Robust Fingerprint Indexing Scheme Using Minutia Neighborhood Structure and Low-Order Delaunay Triangles,” IEEE Trans. Information Forensics and Security, vol. 2, no. 4, pp. 721- 733, Dec. 2007. [19] M. Tico and P. Kuosmanen, “Fingerprint Matching Using an Orientation-Based Minutia Descriptor,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 25, no. 8, pp. 1009- 1014, Aug. 2003. [20] J. Qi and Y. Wang, “A Robust Fingerprint Matching Method,” Pattern Recognition, vol. 38, pp. 1665-1671, 2005. [21] E. Zhu, J. Yin, and G. Zhang, “Fingerprint Matching Based on Global Alignment of Multiple Reference Minutiae,” Pattern Recognition, vol. 38, pp. 1685-1694, 2005. [22] X. Wang, J. Li, and Y. Niu, “Fingerprint Matching Using Orientationcodes and PolyLines,” Pattern Recognition, vol. 40, pp. 3164-3177, 2007. [23] Y. He, J. Tian, L. Li, H. Chen, and X. Yang, “Fingerprint Matching Based on Global Comprehensive Similarity,” IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 28, no. 6, pp. 850-862, June 2006. [24] X. He, J. Tian, L. Li, Y. He, and X. Yang, “Modeling and Analysis of Local Comprehensive Minutia Relation for Fingerprint Matching,” IEEE Trans. Systems, Man, and Cybernetics, Part B, vol. 37, no. 5, pp. 1204- 1211, Oct. 2007. [25] H. Wei, M. Guo, and Z. Ou, “Fingerprint Verification Based on Multistage Minutiae Matching,” Proc. Int’l Conf. Pattern Recognition, pp. 1058-1061, 2006. [26] G.S. Ng, X. Tong, X. Tang, and D. Shi, “Adjacent Orientation Vector Based Fingerprint Minutiae Matching System,” Proc. Int’l Conf. Pattern Recognition, pp. 528-531, 2004. [27] X. Tong, J. Huang, X. Tang, and D. Shi, “Fingerprint Minutiae Matching Using the Adjacent Feature Vector,” Pattern Recognition Letters, vol. 26, pp. 1337-1345, 2005.
  • 7. International Journal of Research in Advent Technology, Vol.3, No.6, June 2015 E-ISSN: 2321-9637 38 [28] L. Sha, F. Zhao, and X. Tang, “Minutiae-Based Fingerprint Matching Using Subset Combination,” Proc. Int’l Conf. Pattern Recognition, pp. 566-569, 2006. [29] J. Feng, Z. Ouyang, and A. Cai, “Fingerprint Matching Using Ridges,” Pattern Recognition, vol. 39, pp. 2131-2140, 2006. [30] Y. Zhang, X. Yang, Q. Su, and J. Tian, “Fingerprint Recognition Based on Combined Features,” Lecture Notes in Computer Science, vol. 4642, p. 281, Springer, 2007. [31] D. Lee, K. Choi, and J. Kim, “A Robust Fingerprint Matching Algorithm Using Local Alignment,” Proc. Int’l Conf. Pattern Recognition, vol. 16, pp. 803-806, 2002. [32] L. Sha and X. Tang, “Orientation-Improved Minutiae for Fingerprint Matching,” Proc. Int’l Conf. Pattern Recognition, vol. 4, pp. 432-435, 2004. [33] Y. Feng, J. Feng, X. Chen, and Z. Song, “A Novel Fingerprint Matching Scheme Based on Local Structure Compatibility,” Proc. Int’l Conf. Pattern Recognition, pp. 374-377, 2006.