Image Processing Based Signature Recognition and Verification Technique Using Artificial Neural Network approach
11 likes•5,152 views
The document presents a dissertation by Priyanka Pradhan focused on developing an offline signature verification system utilizing artificial neural networks. It outlines the need for automatic signature verification due to the impracticality of manual checks for numerous documents and discusses various types of forgery, objectives, research methodologies, and expected results targeting over 98% accuracy. The proposed model captures signature features, processes them through image pre-processing, segmentation, and utilizes neural networks for classification.
![3. Literature Review
9/12/2015Priyanka Pradhan
9
Base Paper
[1]Ali Karounia , Bassam Dayab, Samia Bahlakb,” Offline
signature recognition using neural networks approach”. 1877-0509,
Published by Elsevier Ltd(Dec. 2011)[8].
In this paper, a method for Offline Verification of signatures is presented using
a set of simple shape based geometric features. The features that are used are
Area, Center of gravity, Eccentricity, Kurtosis and Skewness. Before
extracting the features, preprocessing of a scanned image is necessary to
isolate the signature part and to remove any spurious noise present.
The system is initially trained using a database of signatures obtained from
those individuals whose signatures have to be authenticated by the system.](https://image.slidesharecdn.com/presentation2-150912052059-lva1-app6891/85/Image-Processing-Based-Signature-Recognition-and-Verification-Technique-Using-Artificial-Neural-Network-approach-9-320.jpg)
![Contd…
9/12/2015Priyanka Pradhan
10
[2]Nancy Prof. and Gulshan Goyal ,“Signature Processing in
Handwritten Bank Cheque Images”. International Journal on
Recent and Innovation Trends in Computing and Communication
ISSN: 2321-8169 Volume: 2 Issue: 5(May 2015)[2].
The present paper focuses on different steps including browsing a bank
cheque, pre-processing, feature extraction, recognition.
Preprocessing stage includes image resizing, noise elimination, thinning etc.
On the other hand feature extraction is done on the basis of gray level co-
occurrence matrix .
Feature extraction stage includes contrast, homogeneity, energy, entropy,
variance, sum average etc.](https://image.slidesharecdn.com/presentation2-150912052059-lva1-app6891/85/Image-Processing-Based-Signature-Recognition-and-Verification-Technique-Using-Artificial-Neural-Network-approach-10-320.jpg)
![Contd…
9/12/2015Priyanka Pradhan
11
[3] Harpreet Anand and Prof. D.L Bhombe,“Enhanced Signature
Verification AndRecognition Using Matlab”, International Journal
of Innovative Research in Advanced Engineering (IJIRAE) ISSN:
2349-2163Volume 1 Issue 4 (May 2014)[3].
In this paper offline signature verification using neural network is projected.
For authentication of signature, the proposed method is based on geometrical
and statistical feature extraction and then the entire database, features are
trained using neural network .
The extracted features of investigation signature are compared with the
previously trained features of the reference signature.](https://image.slidesharecdn.com/presentation2-150912052059-lva1-app6891/85/Image-Processing-Based-Signature-Recognition-and-Verification-Technique-Using-Artificial-Neural-Network-approach-11-320.jpg)
![Contd…
9/12/2015Priyanka Pradhan
12
[4]A. Vinoth, V. Sujathabai, “A Pixel Based Signature
Authentication System”, International Journal of Innovative
Technology and Exploring Engineering (IJITEE) ISSN: 2278-3075,
Volume-2, Issue-6, May 2013[6] .
In this paper, the off-line signatures is verified by taking a boundary of the
entire signature and do the pixel comparison.
Signature is acquired using a scanner. Detection process is done after the data
acquisition and pre-processing. Pre processing includes noise removal, grey-
scale, manipulation, edge detection. Experimental results show that 50% of
the accurate matching with the existing one from the data base.](https://image.slidesharecdn.com/presentation2-150912052059-lva1-app6891/85/Image-Processing-Based-Signature-Recognition-and-Verification-Technique-Using-Artificial-Neural-Network-approach-12-320.jpg)
![Contd…
9/12/2015Priyanka Pradhan
13
[5] Przemysław Kudłacik and Piotr Porwik, A new approach to
signature recognition using the fuzzy method. Published online:
The Author(s) 2012. This article is published with open access at
Springerlink.com (15 August 2012)[7].
The paper presents a new fuzzy approach to off-line handwritten signature
recognition. The solution is based on characteristic feature extraction. After
finding signature’s center of gravity a number of lines are drawn through it at
different angles.
Cross points of generated lines and signature sample, which are further
grouped and sorted, are treated as the set of features.](https://image.slidesharecdn.com/presentation2-150912052059-lva1-app6891/85/Image-Processing-Based-Signature-Recognition-and-Verification-Technique-Using-Artificial-Neural-Network-approach-13-320.jpg)
![8. References
[1] Shefali Singla and Deepinder Kaur. Signature Verification Using DTI and Guided DTI
Classifiers and Digital Encryption. International Journal of Advanced Research in
Computer Science and Software Engineering. Volume 5 Issue 4 ISSN: 2277 128X (May
2015).
[2] Nancy Prof. and Gulshan Goyal . Signature Processing in Handwritten Bank Cheque
Images. International Journal on Recent and Innovation Trends in Computing and
Communication. ISSN: 2321-8169 Volume: 2 Issue: 5(May 2015).
[3] Harpreet Anand and Prof. D.L Bhombe. Enhanced Signature Verification
AndRecognition Using Matlab. International Journal of Innovative Research in Advanced
Engineering (IJIRAE). ISSN: 2349-2163 Volume 1 Issue: 4 (May 2014)
[4] Sheena, Sheena Mathew. A Study of Multimodal Biometric System. IJRET:
International Journal of Research in Engineering and Technology eISSN: 2319-1163 |
pISSN: 2321-7308. Volume: 03 Special Issue: 15 | IWCPS-2014 | (Dec-2014).
9/12/2015Priyanka Pradhan
20](https://image.slidesharecdn.com/presentation2-150912052059-lva1-app6891/85/Image-Processing-Based-Signature-Recognition-and-Verification-Technique-Using-Artificial-Neural-Network-approach-20-320.jpg)
![Contd…
[5] Harpreet Anand and Prof. D.L Bhombe. Enhanced Signature Verification And
Recognition Using Matlab. International Journal of Innovative Dissertation in Advanced
Engineering (IJIRAE) ISSN: 2349-2163 Volume 1 Issue 4 (May 2014).
[6] A. Vinoth, V. Sujathabai. A Pixel Based Signature Authentication System.
International Journal of Innovative Technology and Exploring Engineering (IJITEE)
ISSN: 2278-3075 Volume-2 Issue-6 (May 2013).
[7] Przemysław Kudłacik and Piotr Porwik, A new approach to signature recognition
using the fuzzy method. Published online: This article is published with open access at
Springerlink.com (15 August 2012).
[8] Ali Karounia , Bassam Dayab, Samia Bahlakb. Offline signature recognition using
neural networks approach. 1877-0509 Published by Elsevier Ltd(Dec. 2011).
9/12/2015Priyanka Pradhan
21](https://image.slidesharecdn.com/presentation2-150912052059-lva1-app6891/85/Image-Processing-Based-Signature-Recognition-and-Verification-Technique-Using-Artificial-Neural-Network-approach-21-320.jpg)
Image Processing Based Signature Recognition and Verification Technique Using Artificial Neural Network approach
- 1. P R E S E N T E D B Y: P R I YA N K A P R A D H A N M . T E C H ( S E ) R O L L N O . - 1 3 0 1 4 0 9 5 0 7 Image Processing Based Signature Recognition and Verification Technique Using Artificial Neural Network approach UNDER THE GUIDANCE OF: ER. L. S. MAURYA HOD(CS/IT) SRMSCET, BAREILLY.
- 2. Outline 1. Introduction 1.1 Signature Verification vs. Signature Recognition 1.2 Types of Signature Forgery 2. Problem statement/objective 3. Literature Review 4. Research Methodology 5. Proposed Model 6. Software and Tools Used 7. Expected Result 8. References 9/12/2015Priyanka Pradhan 2
- 3. Signature has been a distinguishing feature for person identification. When a large number of documents, e.g. bank cheques, have to be authenticated in a limited time, the manual verification of account holders’ signatures is often unrealistic. Signature provides secure means of authentication and authorization. So, there is a need of Automatic Signature Verification and Identification systems. The present dissertation work is done in the field of offline signature verification system by extracting some special features that make a signature difficult to forge. In this dissertation work, existing signature verification systems have been thoroughly studied and a model is designed to develop an offline signature verification system. 1. Introduction 9/12/2015Priyanka Pradhan 3
- 4. 1.1 Signature Verification vs. Signature Recognition Shalini Davies Signature Verification Signature Recognition 9/12/2015Priyanka Pradhan 4
- 6. 1.2 Types of Signature Forgery 9/12/2015Priyanka Pradhan 6 Forgeries can be classified into three main categories. Random forgery: which is written by the person who doesn’t know the shape of original signature. Simple forgery: which is represented by a signature sample, written by the person who knows the shape of original signature without much practice. Skilled forgery: represented by a suitable imitation of the genuine signature model (a) Original (b)Random forgery (c)Skilled forgery
- 7. Contd…. 9/12/2015Priyanka Pradhan 7 FAR (False Acceptance Ratio): It is given by the number of fake signatures accepted by the system with respect to the total number of comparisons made. Calculation of these is show below. Number of forgeries accepted FAR =-------------------------------------- * 100 Number of forgeries tested FRR (False Rejection Ratio): It is the total number of genuine signatures rejected by the system with respect to the total number of comparisons made. Number of originals rejected FRR = ------------------------------------- * 100 Number of originals tested
- 8. 2. Problem Statement/Objective The objectives of this dissertation are: To make sure that only the right people are authorized to access high-security systems The process of signature verification should be able to detect forgeries To use cascading of features for the process of feature extraction of signature from the preprocessed scanned image of a signature that will give more accurate results To cascade and comparison of features 9/12/2015Priyanka Pradhan 8
- 9. 3. Literature Review 9/12/2015Priyanka Pradhan 9 Base Paper [1]Ali Karounia , Bassam Dayab, Samia Bahlakb,” Offline signature recognition using neural networks approach”. 1877-0509, Published by Elsevier Ltd(Dec. 2011)[8]. In this paper, a method for Offline Verification of signatures is presented using a set of simple shape based geometric features. The features that are used are Area, Center of gravity, Eccentricity, Kurtosis and Skewness. Before extracting the features, preprocessing of a scanned image is necessary to isolate the signature part and to remove any spurious noise present. The system is initially trained using a database of signatures obtained from those individuals whose signatures have to be authenticated by the system.
- 10. Contd… 9/12/2015Priyanka Pradhan 10 [2]Nancy Prof. and Gulshan Goyal ,“Signature Processing in Handwritten Bank Cheque Images”. International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169 Volume: 2 Issue: 5(May 2015)[2]. The present paper focuses on different steps including browsing a bank cheque, pre-processing, feature extraction, recognition. Preprocessing stage includes image resizing, noise elimination, thinning etc. On the other hand feature extraction is done on the basis of gray level co- occurrence matrix . Feature extraction stage includes contrast, homogeneity, energy, entropy, variance, sum average etc.
- 11. Contd… 9/12/2015Priyanka Pradhan 11 [3] Harpreet Anand and Prof. D.L Bhombe,“Enhanced Signature Verification AndRecognition Using Matlab”, International Journal of Innovative Research in Advanced Engineering (IJIRAE) ISSN: 2349-2163Volume 1 Issue 4 (May 2014)[3]. In this paper offline signature verification using neural network is projected. For authentication of signature, the proposed method is based on geometrical and statistical feature extraction and then the entire database, features are trained using neural network . The extracted features of investigation signature are compared with the previously trained features of the reference signature.
- 12. Contd… 9/12/2015Priyanka Pradhan 12 [4]A. Vinoth, V. Sujathabai, “A Pixel Based Signature Authentication System”, International Journal of Innovative Technology and Exploring Engineering (IJITEE) ISSN: 2278-3075, Volume-2, Issue-6, May 2013[6] . In this paper, the off-line signatures is verified by taking a boundary of the entire signature and do the pixel comparison. Signature is acquired using a scanner. Detection process is done after the data acquisition and pre-processing. Pre processing includes noise removal, grey- scale, manipulation, edge detection. Experimental results show that 50% of the accurate matching with the existing one from the data base.
- 13. Contd… 9/12/2015Priyanka Pradhan 13 [5] Przemysław Kudłacik and Piotr Porwik, A new approach to signature recognition using the fuzzy method. Published online: The Author(s) 2012. This article is published with open access at Springerlink.com (15 August 2012)[7]. The paper presents a new fuzzy approach to off-line handwritten signature recognition. The solution is based on characteristic feature extraction. After finding signature’s center of gravity a number of lines are drawn through it at different angles. Cross points of generated lines and signature sample, which are further grouped and sorted, are treated as the set of features.
- 14. 4. Research Methodology Stage 1: The signature acquisition process is executed manually using scanner. Stage 2: The Pre-processing phase. After signature acquisition the sampled data is preceded for the Pre-processing phase. In pre-processing phase the images are refined by applying various operations of Digital Image Processing. Pre-Processing phase will filter the images and convert the RGB image into the gray image and then to Black & White image. The threshold Black & White image will execute the Recognition process in an efficient way. Various Pre-Processing phases will be applying on the sampled data. (b)Preprocessed Signature 9/12/2015Priyanka Pradhan 14 (a)Original Gray Scale Image
- 15. Contd… 9/12/2015Priyanka Pradhan 15 Stage 3: After pre-processing the sampled scanned signature documents, the segmentation techniques are applied to it. Various methods for segmentation are available, some of them are: Threshold based segmentation Edge based segmentation Region based segmentation Clustering techniques Matching Stage 4: The segmented images are then used for the feature extraction phase. Various methods for feature extraction are available, some of them are: Global Transformation and Series Expansion Statistical Method Geometrical and Topological Representation
- 16. Contd… 9/12/2015Priyanka Pradhan 16 Stage 5: At final stage, the ANN (ARTIFICIAL NEURAL NETWORK) will be used for the classification process. Various methods for classification are available, some of them are: Hidden Markovs Model Support Vector Machine Artificial Neural Network (a)Supervised Learning (b)Unsupervised learning Fig 4.1: Neural network system
- 17. 5. Proposed Approach 9/12/2015Priyanka Pradhan 17 Fig 5.1:Proposed Model
- 18. 6. Software and Tools Used The sampled signatures are scanned through a normal optical scanner. For morphological operations image processing tool in MATLAB is used. For training the neural network, a toolbox called nntool in MATLAB will be used. 9/12/2015Priyanka Pradhan 18
- 19. 7. Expected Results The motive of this dissertation is to design an offline signature recognition system. The motive is to verify offline scanned signature. The accuracy of signature recognition is targeted to > 98%. 9/12/2015Priyanka Pradhan 19
- 20. 8. References [1] Shefali Singla and Deepinder Kaur. Signature Verification Using DTI and Guided DTI Classifiers and Digital Encryption. International Journal of Advanced Research in Computer Science and Software Engineering. Volume 5 Issue 4 ISSN: 2277 128X (May 2015). [2] Nancy Prof. and Gulshan Goyal . Signature Processing in Handwritten Bank Cheque Images. International Journal on Recent and Innovation Trends in Computing and Communication. ISSN: 2321-8169 Volume: 2 Issue: 5(May 2015). [3] Harpreet Anand and Prof. D.L Bhombe. Enhanced Signature Verification AndRecognition Using Matlab. International Journal of Innovative Research in Advanced Engineering (IJIRAE). ISSN: 2349-2163 Volume 1 Issue: 4 (May 2014) [4] Sheena, Sheena Mathew. A Study of Multimodal Biometric System. IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308. Volume: 03 Special Issue: 15 | IWCPS-2014 | (Dec-2014). 9/12/2015Priyanka Pradhan 20
- 21. Contd… [5] Harpreet Anand and Prof. D.L Bhombe. Enhanced Signature Verification And Recognition Using Matlab. International Journal of Innovative Dissertation in Advanced Engineering (IJIRAE) ISSN: 2349-2163 Volume 1 Issue 4 (May 2014). [6] A. Vinoth, V. Sujathabai. A Pixel Based Signature Authentication System. International Journal of Innovative Technology and Exploring Engineering (IJITEE) ISSN: 2278-3075 Volume-2 Issue-6 (May 2013). [7] Przemysław Kudłacik and Piotr Porwik, A new approach to signature recognition using the fuzzy method. Published online: This article is published with open access at Springerlink.com (15 August 2012). [8] Ali Karounia , Bassam Dayab, Samia Bahlakb. Offline signature recognition using neural networks approach. 1877-0509 Published by Elsevier Ltd(Dec. 2011). 9/12/2015Priyanka Pradhan 21
Editor's Notes
- #19: the classifiers are used for mapping the features of training set to a group of feature vector of training set. There are various approaches which can be used for the classification phase