2. Contents
Introduction
Literature Survey
Problem Definition
Proposed Methodology
Experimentations
Conclusion and future work
References
3. Introduction
Nowadays, person identification (recognition) and verification is
very important in security and resource access control.
Biometrics is the science of automatic recognition of individual
depending on their physiological and behavioral attributes.
For centuries, handwritten signatures have been an integral part of
validating business transaction contracts and agreements.
Among the different forms of biometric recognition systems such as
fingerprint, iris, face, voice, palm etc., signature will be most widely
used.
4. Signature Recognition
Signature Recognition is the procedure of determining to whom a
particular signature belongs to.
Depending on acquiring of signature images, there are two types of
signature recognition systems:
Online Signature Recognition
Offline Signature Recognition
5. Literature Survey
1. Offline Handwritten Signature Recognition(Gulzar A.
Khuwaja and Mohammad S. Laghari)
Biometrics, which refers to identifying an individual based on
his or her physiological or behavioral characteristics, has the
capability to reliably distinguish between an authorized person
and an imposter.
This paper presents a neural network based recognition of
offline handwritten signatures system that is trained with low-
resolution scanned signature images.
6. 2. Off-line Signature Verification Based on Fusion of Grid
and Global Features Using Neural Networks(Shashi Kumar
D R and K B Raja)
In this paper Off-line Signature Verification Based on Fusion of
Grid and Global Features using Neural Network(SVFGNN) is
presented.
The global and grid features are fused to generate set of
features for the verification of signature.
7. 3. DWT based Off-line Signature Verification using Angular
Features (Prashanth C R )
This papers presents DWT based Off-line Signature
Verification using Angular Features (DOSVAF).
The signature is resized and Discrete Wavelet Transform
(DWT) is applied on the blocks to extract the features.
8. 4. Off-Line Signature Recognition Systems(V A Bharadi)
Handwritten signature is one of the most widely used biometric
traits for authentication of person as well as document.
In this paper we discuss issues regarding off-line signature
recognitions.
The performance metrics of typical systems are compared
along with their feature extraction mechanisms.
9. 5. Offline Signature Recognition and Verification Based on
Artificial Neural Network(Mohammed A. Abdala)
In this paper, a problem for Offline Signature Recognition and
Verification is presented.
A system is designed based on two neural networks classifier
and two powerful features (global and grid features).
The designed system consist of three stages which is pre-
processing, feature extraction and neural network stage.
10. 6. Signature Recognition & Verification System Using Back
Propagation Neural Network (Nilesh Y. Choudhary, Dr.
Umesh. Bhadade)
In this paper, off-line signature recognition & verification using
back propagation neural network is proposed which is based
on steps of image processing, invariant central moment &
some global properties and back propagation neural
networks.
11. Problem Definition
Signature Recognition is the procedure of determining to whom a
particular signature belongs to. In this work, the global and grid
features are combined and used to differentiate among the signature
images. These combined features are given to Back Propagation
Neural Network(BPNN) to train it, so that particular signature image
is recognized.
13. Image Acquisition :
Collection of signatures from 50 persons on blank paper.
The collected signatures are scanned to get images in JPG
format to create database.
14. Pre-Processing :
Image pre-processing is a technique to enhance raw images
received from cameras/sensors placed on satellites, space
probes and aircrafts or pictures taken in normal day-to-day life
for various applications.
The techniques for preprocessing used are
RGB to Gray Scale Conversion
Binarization
Thinning
Bounding Box
17. Feature Extraction
Features are the characters to be extracted from the processed
image.
It has used two feature techniques
Global Features
Grid Features
18. Global Features
Height :
Width :
Number of Black Pixels :
Centroid of the signature :
Width
Height
19. Grid Features
The cropped image is divided into 9 rectangular segments i.e.
(3 Χ 3) blocks.
3*3 Blocks of Grid Image
20. DWT(Discrete Wavelet Transform)
DWT applied on 1st block. Each block contributes
horizontal, vertical and diagonal components.
1st Block Horizontal Vertical Diagonal
21. After applying DWT to all 9 blocks, each block is divided into
horizontal, vertical and diagonal components. From each
components two features mainly horizontal and vertical
projection positions are extracted. Total 54 (9 x 3 x 2) features
are extracted.
Grid features extracted from each block are
Horizontal Projection Position:
Vertical Projection Position:
25. Training of BPNN
This involves developing a suitable neural network model
(BPNN). Then the extracted features are presented to
BPNN, which recognizes the different types of signature images.
The training takes place such that the neural network learns that
each entry in the input file has a corresponding entry in the output
file.
27. Algorithm for Training phase
Description: Retrieval of a signature image from a database
Input: Training sample images.
Output: Construction of Back Propagation Neural Network.
Begin
Read the training samples images
Step1: Pre-processing
Convert the image into gray scale image.
Convert the gray scale image into binary image.
Apply thinning process.
Apply bounding box.
28. Step 2: Features Extracted.
Step 3: Back propagation neural network training.
end // end of proposed algorithm
29. Testing using Trained BPNN
In testing, input image from testing set is selected and its
features are extracted and given them to the trained model, the
trained BPNN model classifies given sample and produces
output as type of signature and corresponding pattern
Classification accuracy= Number of recognized signatures
Total number of testing signatures
31. Experimental Results
Experiment 1
The features extracted are listed as:
Height of the signature
Width of the signature
Centroid of X-axis and Y-axis
Number of black pixels of the signature
The image is divided into 9 blocks and DWT is applied to each
block. Energy values of each block were extracted as a feature.
33. Experiment 2
The features are extracted as listed below:
Height of the signature
Width of the signature
Centroid of X-axis and Y-axis
Number of black pixels of the signature
The image is divided into 9 blocks and DWT is applied to each
block. From each block two features, horizontal and vertical
projection positions of horizontal, vertical and diagonal
components are extracted
35. No. of Persons Experiment 1 Experiment 2
10 83.46 % 93.33 %
20 81 % 92.91 %
30 78.28 % 91.38 %
40 76.7 % 90 %
50 74 % 89.47 %
Performance Rate
The performance rate of the two experiments
36. Conclusion
The objective of signature recognition is to recognize the signer
for the purpose of recognition. It has been observed that the
global and grid features extracted using discrete wavelet
transform are found to be efficient for offline signature
recognition. The combination of discrete wavelet transform and
back propagation neural network has given expected results. It
achieved the accuracy rate ranging from 93%-89% for enrollment
of 10 to 50 persons.
37. Future Work
The signature recognition can also be changed by changing the
features that can be extracted from a signature. So, the future
work of the recognition of signature can be done with the same
Neural Network methods but using different signature features
and compares the results with results of the present project.
38. References
Gulzar A. Khuwaja and Mohammad S. Laghari, World Academy of
Science, Engineering and Technology , “Offline Handwritten Signature
Recognition”, 2011
Shashi Kumar D R, K B Raja, R. K Chhotaray, Sabyasachi Pattanaik, “Off-line
Signature Verification Based on Fusion of Grid and Global Features Using
Neural Networks”, 2010
Prashanth C R , K B Raja, Venugopal K R, L M Patnaik, “DWT based Off-line
Signature Verification using Angular Features”, 2012
V A Bharadi, H B Kekre, “Off-Line Signature Recognition Systems”, 2010
Mohammed A. Abdala & Noor Ayad Yousif, “Offline Signature Recognition and
Verification Based on Artificial Neural Network”, 2008
H. Baltzakis, N. Papamarkos, “A New Signature Verification Technique Based On
A Two-Stage Neural Network Classifier”, 2001
Khamael Abbas Al-Dulaimi, “Handwritten Signature Verification Technique
based on Extract Features”, 2011
39. Hemanta Saikia, Kanak Chandra Sarma, “Approaches and Issues in Offline
Signature Verification System”, 2012
Vu Nguyen, Michael Blumenstein, Graham Leedham, “Global Features for the
Off-Line Signature Verification Problem”, 2009
Meenakshi S Arya, Vandana S Inamdar, “A Preliminary Study on Various Off-line
Hand Written Signature Verification Approaches”, 2010
Javed Ahmed Mahar, Prof. Dr. Mumtaz Hussain Mahar, Muhammad Khalid
Khan, “Comparative Study of Feature Extraction Methods with K-NN for Off-
Line Signature Verification”, 2006
Nilesh Y. Choudhary, Mrs. Rupal Patil, Dr. Umesh. Bhadade, Prof. Bhupendra M
Chaudhari,“Signature Recognition & Verification System Using Back
Propagation Neural Network”, 2013
Manoj Kumar, “Signature Verification Using Neural Network”, 2012
Paigwar Shikha and Shukla Shailja,“Neural Network Based Offline Signature
Recognition and Verification System”, 2013
Srikanta Pal, Michael Blumenstein, Umapada Pal, “Off-Line Signature
Verification Systems: A Survey”, 2011