Online Handwritten Signature Research Articles

Signature Verification SaaS Implementation on Microsoft Azure Cloud

Signature recognition is one of the important behavioral biometric trait. Signatures recognition systems can be used to identify precisely user identity by making use of signature information such as x, y variations and pressure from a tablet PC. This makes way for using dynamic, i.e., online handwritten signature based biometric system is more accurate than the static ones, hence can be useful for signature verification applications. In this paper new set of features are proposed for online or dynamic signature recognition. In this research, feature vector and their extraction mechanism is implemented using Webber Local Descriptor (WLD). Thus, helping signature verification applications to detect forgery of signatures. The performance of proposed feature vector is further improved by soft biometric traits of the signature.

On-line Signature Verification Based on GA-SVM

With the development of pen-based mobile device, on-line signature verification is gradually becoming a kind of important biometrics verification. This thesis proposes a method of verification of on-line handwritten signatures using both Support Vector Data Description (SVM) and Genetic Algorithm (GA). A 27-parameter feature set including shape and dynamic features is extracted from the on-line signatures data. The genuine signatures of each subject are treated as target data to train the SVM classifier. As a kernel based one-class classifier, SVM can accurately describe the feature distribution of the genuine signatures and detect the forgeries. To improving the performance of the authentication method, genetic algorithm (GA) is used to optimise classifier parameters and feature subset selection. Signature data form the SVC2013 database is used to carry out verification experiments. The proposed method can achieve an average Equal Error Rate (EER) of 4.93% of the skill forgery database.

Online Handwritten Signature Recognition by Length Normalization using Up-Sampling and Down-Sampling

With the rapid advancement of capture devices like tablet or smart phone, there is a huge potential for online signature applications that are expected to occupy a large field of researches in forthcoming years. Online handwritten signature encounters difficulty in the verification process because an individual rarely produce exactly the same signature whenever he signs. This difference in the produced signature is referred to as intra-user variability. Verification difficulty occurs especially in the case where the feature extraction and classification algorithms are designed to classify a stable length vector of input features. In this paper, we introduce an efficient algorithm for online signature length normalization by using Up-Sampling and Down-Sampling techniques. Furthermore, online signature verification system is also proposed by using both Principal Component Analysis (PCA) for feature extraction and Artificial Neural Network (ANN) for classification. The SIGMA database, which has more than 6,000 genuine and 2,000 forged signature samples taken from 200 individuals, is used to evaluate the effectiveness of the proposed technique. Based on the tests performed, the proposed technique managed to achieve False Accept Rate (FAR) of 5.5% and False Reject Rate (FRR) of 8.75%.

Non-Invertible Online Signature Biometric Template Protection via Shuffling and Trigonometry Transformation

This paper describes a novel approach to a cancelable template protection scheme that secures online handwritten signature samples in the reference database of a biometric verification system. We propose a renewable-noninvertible transformation process named Bio-Trigono comprising two consecutive steps. First, a shuffling scheme is applied to a signature sample to attain the renewability property for template protection. This is followed by the deployment of a cosine function for which its periodic characteristic is exploited to achieve a much desired non-invertible property for additional security. The overall template protection scheme was tested rigorously on signature samples of a SIGMA database through an online signature verification system. Its verification utilized the Principal Component Analysis (PCA) for features' extraction and the Artificial Neural Network (ANN) for user reference modeling and classification processes. Results demonstrated an effective cancelable template protection scheme whereby the best averaged error rates were 10. 3, 10. 5 and 14. 1% for un-transformed first and second transformed signature templates, respectively.

Based on EADTW On-Line Handwriting Signature Handwriting Signature Verification System Design and Implementation

In view of poor accuracy and slow calculation of the traditional on-line handwriting signature verification, an on-line handwriting signature verification based on Early Abandon Dynamic Time Warping (EADTW) was designed and implemented after numerous researched. The training template followed the mechanism of benchmark signature, while the certification part adopted EADTW algorithm. The experimental results showed that compared with on-line handwritten signature system based on DTW (dynamic time warping), this new system not only greatly reduced cumbersome and repeated calculation, but also obviously improved the accuracy, The bigger the test sample is, the more obvious the advantage is.

Online handwritten signature verification using neural network classifier based on principal component analysis.

One of the main difficulties in designing online signature verification (OSV) system is to find the most distinctive features with high discriminating capabilities for the verification, particularly, with regard to the high variability which is inherent in genuine handwritten signatures, coupled with the possibility of skilled forgeries having close resemblance to the original counterparts. In this paper, we proposed a systematic approach to online signature verification through the use of multilayer perceptron (MLP) on a subset of principal component analysis (PCA) features. The proposed approach illustrates a feature selection technique on the usually discarded information from PCA computation, which can be significant in attaining reduced error rates. The experiment is performed using 4000 signature samples from SIGMA database, which yielded a false acceptance rate (FAR) of 7.4% and a false rejection rate (FRR) of 6.4%.

A Proposal for Classification of Multisensor Time Series Data based on Time Delay Embedding

Abstract Multisensor time series data is common in many ap- plications of process industry, medical and health care, biometrics etc.Analysis of multisensor time series data requires analysis of multidimensional time series(MTS) which is challenging as they constitute a huge volume of data of dynamic nature. Traditional machine learning algorithms for classification and clustering developed for static data can not be applied directly to MTS data. Various techniques have been developed to represent MTS data in a suitable manner for analysis by popular machine learning algorithms. Though a plethora of different approaches have been developed so far, 1NN classifier based on dynamic time warping (DTW) has been found to be the most popular due to its simplicity. In this work, an approach for time series classification is proposed based on multidimensional delay vector representation of time series. Multivariate time series is considered here as a group of single time series and each time series is processed separately to be represented by a multidimensional delay vector (MDV). A simple simulation experiment with online handwritten signature data has been done with a similarity measure based on the MDV representation and classification performance is compared with DTW based classifier. The simulation results show that classification accuracy of the proposed approach is satisfactory while computational cost is lower than DTW method.

Detection of alcohol intoxication via online handwritten signature verification

The Internet has recently facilitated various methods of personal authentication based on biometrics. Personal authentication techniques based on human physical features include fingerprint matching, vein matching, facial recognition, speech recognition, and signature recognition. In this study, we investigate the detection of alcohol intoxication on the basis of handwritten signatures. As in the case of previous related studies, we found that the signature attestation rate varies at an individual level according to sex, age, acetaldehyde removal efficiency, and individual constitution.In this study, we employed 30 people to evaluate the change in a handwritten signature before and after alcoholic intake. First, we measured the signature verification rate using the online signature verification system. The signature verification rate measured using the WACOM Tablet pen before alcohol consumption was 97.0%. Moreover, the size of the characters and the interval between the characters were measured for signatures collected after alcohol consumption. We detected the level of alcohol intoxication on the basis of the total time taken for writing the signature, the average pressure of the brush, the two-dimensional writing speed, and internal angle of stroke turns. The maximum alcohol detection rate of this method was 95.1%, which was achieved when the examinees were tested 35min after alcohol consumption. The rate of alcohol detection increases with the alcohol density in an examinee’s breath.

Efficient On-line Handwritten Signature Verification by Integrating DTW and Proposed Local Dynamic Similarity Measure

Efficient On-line Handwritten Signature Verification by Integrating DTW and Proposed Local Dynamic Similarity Measure

Online handwritten signature verification system based on DWT features extraction and neural network classification

Handwritten signature is the most widely accepted biometric to identity verification. The target of research is to present online handwritten signature verification system based on discrete wavelet transform (DWT) features extraction and feed forward back propagation error neural network recognition. Steps for verifying online handwritten signature in this system start with extracting pen position data ( x and y positions) of points that forming the signature. Pen-movement angles are then derived from pen position data. To reduce variations in pen-position and pen-movement angles dimensionality, data are normalized and resampled. To enhance the difference between a genuine signature and its forgery, the signature is verified in DWT domain. Low frequency sub-band signals (approximations) of pen-position parameter and pen-movement angle parameter are considered as intrapersonal features. These are used for suppressing variations between different genuine signatures and enhancing the interpersonal variations, hence are given higher scores within total recognition process. Both of pen-position and pen-movement angle features are then associated for obtaining a decision about online handwritten signature verification. A multi-matcher consists of six neural networks which use multiple representations and matching for the same input biometric signal is used to verify signature. The recognition rate for each of these neural network recognizers is discussed and a comparison of those rates is performed. Experiments are carried on signature database for five users each of 20 genuine and 20 skilled forgery signatures. Recognition success rate for genuine signatures is 95%.

The algorithm of online handwritten signature verification based on improved DTW

Online handwritten signature verification is widely used because of its celerity and convenience. It is the personal verification based on biometrics technology. It is the verification using information about the movement during writing signature, so it becomes an accepted method of personal verification. Basing on this thesis, an online handwritten signature verification algorithm using improved DTW is presented. The weak point of classic DTW is that only the data points by warping the x-axis are considered but the differences on y-axis of two series are not considered. The two situations above are considered in the improved DTW method in order to prevent this problem and the distance measure of DTW is modified based on improved DTW in this paper particularly. The result of experiment proves that this algorithm obtains a good verification effect.

Using position extrema points to capture shape in on-line handwritten signature verification

There is considerable interest in authentication based on handwritten signature verification (HSV) because of the long-standing tradition of its use in many common authentication tasks. HSV may be considered superior to many other biometric authentication techniques, for example fingerprints or retinal patterns, which are more reliable but also more intrusive. Furthermore, they require special and relatively expensive hardware to capture the image. The present paper is an attempt to develop a reliable HSV technique by capturing the shape of the signature using the position extrema points of a signature. The technique presented essentially captures the directions of pen motion during the writing of the signature and this is represented in a simple way by a string. The technique is evaluated and shown to be promising.

UNIVERSAL WRITING MODEL FOR RECOVERY OF WRITING SEQUENCE OF STATIC HANDWRITING IMAGES

Online features have been proven to be more robust information for handwriting recognition than an offline static image due to dynamic aspects, such as the writing sequence of strokes. The estimation of temporal information from a static image becomes an important issue. This paper presents a new statistical method to reconstruct the writing order of a handwritten signature from a two-dimensional static image. The reconstruction process consists of two phases, namely the training phase and the testing phase. In the training phase, the writing order with other attributes, such as length and direction, are extracted and analyzed from a set of training online handwritten signatures. A Universal Writing Model (UWM), which consists of a set of distribution functions, is then constructed. In the testing phase, the UWM is applied to reconstruct the writing order of an offline signature. 300 offline signatures with ground truth are used for evaluation. Experimental results show that about one-eighth of the reconstructed writing sequences are the same as the actual writing sequences.

Identity authentication using improved online signature verification method

We present a system for online handwritten signature verification, approaching the problem as a two-class pattern recognition problem. A test signature’s authenticity is established by first aligning it with each reference signature for the claimed user, using dynamic time warping. The distances of the test signature to the nearest, farthest and template reference signatures are normalized by the corresponding mean values obtained from the reference set, to form a three-dimensional feature vector. This feature vector is then classified into one of the two classes (genuine or forgery). A linear classifier used in conjunction with the principal component analysis obtained a 1.4% error rate for a data set of 94 people and 619 test signatures (genuine signatures and skilled forgeries). Our method received the first place at SVC2004 with a 2.8% error rate.

On-line signature verification

We describe a method for on-line handwritten signature verification. The signatures are acquired using a digitizing tablet which captures both dynamic and spatial information of the writing. After preprocessing the signature, several features are extracted. The authenticity of a writer is determined by comparing an input signature to a stored reference set (template) consisting of three signatures. The similarity between an input signature and the reference set is computed using string matching and the similarity value is compared to a threshold. Several approaches for obtaining the optimal threshold value from the reference set are investigated. The best result yields a false reject rate of 2.8% and a false accept rate of 1.6%. Experiments on a database containing a total of 1232 signatures of 102 individuals show that writer-dependent thresholds yield better results than using a common threshold.

A Hidden Markov Model approach to online handwritten signature verification

A method for the automatic verification of online handwritten signatures using both global and local features is described. The global and local features capture various aspects of signature shape and dynamics of signature production. We demonstrate that adding a local feature based on the signature likelihood obtained from Hidden Markov Models (HMM), to the global features of a signature, significantly improves the performance of verification. The current version of the program has 2.5% equal error rate. At the 1% false rejection (FR) point, the addition of the local information to the algorithm with only global features reduced the false acceptance (FA) rate from 13% to 5%.