Oscar Miguel-Hurtado

On-Line Signature Verification by Dynamic Time Warping and Gaussian Mixture Models

Handwriting signature is the most diffuse mean for personal identification. Lots of works have been carried out to get reasonable errors rates within automatic signature verification on-line. Most of the algorithms that have been used for matching work by features extraction. This paper deals with the analysis of discriminative powers of the features that can be extracted from an on-line signature, how its possible to increase those discriminative powers by dynamic time warping as a step in the preprocessing of the signal coming from the tablet. Also it will be covered the influence of this new step in the performance of the Gaussian mixture models algorithm, which has been shown as a successfully algorithm for on-line automatic signature verification in recent studies. A complete experimental evaluation of the algorithm base on dynamic time warping and Gaussian Mixture Models has been conducted on 2500 genuine signatures samples and 2500 skilled forgery samples from 100 users. Those samples are included at the public access MCyT-Signature-Corpus Database.

Performance evaluation of handwritten signature recognition in mobile environments

The utilisation of biometrics in mobile scenarios is increasing remarkably. At the same time, handwritten signature recognition is one of the modalities with highest potential of use for those applications where customers are used to sign in those traditional processes. However, several improvements have to be made in order to reach acceptable levels of performance, reliability and interoperability. The evaluation carried out in this study contributes with multiple results obtained from 43 users signing 60 times, divided in three sessions, in eight different capture devices, being six of them mobile devices and the other two digitisers specially made for signing and used as a baseline. At each session, a total of 20 signatures per user are captured by each device, so that the evaluation here reported a total of 20 640 signatures, stored in ISO/IEC 19794–7 format. The algorithm applied is a DTW-based one, particularly modified for mobile environments. The results analysed include inter-operability, visual feedback and modality tests. One of the big challenges of this research was to discover if the handwritten signature modality in mobile devices should be split into two different modalities, one for those cases when the signature is performed with a stylus, and another when the fingertip is used for signing. Many relevant conclusions have been collected and, over all, multiple improvements have been reached contributing to future deployments of biometrics in mobile environments.

FPGA implementation for an iris biometric processor

Biometrics is nowadays one of the most promising techniques in authentication. Biometrics intends to identify a user by his/her physical and/or behavioural characteristic. Among all Biometric techniques, Iris recognition stands out, as its error rates are one of the lowest. The authors propose in this paper a hardware implementation based on FPGA for an iris biometric processor. By this solution a reduction of the processing time is obtained and security levels of the whole system are increased due to the reduction of software involved

Analysis of handwritten signature performances using mobile devices

This paper continue our previous work on on-line signature verification on portable devices. A database has been collected which contains data for 25 users, with 28 genuine and 25 skilled forged signatures per user, in 5 different devices. 4 portable devices of different sizes and screen technologies (capacitive and resistive) have been used. The 5th device is a traditional digital pen tablet, to use as a baseline for results comparison. Two different algorithms, DTW and SVM, have been used to asses the performance of the signals captured on portable devices. Two main experiments using both algorithms and the 5 sub-databases have been done. The first experiment uses each database independently. The user model is created and tested with signatures of the same database. The second experiment creates the user model using signatures acquired with the digital pen tablet. Results of the first experiment achieve good error rates for random forgeries.

Analysis on compact data formats for the performance of handwritten signature biometrics

This paper deals with the Signature Data Formats proposed by ISO 19794 project: 19794 part 7 Full Format and Compact Format (published in 2007) and the new 19794 part 11, which is under development. It will be shown how these formats handle the raw data coming from a Signature Input Device, and what the size of a Biometric Information Record is for each one. Another compression method, using LZ77 compression algorithm, is proposed and tested. The paper will also show the impact of using these compact formats on the performance of two different algorithms: Dynamic Time Warping and Gaussian Mixture Models. MCyT and SVC2004 signature databases have been used to carry out all tests.

A new algorithm for signature verification system based on DTW and GMM

This paper continue our previous work on combining dynamic time warping and Gaussian mixture modelling for automatic signature verification. MCyT signature database which contains data for 100 users, with 25 genuine and 25 forged signatures per users, has been used to test the proposed methods.

Usability analysis of a handwritten signature recognition system applied to mobile scenarios

Along with the necessity to solve the problems due to the misuse of biometric systems and thus the consistent increase in the final products error rates, a usability evaluation on handwritten signature recognition was carried out. Furthermore, according to the popularity of mobile devices and the market trends, the evaluation was performed signing in mobile scenarios with smart phones, tablets and other common mobile devices. This study reveals interesting results correlating habituation and preferences with better or worst results and it shows the need of involve the user more incisively in the development of biometric solutions, not only for comfort issues but for better systems throughput.

Analysis on the resolution of the different signals in an on-line handwritten signature verification system applied to portable devices

Nowadays, the use of portable devices with touch screens has reached an extended use among the population. With this technology, it is possible to incorporate to such devices the possibility of using handwritten signature to authenticate the user. In order to study the possibilities to carry on this authentication, a previous analysis has been carried out in this paper to evaluate how an already implemented on-line signature authentication algorithm will work in portable devices. Signature database signals have been re-scaled to emulate signals taken from a portable device, emulating two technologies, capacitive and resistive screens. For this, using only 3 of the 5 standard signals captured by graphic tablets: X, Y and P. Support Vector Machines have been used as an algorithm to test against the modified database, and results show good performance of the algorithms, obtaining rates around EER=3%, showing that the algorithms will give good results implemented in real smart phones.

Optimisation of biometric ID tokens by using hardware/software co-design

In current society, the necessity of recognising people is increasing every day. Logical or physical access is restricted to authorised users, which in many cases have to provide tokens where their personal information is stored. At the same time, biometrics proposes a feasible solution for the recognition problem. The combination of both solutions is coming up front. However, up till now, owing to processing restrictions, these tokens are just able to store data and perform the last steps of the biometric recognition process. In this study, the authors propose a new system where tokens are based on hardware/software (HW/SW) co-design, which allows computing most of the biometric process in them. This proposal covers several aspects which these systems are subject to, taking advantages of the two platforms they use for reducing computational time or HW area, and also to increase security or minimise misidentification errors. For testing this proposal, an Iris ID token has been implemented, showing different design alternatives adapted to different work scenarios.

Hardware/Software Codesign for an Iris Biometric Search Engine

Many airports are employing Biometric Technology to check the user identity nowadays. Till nowadays, the use of Biometric in this situation is restricted to verification, i.e., the user at the same time he/she provides his/her Biometric data, provides also his/her identity. However, in some cases, it is needed to search among all users stored in the database for the identity of the user. For this purpose, search engines are used. In this paper, authors propose an architecture for a Biometric search engine. The architecture proposal performs part of the algorithm using dedicated hardware and dataflow control is done by a microprocessor, obtaining with this configuration high speed and reliability.