Figure 1 - uploaded by Derlin Morocho
Content may be subject to copyright.
Genuine (from two different signers) and forgeries (made by other people after practicing for 2 minutes). Which signatures are genuine? See solution at 1
Source publication
This work explores crowdsourcing for the establishment of human baseline performance on signature recognition. We present five experiments according to three different scenarios in which laymen, people without Forensic Document Examiner experience, have to decide about the authenticity of a given signature. The scenarios include single comparisons...
Context in source publication
Context 1
... is expected a drop of performance caused by the lack of information about the variability of the owner. In addition, we have to consider motivation as an important factor to be considered. Without specific training and considering that signature recognition is not the principal job assignment of the laymen, their performance is an open question. Fig. 1 tries to illustrate the difficulties related with this ...
Similar publications
![Figure 2. Crowdsourcing Process [12]](profile/Derlin-Morocho/publication/318123425/figure/fig7/AS:668533331619850@1536402175337/Crowdsourcing-Process-12_Q320.jpg)
![Figure 3. Signature Recognition via Crowdsourcing [13]](profile/Derlin-Morocho/publication/318123425/figure/fig2/AS:580933630005248@1515516778721/Signature-Recognition-via-Crowdsourcing-13_Q320.jpg)
![Figure 5. Features used in the test. [14]](profile/Derlin-Morocho/publication/318123425/figure/fig4/AS:580933629165569@1515516778965/Features-used-in-the-test-14_Q320.jpg)
Citations
... It is reasonable to assume that the analysis performed by a non-FDE human (excluding the experience and training) is mostly based on the personal subjectivity of each subject. While the baseline performance of FDEs has been analysed in the literature [18,19], to the best of our knowledge, the literature lacks of studies analysing the baseline performance of laymen [13,[20][21][22]. Crowdsourcing was employed in [21] in order to establish a human baseline performance on signature authentication. ...
... While the baseline performance of FDEs has been analysed in the literature [18,19], to the best of our knowledge, the literature lacks of studies analysing the baseline performance of laymen [13,[20][21][22]. Crowdsourcing was employed in [21] in order to establish a human baseline performance on signature authentication. The experiments reported over responses from 150 laymen shown performances ranging from 7% (false acceptance rate) to 80% (false rejection rate) depending on the scenario and the information provided to the users. ...
... The results reported suggest the potential of human annotations to improve The present work is a step forward in the analysis of the potential of human intervention to improve AS authentication. The contributions of this work are three-fold: (i) we extended the experiments presented in [21] with responses from 500 laymen, providing new insights in the performance of humans in signature authentication tasks based on crowdsourcing experiments; (ii) we present a deep analysis of the attribute-based signature authentication system proposed in [13] via human interventions made by 11 different annotators and the complete BiosecurID-UAM corpus (3696 signatures from 132 different signers); and (iii) we analyse the performance of combined schemes incorporating the proposed attribute-based manual approach to online and offline state-of-the-art ASV systems. ...
This work explores human intervention to improve automatic signature authentication systems. Significant efforts have been made in order to improve the performance of automatic signature authentication algorithms over the last decades. This work analyzes how human actions can be used to complement automatic systems. Which actions to take and to what extend those actions can help state-of-the-art Automatic Signature Verification systems (ASV) is the final aim of this research line. This work studies human intervention at classification and feature extraction levels. The analysis at classification level comprises experiments with responses from 500 people based on crowdsourcing signature authentication tasks. The results allow to establish a human baseline performance and comparison with automatic systems. Intervention at feature extraction level is evaluated using a self-developed tool for the manual annotation of signature attributes inspired in FDE analysis. We analyze the performance of attribute-based human signature authentication and its complementarity with automatic systems. The experiments are carried out over a public database including the two most popular signature authentication scenarios based on both online (dynamic time sequences including position and pressure) and offline (static images) information. The results demonstrate the potential of human interventions at feature extraction level (by manually annotating signature attributes) and encourage to further research in its capabilities to improve the performance of Automatic Signature Verification systems.
... Often, this variability is attributed to the several sources of noise (?) that distort the measured trait. According to Figure 1.1, the intra-personal variability which affects to the measured sample (M) could be characterize by: sensor limitations like resolution or sample rate; biological aging effects or cognitive-motor impairments; user interaction with Morocho et al., 2016) the sensor; environment changes like background noise and; other factors as consequence of the individuals' mood, hurry or willingness to cooperate. Another greatest challenge faced by signature-based biometric systems is the unpredictable inter-personal variability. ...
... In order to validate the correctness of our HMM implementation, we also tested the system on MCYT-100 using the first 10 genuine signatures of each user for training and achieved an equal error rate of 0.80 % for random forgeries and 3.76 % for skilled forgeries. These results were indeed very similar to the results reported in (Fierrez et al., 2007Morocho et al., 2016) Aspectos emergentes en la verificaci?n autom?tica de firmas La verificaci?n autom?tica de firmas (ASV) tiende a centrarse en la mejora de la precisi?n del reconocimiento, aunque temas como la interoperabilidad, las normas, la escalabilidad y la protecci?n est?n tambi?n ganando atenci?n en la comunidad cient?fica. ...
Learning to write is complex and usually starts with lines and scribbles. Before reaching a mature handwriting, children start to know the letters' shapes and their sequence, although the children's motor control is not yet accurate. Modeling this behavior in a mathematically way would allow to understand the mechanical processes from the initial thought of signing to its complete fulfillment.
For instance, statistical models of a particular muscle could gain a better understanding of its general behavior when a stimulus is applied. The kinematical response of an executed movement is also a source of information about the human reaction. Indeed, these characteristics could be mathematically modeled according to the literature in order to design synthetically human movements.
On the other hand, handwriting signature is used as a biometric trait to authenticate the user identity. However, the signature-based biometric systems are not used in practical applications due to their lower performance compared to other biometric technologies. Therefore, it is often preferred to use other traits such as iris, fingerprint or face.
As a bridge between synthesis of biometric data and human modeling, innovative methods are addressed in this dissertation to generate synthetic handwriting signatures following the insights learnt from the motor equivalence theory.
As such, in this Thesis several procedures are proposed to generate i) fully synthetic signature databases and ii) duplicated signatures from a single real specimen. The goal of the proposed methods is to verify whether the generated signatures are able to introduce realistic intra and inter-personal variability in signature-based biometric systems as well as to certify their human-like appearance. For these purposes, machine-oriented and human-oriented evaluations are discussed in the frameworks used in this document.
... These protocols are mostly based on the analysis of local features related with the personal characteristics of each signer (see Fig. 1). Previous studies in human performance focused on signature recognition via crowdsourcing [6]. These studies analyze the performance of humans without FDE experience (called layman). ...
... The parameters used in this protocol are based on selecting discriminating features inspired by FDEs, in order to assess their applicability, and their performance in signature recognition [2,6,7,8,9]. Inspired by the FDEs analisys we choose nine popular characteristics employed in signature recognition. ...
... 1). Previous studies in human performance focused on signature recognition via crowdsourcing[6]. These studies analyze the performance of humans without FDE experience (called layman). ...
This work explores the human ability to recognize the authenticity of signatures. We use crowdsourcing to analyze the different factors affecting the performance of humans without Forensic Document Examiner experience. We present different experiments according to different scenarios in which laymen, people without Forensic Document Examiner experience, provide similarity measures related with the perceived authenticity of a given signature. The human responses are used to analyze the performance of humans according to each of the scenarios and main factors. The experiments comprise 240 signatures from BiosecurID public database and responses from more than 400 people. The results shows the difficulties associated to these tasks, with special attention to the false acceptance of forgeries with rates ranging from 50% to 75%. The results suggest that human recognition abilities in this scenario are strongly dependent on the characteristics considered and the signature at hand. Finally the combination of human ratings clearly outperfoms the individual performance and and a state-of-the-art automatic signature verification system.
... The performance of people doing signature recognition is studied in [8,14] and shows promising results in human performance as they are being able to distinguish a genuine signature from a forged one. Due to the positive results of these studies, it is proposed to use the tool of crowdsourcing implemented in MTurk to attack the problem of intravariability to make handwritten signature recognition with human intervention. ...
This paper presents discriminative features as a result of comparing the authenticity of signatures, between standardized responses from a group of people with no experience in signature recognition through a manual system based on crowdsourcing, as well as the performance of the human vs. an automatic system with two classifiers. For which an experimental protocol is implemented through interfaces programmed in HTML and published on the platform Amazon Mechanical Turk. This platform allows obtaining responses from 500 workers on the veracity of signatures shown to them. By normalizing the responses , several general features which serve for the extraction of discriminative features are obtained in signature recognition. The comparison analysis in terms of False Acceptance Rate and False Rejection Rate founds the presented features, which will serve as a future study of performance analysis in the implementation of automatic and semiautomatic signature recognition systems that will support financial, legal and security applications.
Finger vein biometrics is becoming an important source of human authentication due to its advantages in terms of liveness detection, high security, and user convenience. Although there exist a lot of deep learning-based methods for finger vein authentication, they only extract features from finger vein images in the spatial domain and may lose some important information that is present in other domains, such as the frequency domain. Motivated by this conjecture and the remarkable performance of image feature extraction in the frequency domain, this work explores a method capable of extracting finger vein features in both the spatial and frequency domains. Therefore, the features extracted from different domains can complement each other. In addition, we propose a novel frequency-spatial coupling network (FVFSNet) for finger vein authentication. FVFSNet is mainly composed of three parts: (1) the frequency domain processing module (FDPM), (2) the spatial domain processing module (SDPM), and (3) the frequency-spatial coupling module (FSCM). The FDPM is used to extract the finger vein features present in the frequency domain, which is mainly composed of the frequency-spatial domain transformation and the frequency domain convolution layer. The SDPM is used to extract the finger vein features present in the spatial domain, which is mainly composed of convolution layers with an efficient design. The FSCM is used to couple the features extracted from the FDPM and SDPM, which is mainly composed of the channel and spatial attention mechanisms. To validate our conjecture and the performances of FVFSNet, extensive experiments are conducted on nine commonly used publicly available finger vein datasets. Experimental results show that the frequency domain constitutional neural network has a surprising effect on finger vein authentication, and the proposed FVFSNet achieves the state-of-the-art performance with the advantages of lightweight and low computational cost.
