Wyniki wyszukiwania - BazTech

1

Dynamic handwritten signature identification using spiking neural network

Kutsman Vladislav, Kolesnytskyj Oleh

Informatyka, Automatyka, Pomiary w Gospodarce i Ochronie Środowiska

|

2021

|

T. 11, nr 3

34--39

EN

The article proposes a method for dynamic signature identification based on a spiking neural network. Three dynamic signatureparameters l(t), xy(t), p(t) are used, which are invariant to the signature slope angle, and after their normalization, also to the signature spatial and temporal scales. These dynamic parameters are fed to the spiking neural network for recognition simultaneously in the form of time series without preliminary transformation into a vector of static features, which, on the one hand, simplifies the method due to the absence of complex computational transformation procedures,and on the other hand, prevents the loss of useful information, and therefore increases the accuracy and reliability of signature identificationand recognition (especially when recognizing forged signatures that are highly correlated with the genuine). The spiking neural network used has a simple training procedure, and not all neurons of the network are trained, but only the output ones. If it is necessary to add new signatures, it is not necessaryto retrain the entire network as a whole, but it is enough to add several output neurons and learn only their connections. Inthe results of experimental studies of the software implementation of the proposed system, it’s EER = 3.9% was found when identifying skilled forgeries and EER = 0.17% when identifying random forgeries.

PL

W artykule zaproponowano metodę dynamicznej identyfikacji podpisów opartą na pulsującej sieci neuronowej. Wykorzystywane są trzy parametry dynamiczne podpisu l(t), xy(t), p(t), które są niezmienne względem kąta nachylenia podpisu, a po ich normalizacji –także do skali przestrzennej i czasowej podpisu. Te dynamiczne parametry są podawane do sieci neuronowej w celu rozpoznania jednocześnie jako szeregi czasowe bez uprzedniej konwersji na wektor cech statycznych, co z jednej strony upraszcza metodę ze względu na brak skomplikowanych procedur konwersji obliczeniowej,a z drugiej ręka zapobiega utracie przydatnych informacji –zwiększa dokładność i wiarygodność identyfikacjii rozpoznawania podpisów (zwłaszczaw rozpoznawaniu podpisów sfałszowanych, które są silnie skorelowane z autentycznymi).Zastosowana sieć neuronowa typu spiking ma prostą procedurę treningu, przy czym nie wszystkie neurony sieci są trenowane, a jedynie te wyjściowe.Jeśli konieczne jest dodanie nowych sygnatur, nie jest konieczne trenowanie całej sieci, ale wystarczy dodać kilka neuronów wyjściowych i uczyć tylko te połączenia.W wyniku eksperymentu programowego zaproponowanego systemu otrzymano EER = 3,9% przy identyfikacji sfałszowanych podpisów i EER = 0,17% przy identyfikacji fałszerstw losowych.

2

Determining valuable ranges of handwritten signature using fuzzy approach and window method

Kudłacik P., Doroz R.

Journal of Medical Informatics & Technologies

|

2015

|

Vol. 24

23--30

EN

The paper proposes possible improvements in signature recognition approach based on window method. The analysis focuses on a stage of window preprocessing using fuzzy sets in order to choose significant ranges of each signature. Proposed extension allows the solution to improve in two areas. First of all minimizing a number of processed windows significantly reduces computation time. Secondly, filtered signatures with valuable information about significant ranges allow the system to recognize signatures of a poor or good quality. Developed method of signature quality assessment can be used in any signature recognition system, regardless of used method of analysis. Merging the information about signature quality and choosing only important signature ranges should also improve the overall detection results, however, more examinations are needed to confirm this statement.

3

Method of signature recognition with the use of the complex features

Doroz R., Pałys M., Orczyk T., Safaverdi H.

Journal of Medical Informatics & Technologies

|

2014

|

Vol. 23

155--162

EN

In this paper a new method of handwritten signatures verification has been proposed. This method, for each signature, creates complex features which are describing this signature. These features are based on dependencies analysis between dynamic features registered by tablets. These complex features are then used to create vectors describing the signature. Elements of these vectors are calculated using measures proposed in this work. The similarity between signatures is assessed by determining the similarity of vectors in the compared signatures. Research, whose results will be presented in the further part of this work, have shown a high efficiency of verification using proposed method.

4

Using hidden Markov models in signature recognition process

Doroz R., Wróbel K.

Journal of Medical Informatics & Technologies

|

2012

|

Vol. 21

75--84

EN

This paper presents a method of recognition of handwritten signatures with the use of Hidden Markov Models (HMM). The method in question consists in describing each signature with a sequence of symbols. Sequences of symbols were generated on the basis of an analysis of local extremes determined on diagrams of dynamic features of signatures. For this purpose, the method proposed by G.K. Gupta and R.C. Joyce has been modified. The determined sequences were then used as input data for the HMM method. The studies were conducted with the use of the SVC2004 database. The results are competitive in relation to other methods known from the literature.

5

The use of methods of statistical analysis in signature recognition system based on Levenshtein distance

Pałys M., Doroz R., Porwik P.

Journal of Medical Informatics & Technologies

|

2012

|

Vol. 21

67--73

EN

The study being presented is a continuation of the previous studies that consisted in the adaptation and use of the Levenshtein method in a signature recognition process. Three methods based on the normalized Levenshtein measure were taken into consideration. The studies included an analysis and selection of appropriate signature features, on the basis of which the authenticity of a signature was verified later. A statistical apparatus was used to perform a comprehensive analysis. The independence test ◈ was applied. It allowed determining the relationship between signature features and the error returned by the classifier.

6

Improving a signature recognition method using the fuzzy approach

Kudłacik P.

Journal of Medical Informatics & Technologies

|

2012

|

Vol. 21

85--93

EN

The paper introduces a significant improvement of the signature recognition method based on characteristic preprocessing of an input data. The original approach transforms an input data into a sorted set of points obtained from intersections of a signature with generated lines going through it's center point. For further analysis the discrete Walsh transform was used. The solution presented in this paper divides points obtained in the preprocessing phase into groups. This step allows the method to preserve more unique features, which positively reflects on the results. Preprocessed data is used to build a fuzzy structure called the fuzzy signature. The method considering a natural imprecision makes the verification system flexible.

7

The method of signature recognition based on least squares contour alignment and windows technique

Wróbel K., Doroz R.

Journal of Medical Informatics & Technologies

|

2010

|

Vol. 15

35--39

EN

This paper presents a new method of recognizing handwritten signatures. Signature was treated as a collection of features of specific values. As features the values of x, y coordinates of signature points have been used. The method discussed in the paper is a modification of the method based on least squares contour alignment. This modification consists of dividing signatures into windows of the preset size and measuring the value of similarity between the windows according to their position in the signature. The effectiveness of the method was verified in practice. During the study, the influence of the parameters of the method on the obtained results was determined.

8

The method for determining the characteristic points of signatures based on IPAN99 algorithm

Doroz R., Kostorz I.

Journal of Medical Informatics & Technologies

|

2010

|

Vol. 15

41-46

EN

The paper puts forward a new method of determination of signatures' characteristic points. The method is based on seeking points of the highest curvature using the IPAN99 algorithm. The way of IPAN99 algorithm parameters' automatic selection for a particular signature has been fully described. Moreover, the way of determination of additional characteristic points, important for a signatures analysis, has been shown. The presented results of carried out experiments confirm that the proposed method is useful for signature recognition and verification.

9

Signatures recognition method by using the normalized Levenshtein distances

Doroz R., Wróbel K., Porwik P.

Journal of Medical Informatics & Technologies

|

2009

|

Vol. 13

73--77

EN

This study examines the effectiveness of normalized Levenshtein metrics in the process of recognition of handwritten signatures. Three methods of normalization of the Levenshtein metric were taken into consideration. In addition, it was determined, which signature features are most important during their comparisons with the use of the aforementioned metric. The following signature features were examined: coordinates of signature points, pen pressure in successive points, and different types of pen speed. The influence of individual parameters of the Levenshtein algorithm on the obtained results was also determined, and the best method of normalization was selected.

10

Signature pre - procesing based on walsh coefficients

Porwik P., Wróbel K.

Journal of Medical Informatics & Technologies

|

2008

|

Vol. 12

57--61

EN

Recognition and verification of persons are difficult and important tasks today. In many fields of human activities (driver's licenses, passports, electronic cards, etc.), signature recognition of person is needed. Hence, it inspires the development of a wide range of automatic identification systems. Signatures have been used for many centuries as a method of people's identification. Signatures recognition was performed manually by experts in the past. Nowadays, these procedures are very often automatically applied. In this paper the system that automatically authenticates documents based on the owner's handwritten signature is presented.

11

On some optimalization of signature recognition

Doroz R., Bargielski M.J.

Journal of Medical Informatics & Technologies

|

2008

|

Vol. 12

201--206

EN

Signature recognition is one of the important problems nowadays. In paper we present known method of pattern (curves) recognition, i.e. algorithm IPAN99 and researches over its optimization; there are many control parameters which influence on recognition results. We present some quasi-optimal set of control parameter. Our next aim is to automatically find proper parameters. Thus some optimum seeking method for unimodale and multimodale function is proposed.

12

New signature similarity measure based on average differences

Wróbel K., Doroz R.

Journal of Medical Informatics & Technologies

|

2008

|

Vol. 12

51--56

EN

This paper presents a new method of recognizing handwritten signatures. Signature was treated as a collection of features of specific values. As features the values of x, y coordinates of signature points and the pressure p in its consecutive points have been used. Additionally, before comparing them, the signatures were properly prepared. The method discussed in the paper is a modification of the method based on average differences. This modification consists in dividing signatures into windows of the preset size and measuring the value of similarity between the windows according to their position in the signature. The paper shows the construction of a new similarity measure taking into consideration the modifications introduced and the results of the research obtained by means of this similarity measure.

13

Determining signatures' characteristic features using statistical methods

Para T., Mitas M.

Journal of Medical Informatics & Technologies

|

2008

|

Vol. 12

41--49

EN

Many signatures verification systems have been developed so far. Most of them have some common algorithms and solutions. The problem is that authors of the solutions usually present the final results of a working system. They do not reveal effects of the particular components. This is why other researchers do not know which element improves the results and is worth using. This paper shows how to estimate, in an easy way, if the selected component/set of data/feature gives good results.

14

Signature image recognition by shape context image matching

Adamski M., Saeed K.

Journal of Medical Informatics & Technologies

|

2007

|

Vol. 11

89--95

EN

This paper presents experiments on recognition of signature images. In preprocessing stage a thinning algorithm is used followed by a sampling technique. Sampled points are used to calculate shape context histograms and based on their values corresponding pairs of points from reference and tested signature objects are selected. A distance measure based on shape contexts is used to classify analysed signatures.

15

On-line signature recognition method based on linear regression

Para T., Porwik P., Wróbel K.

Journal of Medical Informatics & Technologies

|

2007

|

Vol. 11

97--104

EN

Nowadays, automatic signature verification is an active area of researches in numerous applications such as bank check verification, access restriction or special areas such as police investigations. In our researches signature was captured by Topaz SigLite T-LBK750-HSB device, where some dynamic features of signature can be also registered. In many transactions, the electronic verification of a person's identity is beneficial, hence it inspires the development of a wide range of automatic identification systems. In this paper the system that automatically authenticates documents based on the owner's handwritten signature is presented.

16

Combined off-line type signature recognition method

Porwik P., Para T., Smacki Ł., Żuławiński S.

Journal of Medical Informatics & Technologies

|

2006

|

Vol. 10

81--90

EN

In this paper the off-line type signature analysis have been presented. The signature recognition is composed of some features. Different influences of such features were tested and stated. Proposed approach gives good signature recognition level, hence described method can be used in many areas, for example in biometric authentication, as biometric computer protection or as method of the analysis of person's behaviour changes.