Signature recognition

About: Signature recognition is a research topic. Over the lifetime, 2138 publications have been published within this topic receiving 37605 citations.

An individuality model for online signatures using global Fourier descriptors

Abstract: The discriminative capability of a biometric is based on its individuality/uniqueness and is an important factor in choosing a biometric for a large-scale deployment. Individuality studies have been carried out rigorously for only certain biometrics, in particular fingerprint and iris, while work on establishing handwriting and signature individuality has been mainly on feature level. In this study, we present a preliminary individuality model for online signatures using the Fourier domain representation of the signature. Using the normalized Fourier coefficients as global features describing the signature, we derive a formula for the probability of coincidentally matching a given signature. Estimating model parameters from a large database and making certain simplifying assumptions, the probability of two arbitrary signatures to match in 13 of the coefficients is calculated as 4.7x10 -4 . When compared with the results of a verification algorithm that parallels the theoretical model, the results show that the theoretical model fits the random forgery test results fairly well. While online signatures are sometimes dismissed as not very secure, our results show that the probability of successfully guessing an online signature is very low. Combined with the fact that signature is a behavioral biometric with adjustable complexity, these results support the use of online signatures for biometric authentication.

Efficient Maximum Entropy Training for Statistical Object Recognition

Abstract: In statistical pattern recognition, we use probabilistic models within the task of assigning observations to one of a set of predefined classes, like e.g. images of handwritten digits to one of the classes ‘0’ to ‘9’. The principle of maximum entropy is a powerful framework that can be used to estimate class posterior probabilities for pattern recognition tasks. It is a conceptually simple and easily extensible model that allows to estimate a large number of free parameters reliably. We show how to apply this framework to object recognition and compare the results to other state-of-the-art approaches in experiments with the well known US Postal Service handwritten digits recognition task. We also introduce a simple but effective heuristic method for speeding up the algorithms used to determine the model parameters.

Online signature slant feature identification algorithm

Abstract: According to the American National Science and Technology Council (NSTC), the first signature recognition system was developed in 1965. Then the research continued in 1970 focusing on the potential of geometric characteristic of a signature rather than dynamic characteristic. Nowadays, signature is a commonly used identification procedure. Everyone would be required having a signature for authorization and other important tasks that needs identification. Thus, signature has become one of a method to represent its writer uniquely. Signature has many hidden features that are difficult to extract. Some of the identified features that a signature should have are slanting, baseline, proportion and size. This paper covers the area of signature slant identification. Signatures are captured using a tablet and saved in a digitized format of x and y values. Then it is filtered and calculated for its angle and degree. In the end the signature will be classified to its slant category. A slant algorithm is created and coded into a functional system. An experiment consisting of 50 signatures are tested and the finding shows the angle and degree of the slant in every signature. The result is then tested for its accuracy with an available 10 sample of created proofed signatures. The result shows a favorable accuracy of 80% correct slant identification. The creation of this algorithm would be able to give some degree of contribution in the area of signature recognition.

Designing a Fuzzy RBF Neural Network with Optimal Number of Neuron in Hidden Layer and Effect of Signature Shape for Persian Signature Recognition by Zernike Moments and PCA

Abstract: This paper presents an efficient method for Persian signature recognition based on Fuzzy RBF neural network (FRBF). A new training method will be presented which had a very low error rates in Persian signature recognition. In this training algorithm, connection weights, centers, width and number of RBF units will be determined during training phase. FCM algorithm will be used for initializing parameters. The membership of input patterns and distance from centers in each RBF unit calculate cost function for each input pattern. In this study Zernike Moment (ZM) and Principle Component Analysis (PCA) have been used as features. Also has been inspected effect of signature shape in system error. Simulation results on signature database from Persian peoples which contains 200 pictures indicate that the proposed system not only has a low error rate, but also determine the optimal number of RBF units.