An Efficient Signature Verification Method Based on an Interval Symbolic Representation and a Fuzzy Similarity Measure

Abstract

In this paper an efficient off-line signature verification method based on an interval symbolic representation and a fuzzy similarity measure is proposed. In the feature extraction step, a set of Local Binary Pattern (LBP) based features is computed from both the signature image and its under-sampled bitmap. Interval-valued symbolic data is then created for each feature in every signature class. As a result, a signature model composed of a set of interval values (corresponding to the number of features) is obtained for each individual's handwritten signature class. A novel fuzzy similarity measure is further proposed to compute the similarity between a test sample signature and the corresponding interval-valued symbolic model for the verification of the test sample. To evaluate the proposed verification approach, a benchmark offline English signature dataset (GPDS-300) and a large dataset (BHSig260) composed of Bangla and Hindi off-line signatures were used. A comparison of our results with some recent signature verification methods available in the literature was provided in terms of average error rate and we noted that the proposed method always outperforms when the number of training samples is eight or more.

A Comprehensive Study of Sparse Representation Techniques for Offline Signature Verification

“…For this purpose, two different types of noise, namely Salt & Pepper, and Gaussian white noise were employed on the CEDAR dataset with various noise levels. It has been reported that these two types of noise commonly appear in images during the data collection process [53]. Thus, we added noise in the original signature images as follows: a) Salt and Pepper with parameters d=0.001, 0.01 and 0.1 and b) Gaussian with mean and variance parameters i) m=0 & v=0.01, ii) m=0 & v=0.1 and iii) m=0.2 & v=0.01.…”

A Perspective Analysis of Handwritten Signature Technology

“…The classification was carried out through an NN technique for Hindi and Bengali signatures separately. By using the same 117:15 dataset, promising results were found in [7] using a symbolic interval representation of signatures and fuzzy similarities. In this case, the authors performed experiments without separating Bengali and Hindi signatures.…”

“…Number of strokes/lognormals [83,248], * mathematical transformations (Wavelet [98,200,202,203], * Radon [268], * fractal [84], ** or others like total number, mean, and maximum of intra-/interstroke intersections, the number of x-axes, zero-crossings and the signature length [81,91,93,154,202,246,250]*), symbolic representation [7], * code vectors [234], * Sigma-lognormal based [83,102], * ...…”

“…Additionally, fuzzy membership functions are still used in signature verification. For example, in [7], a fuzzy similarity measure is used to classify a symbolic representation of offline signatures, and this method also achieves a competitive performance.…”

Fixed-sized representation learning from offline handwritten signatures of different sizes

“…These features range from simple descriptors such as the size of the signature and inclination [16], graphometric features [17], [18], texture-based [19,20], interest point-based [21], among others. Recent advancements in this field include using multiple classifiers trained with different representations [20], using interval symbolic representations [22] and augmenting datasets by duplicating existing signatures or creating synthetic ones [23][24][25]. More recently, methods for learning features from signature images have been proposed [5][6][7][8].…”