Reducing forgeries in writer-independent off-line signature verification through ensemble of classifiers

Texture Analysis for Off-Line Signature Verification

“…Random forgeries were used as negative samples (genuine samples of other signers). The use of these samples was proposed in Bertolini et al (2009), and for this work it was taken a genuine sample of each of the other signers of the database, ie 74. Given the small number of samples available for training was used the Leaving-one-out cross-validation procedure (LOOCV), to determine the value of the parameters (γ, C) for LS-SVM classifier with RBF kernel.…”

Forensic Analysis of Offline Signatures Using Multilayer Perceptron and Random Forest

“…Bertolini et al in [11], identified new feature set containing curvature of the most important segments of signature images. The Bezier curve was simulated and the features of signature images were extracted from that curve.…”

Approaches and Issues in Offline Signature Verification System

“…Bertolini et al [36] proposed a method of off-line signature verification through ensemble of classifiers. They tried to simulate the shape of the signature by using Bezier curves and then extracted features from those curves.…”