“…Different learning strategies, such as machine learning and similarity-based approaches, have been proposed in the literature to solve the problem of signature verification. Machine learning approaches include Neural Networks (NNs) [4,7,19,42], Bayes classifier [3], Hidden Markov Models (HMMs) [30,34,36,38,41], Support Vector Machines (SVMs) [5, 10-14, 27-28, 31, 37, 38, 40-42], Gaussian Mixture Models (GMMs) [6], Gentle AdaBoost algorithm [39], and Ensembles of classifiers [41]. Similarity-based approaches comprise the following: k-Nearest Neighbour (kNN), Dynamic Time Warping, and point matching [3,6,8,12,19,29,[32][33][34][35][36].…”