Narrow Results

Biometric authentication based on fingerprints, voice, hand shape, facial measurements and iris analysis, among others, are quite common nowadays. In a similar manner, the analysis of acoustic patterns generated during the friction between pen and paper at the time a person subscribes has been shown to be a feasible, adequate, and non-invasive alternative to those techniques. An interesting implementation for such an approach, which is described in this paper, is based on the association of the time-frequency analysis supported by the discrete wavelet-packet transform with one of two pattern-matching classifiers, namely Euclidian norma and an original scoring equation derived from correlation, acting semantically. Valuable results were obtained during the tests, motivating further research. The proposed technique is novel on literature, offering a contribution to the state-of-the-art.