BIOMETRIK SUARA DENGAN TRANSFORMASI WAVELET BERBASIS ORTHOGONAL DAUBENCHIES

Abstract

Biometric as one of identification or recognition person techniques that based on  uniquely part of human body. Voice one of uniquely human has. Voice signal that come out from different speakers give different voice pattern. Because of high variations used neural network (NNW)for matching the patterns. Before voice data is processed using NNW it’s processed using digital processed through feature extraction phase using discrit wavelet orthogonal base 4 orders with 10 and 15 decomposition rate. The result of NNW is processed by model decision maker that determine speaker identification. The result of experiment shows system biometric built can identify as high as 86%