User authentication scheme based on Fast-Walsh Hadamard transform

Abstract

User authentication scheme in context of audio can be achieved by labeling the speech signal with key signal dependent physical features. The key features for the development of the label are identified by using the common classification schemes such as ANN, kNN and SVM. The relevant features are given as input to an online QR code generator to obtain the QR code which is employed as the watermark. Using the FWHT transform, speech signal is then labeled and the confirmation along with comparison at the receiving end guarantees speaker authentication.