Introducing the Discrete Path Transform (DPT) and its applications in signal analysis, artefact removal, and spoken word recognition

Abstract

This article introduces the Discrete Path Transform (DPT). Designed to serve as a new tool for handcraft feature extraction (FE), it improves the elementary analysis provided by signal energy (E) and enhances the humble spectral investigation granted by zero-crossing rates (ZCRs). C/C++ source-codes to realize both the DPT direct and inverse (IDPT) forms are presented together with a few hypothetical numerical examples and an application involving general signal analysis, artefact removal from biomedical signals, and spoken word recognition (SWR), thus demonstrating how useful and effective the proposed transform is. Brief comparisons with Teager Energy Operator (TEO) and a list of important references were also included.