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

Guido, Rodrigo Capobianco [UNESP]; Pedroso, Fernando [UNESP]; Contreras, Rodrigo Colnago [UNESP]; Rodrigues, Luciene Cavalcanti [UNESP]; Guariglia, Emanuel [UNESP]; Neto, Jogi Suda [UNESP]

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

Data

2021-10-01

Autores

Guido, Rodrigo Capobianco

Pedroso, Fernando

Contreras, Rodrigo Colnago

Rodrigues, Luciene Cavalcanti

Guariglia, Emanuel

Neto, Jogi Suda

Tipo

Artigo

Resumo

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.