The Maximal Overlap Discrete Wavelet Scattering Transform and Its Application in Classification Tasks

TL;DR

The paper introduces the Maximal Overlap Discrete Wavelet Scattering Transform (MODWST), a new method for signal classification that performs well with limited training data, offering an alternative to CNNs.

Contribution

It proposes the MODWST, combining ideas from MODWT and WST, and demonstrates its effectiveness in stationary and ECG signal classification tasks.

Findings

MODWST achieved good classification performance.

It is effective with limited training data.

It competes well with CNNs in certain applications.

Abstract

We present the Maximal Overlap Discrete Wavelet Scattering Transform (MODWST), whose construction is inspired by the combination of the Maximal Overlap Discrete Wavelet Transform (MODWT) and the Scattering Wavelet Transform (WST). We also discuss the use of MODWST in classification tasks, evaluating its performance in two applications: stationary signal classification and ECG signal classification. The results demonstrate that MODWST achieved good performance in both applications, positioning itself as a viable alternative to popular methods like Convolutional Neural Networks (CNNs), particularly when the training data set is limited.