Wavelet denoising as signal feature-dependent kernel convolution
Mohd Rozni Md Yusof, Ahmad Kamal bin Ariffin, Mohd Tahir Ismail · Journal of Physics Conference Series · 2019
This work presents a view of wavelet threshold denoising as a convolution process between a noisy signal and a low-pass kernel whose width varies depending on the features of the signal. Wavelet denoising has been widely used for many purposes across a wide range of disciplines, such as in Medical Imaging. However, the procedures are usually presented in a specialized mathematical notation that can be daunting to many researchers, especially those who are not specializing in mathematics. This work seeks to present wavelet denoising in a form more accessible to researchers in other disciplines such as Geographical Information Systems, image processing, and remote sensing.