Sparse transform based online video denoising using block-matching and ICA

Authors

  • Archana K, Dr. Arathi T

Abstract

As we live in a smartphone era, we want to store every moment of our day-to-day life, that
is recorded using camera, low-quality smartphone cameras have opened demand for video denoising. Video
denoising is mainly done by utilizing the properties of spatial and temporal redundancy of noisy video to
reconstruct the estimated video. Here, introduces a new method of transform domain video denoising using
sparse processing and independent component analysis (ICA). This work aims to develop an online transform
learning algorithm for denoising, using the sparse coding and transform updation. This method of denoising
uses streaming videos for processing and the reconstructed videos can be used in application like video
communication, forensics and CCTV surveillance etc. Online sparse transform learning from streaming
signals involve cheap computation cost and good convergence guarantee. Moreover, this method increases the
denoising performance over the existing spatial and transform domain methods.

Published

2020-10-16

Issue

Section

Articles