Noise Suppression and Contrast Enhancement via Bayesian Residual Transform (BRT) in Low-Light Conditions

Very low-light conditions are problematic for current robotic visionalgorithms as captured images are subject to high levels of ISOnoise. We propose a Bayesian Residual Transform (BRT) model forjoint noise suppression and image enhancement for images capturedunder these low-light conditions via a Ba...

Full description

Saved in:
Bibliographic Details
Published in:Journal of Computational Vision and Imaging Systems 2016-10, Vol.2 (1)
Main Authors: Chung, Audrey G., Wong, Alexander
Format: Article
Language:English
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Very low-light conditions are problematic for current robotic visionalgorithms as captured images are subject to high levels of ISOnoise. We propose a Bayesian Residual Transform (BRT) model forjoint noise suppression and image enhancement for images capturedunder these low-light conditions via a Bayesian-based multiscaleimage decomposition. The BRT models a given image as thesum of residual images, and the denoised image is reconstructedusing a weighted summation of these residual images. We evaluatethe efficacy of the proposed BRT model using the VIP-LowLightdataset, and preliminary results show a notable visual improvementover state-of-the-art denoising methods.
ISSN:2562-0444
2562-0444
DOI:10.15353/vsnl.v2i1.110