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

  • Audrey G. Chung
  • Alexander Wong

Abstract

Very low-light conditions are problematic for current robotic vision
algorithms as captured images are subject to high levels of ISO
noise. We propose a Bayesian Residual Transform (BRT) model for
joint noise suppression and image enhancement for images captured
under these low-light conditions via a Bayesian-based multiscale
image decomposition. The BRT models a given image as the
sum of residual images, and the denoised image is reconstructed
using a weighted summation of these residual images. We evaluate
the efficacy of the proposed BRT model using the VIP-LowLight
dataset, and preliminary results show a notable visual improvement
over state-of-the-art denoising methods.

Published
2016-10-03
How to Cite
Chung, A., & Wong, A. (2016). Noise Suppression and Contrast Enhancement via Bayesian Residual Transform (BRT) in Low-Light Conditions. Journal of Computational Vision and Imaging Systems, 2(1). https://doi.org/10.15353/vsnl.v2i1.110
Section
Articles