Local denoising based on curvature smoothing can visually outperform non-local methods on photographs with actual noise

Ghimpeteanu, Gabriela; Kane, David; Batard, Thomas; Levine, Stacey; Bertalmío, Marcelo

doi:http://dx.doi.org/10.1109/ICIP.2016.7532932

Local denoising based on curvature smoothing can visually outperform non-local methods on photographs with actual noise

Citation

Ghimpeteanu G, Kane D, Batard T, Levine S, Bertalmío M. 2016. Local denoising based on curvature smoothing can visually outperform non-local methods on photographs with actual noise. In: 2016 IEEE International Conference on Image Processing Proceedings; 2016 Sep 25-28; Phoenix, Arizona, USA. New York: IEEE; 2016. p. 3111-5. DOI: 10.1109/ICIP.2016.7532932

Abstract

We propose a fast, local denoising method where the Euclidean curvature of the noisy image is approximated in a regularizing manner and a clean image is reconstructed from this smoothed curvature. User preference tests show that when denoising real photographs with actual noise our method produces results with the same visual quality as the more sophisticated, nonlocal algorithms Non-local Means and BM3D, but at a fraction of their computational cost. These tests also highlight the limitations of objective image quality metrics like PSNR and SSIM, which correlate poorly with user preference.

Description

Comunicació presentada a: IEEE International Conference on Image Processing 2016, celebrada del 25 l 28 de setembre de 2016 a Phoenix, Estats Units d'Amèrica.