On the Accuracy of Vector Metrics for Quality Assessment in Image Filtering

Fabrizio Russo

Abstract

Noise cancellation and detail preservation are two conflicting goals in the design of any denoising filter. In order to specifically measure these features, vector metrics have been recently introduced for grayscale and color images. This paper provides, for the first time, a study of the accuracy of the main algorithms in the literature focusing on the case of Gaussian noise. A novel vector method is also proposed to overcome the inaccuracies of current techniques.

Keywords

image processing, denoising filters, image quality, vector metrics, Gaussian noise.

Download

IMEKO-TC4-2014-304.pdf

DOI

-

IMEKO TC

TC4 - Measurement of Electrical Quantities

Back to the proceedings