Objective image quality measures by planar glyph representation

The commonly used non-perceptual image quality measures are mainly derived from pixel-to-pixel differences and closely relate to the mean squared error. Recently, there has been an increased interest in the evaluation of neighborhood-based differences between images, that take into account the local (spatial) image context. This contribution proposes the use of such local context information for the evaluation of image quality or inter-image differences by the planar glyph representation of multivariate data and the induced geometric-inspired difference measures.