Abstract—In this paper we present and examine a general SVD-based mathematical framework for the manipulation and transformation of images for traditional operations such as image compression and super-resolution to computational photography applications, e.g., to achieve novel aesthetic effects for the rapidly growing social media demand for image filter apps. Examples are given to demonstrate how images can be perturbed from realistic toward fanciful or from airbrushed gloss toward gritty realism. A primary goal of this paper is to draw greater attention to algorithmic methods that go beyond the traditional parameters of conventional image manipulation.

Index Terms—computational photography, singular value decomposition, SVD, orthogonal vector interpolation, OVI, image processing, image enhancement, image compression, image super-resolution

Cite: William A. Starms and Jeffrey K. Uhlmann, "Orthogonal Vector Interpolation for Aesthetic Image Transformations," Journal of Image and Graphics, Vol. 5, No. 2, pp. 47-51, December 2017. doi: 10.18178/joig.5.2.47-51