“Colour Image Processing with Hypercomplex Algebra” (YouTube movie*) by Dr. Steve Sangwine (School of Computer Science and Electronic Engineering, University of Essex, United Kingdom), invited presentation on 9th March 2015 (Mo.) 10:30-11:30am, at Kogakuin University, Tokyo, Japan, hosted by Ass. Prof. Dr. Kanta Tachibana. Recorded by Sen. Ass. Prof. Eckhard Hitzer, International Christian University, Tokyo, Japan.
*NOTE: Beginning – 1min25sec … Japanese Introduction (Dr. K. Tachibana), 1min25sec – end … English Presentation (Dr. S. Sangwine).
Presentation slides: http://privatewww.essex.ac.uk/~sjs/research/Slides_ICU_March_2015.pdf
In presentation videos:
1) Lena: A Fourier decomposition based on hypercomplex transforms.
2) Planar Ellipse Movie by Dr. Steve Sangwine, University of Essex, UK
Related textbook co-authored by speaker: Todd A. Ell, Nicolas Le Bihan, Stephen J. Sangwine, Quaternion Fourier Transforms for Signal and Image Processing, Series FOCUS Digital Signal and Image Processing Series, iSTE/Wiley, London and Hoboken, 2014.
Abstract: The speaker has worked on colour image processing since 1992, and has studied hypercomplex algebras (principally quaternions) applied to colour images since 1996. The talk will provide an overview of colour science as it applies to image processing, to explain what colour is, and how it is represented electronically, and measured. This part of the talk is based on presentations that the author has given to scientifically literate but not research-level audiences.
The second part of the talk will introduce the idea of quaternion Fourier transforms and the way they represent colour images, in particular in terms of oscillations within a three-dimensional ‘colour space’. Recent work on this has shown that simple harmonic motion in any number of dimensions consists of elliptical oscillation in a plane, and this finding has ramifications for both quaternion and other types of hypercomplex transforms, indeed any Fourier domain representation of an oscillation in N-dimensions.