Abstract
In this paper, a procedure for computer-based detection of skin cancer, and in particular basal cell carcinoma (BCC), to assist dermatologists is investigated. The tissue compartments, which discriminate healthy and cancerous skins from an optical properties point of view, are studied. The application of an image-processing algorithm on a three-dimensional (3D) optical coherence tomography (OCT) image is explained. The algorithm finds the differences between healthy skin and BCC lesion by extracting scattering coefficient μs, absorption coefficient μa, and anisotropy factor g, from the 3D image of skin. We present the essential stages required to design a computer-based skin cancer detection algorithm using OCT and evaluate the performance of the algorithm using a phantom. The procedure to design the phantom and the choice of material used to model skin tissue based on BCC discriminators are discussed in detail.
Acknowledgements
We would like to thank Professor Peter Anderson, Dr David Levitz and Dr Mette Mogensen from RISO, Denmark, Professor Ian Bruce from the nanoparticle group of the University of Kent, Dr Teresa Correia from University College, London and Professor Julia Welzel from Klinikum Augsburg for their kind comments.