Abstract

The Spiral Architecture has been developed as a fast way ofindexing a hexagonal pixel-based image. In combinationwith spiral addition and spiral multiplication, methods havebeen developed for hexagonal image processing operationssuch as translation and rotation. Using the SpiralArchitecture as the basis for our operator structure, wepresent a general approach to the computation of adaptivecoarse scale Laplacian operators for use on hexagonal pixelbasedimages. We evaluate the proposed operators usingsimulated hexagonal images and demonstrate improvedperformance when compared with rectangular Laplacianoperators such as Marr-Hildreth.