Biologically Inspired Hexagonal Image Structure for Computer Vision

Representation of digital images in hexagonal grid has been under investigation for past 40 years. The common digital image representation makes use of rectangular grid of square-structured pixel elements known as pixels. Hence, normally, rectangular grid for displaying and processing of images is used. Advanced processing power of modern graphic devices and innovative advancements incorporated in charge-coupled device (CCD) technology have made hexagonal sampling lattice a more interesting area for research. The hexagonal grid is superior to square structure due to high packing density, equidistant pixels, higher symmetry, less aliasing, good angular resolution and consistent connectivity. In addition to these benefits, an interesting inspiration behind using a hexagonal sampled lattice image is inspired from the human visual perception as this structure closely resembles human eye. Also, hexagon structure encompasses more area than any other closed planar geometry of identical perimeter rather than a circle. Hence, this structure will have high sampling density. Since there are no full-fledged hardware devices for capturing and displaying hexagonal image structure, image conversions have to be done as the preliminary step before proceeding with hexagonal image processing. This research review will give an overview of three different methods for hexagonal image sampling and various hexagonal software simulation schemes simulated images. Finally, to show the computational efficiency, run times of different methods are taken for different sized images.