Hexagonal Image Processing for Computer Vision With Hexnet: A Hexagonal Image Processing Data Set and Generator

In the domains of image processing and computer vision, the exploration of hexagonal image processing systems has emerged as a fundamentally innovative yet nascent methodology that is motivated by the occurrence of hexagonal structures in the human visual perception system and nature itself. However, despite the possible benefits of hexagonal over conventional square approaches for image processing systems—which commonly utilize square pixels—no known publicly available hexagonal image data sets exist that would enable the evaluation of hexagonal approaches that have been developed within image processing and computer vision for tasks such as object detection and classification. For this purpose, this contribution proposes a foundation for hexagonal image data sets and their development: The Hexnet Hexagonal Image Processing Data Set (short Hexnet Dataset), which is based on The Hexagonal Image Processing Framework Hexnet (Hexnet Framework). As a baseline, three data subsets are introduced: 1) geometric primitives for the evaluation of hexagonal structures; 2) astronomical image processing, in which the descriptions of sensory elements of hexagonal telescope arrays have been leveraged for the detection and classification of synthesized atmospheric events; and 3) conventional image data sets, which provides hexagonally transformed versions of commonly evaluated square imagery.