Image processing algorithm for improving the identification of patterns on Diptera wings

Abstract

This paper presents an image processing algorithm to highlight important features from Diptera wings, as a first step towards the development of an automated analysis tool for the identification, at a species level, of certain genres of Diptera based on their wings morphological characteristics. The present algorithm includes: patterns, landmarks, and spot cells, identification through filters, edge and round geometries detectors as well as their respective centers and radii. The aim was to improve the usually manual Diptera identification task by stressing on the most important features of the wings and use this as a training tool for non-experienced users.