Insect Soup Challenge: Segmentation, Counting, and Simple Classification

In the paper we present a method for segmentation of insects from the Insect Soup images. The method enables reliable segmentation of insects of variable size, shape and color. After segmentation, a set of properties are assigned to each segmented insect which enables classification into different categories. The approach was successfully applied on two different types of real life images: images from the Insect Soup Challenge and images acquired from traps in the field using low resolution cameras.