Classification of Caenorhabditis Elegans Behavioural Phenotypes Using an Improved Binarization Method

Abstract

Because of simple model organisms, Caenorhabditis (C.) elegans is often used in genetic analysis in neuroscience. The classification and analysis of C. elegans was previously performed subjectively. So the result of classification is not reliable and often imprecise. For this reason, automated video capture and analysis systems appeared. In this paper, we propose an improved binarization method using a hole detection algorithm. Using our method, we can preserve the hole and remove the noise, so that the accuracy of features is improved. In order to improve the classification success rate, we add new feature sets to the features of previous work. We also add 3 more mutant types of worms to the previous 6 types, and then analyze their behavioural characteristics.