Automated comet assay segmentation using combined dot enhancement filters and extended-maxima transform watershed segmentation

Abstract

The comet assay, also known as single-cell gel electrophoresis, is a widely used and trustworthy method for assessing DNA damage and repair in individual cells. It plays a crucial role in assessment of genetic damage potential and human biomonitoring studies in the medical and biological fields. Ensemble of comet assay individual cells and establishing accurate information on the occurrence of cellular injury followed by the process of cellular restoration is a challenging task. This paper introduces an algorithm for the detection of a distinct head, composed of undamaged DNA, and a tail, comprising damaged or fragmented DNA, in genotoxicity testing images, and provides information on the region properties of such images. The proposed approach combines a dot enhancement filter to distinguish and help in the detection of the head in each cell combined with a multilevel segmentation approach consisting of a watershed-geodesic active contour model that is capable to refine the tail estimation. The effectiveness of the suggested algorithm is quantitatively evaluated with annotation data provided by biologists, and its results are compared with those obtained from previous works. The proposed system exhibits comparable or superior performance to the existing systems while avoiding excessive computational cost.