Automatic Detection and Surface Measurements of Micronucleus by a Computer Vision Approach

This paper reports a novel method for nucleus and micronucleus segmentation. These biological structures are useful to biologists for relieving structural chromosome aberration. The adopted method consists of a pipeline of advanced computer vision algorithms, some of them were specifically tailored for the current segmentation problem. Starting from the weak hypotheses on size, shape, and color of micronucleus, it is possible to efficiently segment and measure the image features of interest by a com- puter vision approach. We report experimental results with a new flow cytometer architecture specifically developed to recognize and measure micronucleus of human lymphocyte. The robustness of the algorithm with respect to various kinds of noise is also reported.