Deformable templates for recognizing the shape of the zebra fish egg cell

Abstract

In biology, manipulating a micro-scale object such as chromosome, nucleus or embryo has been an important issue. For instance, skillful manipulation of the embryo cell in the biological experiment requires many years experience with a complex setup. Moreover, such process is usually very slow and requires many hours of intense operations such as trying to find the position of the cell within a petri dish and injecting a pipette to the cell from the best orientation. We have designed a new vision system, by which it finds the region of the zebra fish egg cell, and then tracks the yolk as well as the inner line that divides between protoplasm and vitellin within the egg cell, using two different deformable templates. Result suggests that recognition performance of the system varies depending upon the magnification ratio of the microscope, and it reaches to the maximum at 80:1.