Abstract
The process of detecting and tracking biological features such as bacteria and nuclei is complicated by the fact that they constantly change their shape. Shape changes happen both continuously as the biological features grow and discontinuously as they divide or die. In this paper I present a new method of tracking such features for the case that they can be reasonably approximated by a relatively simple mathematical shape such as a cylinder or an ellipse. Using contour plots with multiple levels to detect the features and their shapes, rather than the commonly used single contour detection technique, this method can efficiently detect multiple features even if they have large differences in brightness, as well as reliably track divisions when both brightness and size drop dramatically.
