Evolutionary snake algorithms for biological applications

Abstract

The automated segmentation of moving objects in image sequences is becoming increasingly important and has numerous applications to time-lapse imaging in medical and biological research. With advances in techniques such as fluorescence microscopy, the amount of medical image data to be analyzed is ever increasing and so is the demand for segmentation software. Broadly, there are two methods for tracking objects: tracking by detection and tracking by model evolution. The first method calculates centroids of specific objects and eliminates the background with a threshold based algorithm. The second one is more advanced and uses active contour/surface models that are optimal with respect to a chosen energy function. Active contours, also known as snakes, have been investigated and optimized for image segmentation in two dimensions. However, there is little or no research on snakes that evolve in time. The purpose of this project is to develop such a technique and demonstrate its performance on real data.

Legend