Comparison of Segmentation Algorithms for the Zebrafish Heart in Fluorescent Microscopy Images

The zebrafish embryo is a common model organism for cardiac development and genetics. However, the current method of analyzing the embryo heart images is still mainly the manual and visual inspection through the microscope by scoring embryos visually a very laborious and expensive task for the biologist. We propose to automatically segment the embryo cardiac chambers from fluorescent microscopic video sequences, allowing morphological and functional quantitative features of cardiac activity to be extracted. Several methods are presented and compared within a large range of images, varying in quality, acquisition parameters, and embryos position. Despite such variability in the images, the best method reaches a 70% of accuracy, allowing reducing biologists workload by automating some of the tedious manual segmentation tasks.