A computer imaging system for noninvasively determining blood pool volume in the chick embryo heart

This paper introduces a noninvasive semi-autonomous computer imaging based system for quantifying ventricular blood pool volume in embryonic chick models. The technique proceeds by first obtaining two reference images from a recorded sequence of the functioning chick heart. These reference images, which are representative of the maximum end diastolic and the minimum end systolic phases for a set of images, are used as the basis for automating subsequent operations. A noise reduction filter, an automatic thresholding algorithm, and region growing are applied to selected end phase images in order to determine the extent of the blood pool in the ventricle. The area of the isolated ventricular blood pool in each of the resulting binary images is computed and used to calculate the volume. A statistical analysis of the data obtained by this system indicates substantial improvement over previous methods