Automatic Segmentation of Zona Pellucida in Human Embryo Images Applying an Active Contour Model

When evaluating the feasibility of a human embryo to implant in a woman uterus on In-Vitro Fertilization (IVF), one of the most important morphological parameters taken into account in the embryo selection process is the thickness of zona pellucida (i.e. the external layer of the embryo). Furthermore, the measurement of zona pellucida thickness variation is a potentially reliable parameter to evaluate the ability of embryo hatching and, it is directly related with the implantation rate. In ordinary embryo evaluation procedures this thickness is measured by human embryologists, which leads to a simple classification between thin, medium and thick values. Aiming at improving this analysis due to the potential loss of information that this implies, we introduce in this paper a vision and segmentation automatic system system which identifies on embryo images the external and internal membranes of zona pellucida applying the active contour algorithm, which allows to record and quantify the thickness variation of zona pellucida. We compare the performance of our system to manual segmentation of an embryologist expert to show the validity of our approach.