Image Processing in an Automated Live Cell Imaging System

In vitro embryo production and selection is a key step of infertility treatment. Time-lapse technology was introduced in embryology in order to increase the efficiency of embryo selection. This paper describes the image processing algorithms used to detect embryos and the accommodating microwells in an automated time-lapse imaging system specialized for this purpose. The system consists of compact digital inverted microscope units, microwell culture dishes and a controller, and includes software for controlling the time-lapse cycles and managing the visualization of the observed cells/tissues. The purpose of the described algorithms is to identify the regions of interest on the photos: the well detection algorithm detects the microwells of the applied culture dish holding the embryos, the embryo detection algorithm locates the embryos within the microwells. The algorithms proved to be robust during test runs and were integrated to the software system, which allows providing views for the users automatically that show only the relevant regions of the images taken by the microscope units.