In vitro fertilization requires the selection of good sperm. Traditionally, embryologists have performed visual search under a microscope and sorted sperm manually, but the knowledge, skill, and experience of embryologists have a large impact on this process. If there were a method for efficiently searching for and retrieving sperm that was independent of the embryologist’s skill level, it would reduce the burden on the patient, embryologists, and physicians, and improve the fertilization rate.

To improve the efficiency of sperm retrieval by embryologists, we are developing a sperm sorting support system that combines a microscope system with an Augmented Reality (AR) function and real-time sperm analysis machine learning software.

The software detects good sperm based on previously learned data, analyzes their motility, and superimposes a display of the sperm that should be candidates for retrieval through the microscope eyepiece to assist the embryologist in the retrieval process. The system is designed to be compact and can be easily attached to existing microscopes.