Event № 1071
Topic: MSc Seminar Time: Jun 4, 2020 11:00 AM Jerusalem
Join Zoom Meeting https://technion.zoom.us/j/94090731867
Meeting ID: 940 9073 1867
Advisor: Savir Yonatan
Abstract: Assisted reproductive technology procedures, performed in the US and Europe, have increased by more than 30% in the last decade. Selecting embryos with the highest potential for implantation is a great challenge in current treatments. Several criteria exist to correctly validate the embryos, all relying on morphological aspects such as growth rate, degree of fragmentation and blastocyst formation. As manual approaches are labor-intensive and time-consuming, a considerable effort was dedicated in the last decade to the standardization and automation of the embryo scoring, however, with limited success. In this work, we developed a decision support system, based on two deep residual networks and computer vision techniques, that allows inferring the embryo state in real-time, extract morphological features of the embryos and predict blastocyst development potential. This platform will aid in improving artificial reproduction success rate and reducing its cost.