Machine Learning Techniques to Improve the Success Rate in In-Vitro Fertilization (IVF) Procedure

Abstract

Abstract In Vitro Fertilization (IVF) usually assesses the embryo quality by visual morphological methods to transfer the potential embryo. But the success rate of IVF still remains low because of variations in selection process. The main objective is to improve the implantation rate by predicting the quality of embryos transferred from Day-2 to Day-3. Here using the Machine Learning techniques, thousands of the images trained together for the Day-2, the selection of embryos to come for the further assessment i.e. for Day-3. This will assist the doctors to check for the quality embryo without human intervention. We have also compared the results obtained by our Artificial Intelligence methods precision of >0.98 and also generalizes the method for potential embryo selection. Around 3000 plus embryo images are trained by CNN based Azure model and the results were validated using the Machine Learning techniques. Potentially viable embryo will help improve the implantation and success rate.