| Summary: | In Vitro Fertilization (IVF) is an Assisted Reproductive Technology (ART) that is aimed to overcome fertility problems among couples to get conceived through clinical procedure by obtaining sperm and ovum for fertilization. Selection of the fertilized ovum or embryo is crucial since the growth is visually observed for implantation. To ensure the success of implantation and healthy baby is born, selection of the best embryo quality was performed manually by at least two embryologists based on embryo morphological appearance. However, the conventional method could result in non-uniformity grading issues and increase the validation time. Therefore, this study is proposed a new grading method namely as EmbryoSys to optimize the grading and selection process of the embryo. A deep learning of Blast-Net segmentation model is trained to automatically grade and select the best embryo for implantation. The proposed method is developed in form of web-based application as the front-end with patient's medical information and IVF records database. Promising performances are reported in predicting the embryo grade from IVF microscopic images with degree of expansion has accuracy of 79%, ICM 69%, and TE 61%. The outcome from this research is expected to serve as a prognosis system for the embryologist in providing accurate grading for fertilised IVF embryo in microscopic images. © 2023 IEEE.
|
|---|