Evaluation of deep convolutional neural networks in classifying human embryo images based on their morphological quality

• DOI: 10.1016/j.heliyon.2021.e06298
• Publication Date: 2021-02-23T09:03:30Z
• Authors: Prudhvi Thirumala...; Manoj Kumar Kanak...; Charles L. Bormann; Raghav Gupta; Rohan Pooniwala; Hemanth Kandula; Irene Souter; Irene Dimitriadis; Hadi Shafiee
• Keywords: H1-99; FOS: Computer and information sciences; Computer Science - Machine Learning; Science (General); Computer Vision and Pattern Recognition (cs.CV); Image and Video Processing (eess.IV); Computer Science - Computer Vision and Pattern Recognition; Electrical Engineering and Systems Science - Image and Video Processing; In-vitro fertilization; Machine Learning (cs.LG); Social sciences (General); Q1-390; 03 medical and health sciences; 0302 clinical medicine; Human embryos; Deep neural networks; FOS: Electrical engineering, electronic engineering, information engineering; Convolutional neural networks; Research Article

A critical factor that influences the success of an in-vitro fertilization (IVF) treatment cycle is quality transferred embryo. Embryo morphology assessments, conventionally performed through manual microscopic analysis suffer from disparities in practice, selection criteria, and subjectivity due to experience embryologist. Convolutional neural networks (CNNs) are powerful, promising algorithms with significant potential for accurate classifications across many object categories. Network architectures hyper-parameters affect efficiency CNNs any given task. Here, we evaluate multi-layered developed scratch popular deep-learning such as Inception v3, ResNET-50, Inception-ResNET-v2, NASNetLarge, ResNeXt-101, ResNeXt-50, Xception differentiating between embryos based on their morphological at 113 h post insemination (hpi). best quality.