An image processing protocol to extract variables predictive of human embryo fitness for assisted reproduction

Abstract

ABSTRACTOBJECTIVETo study the application of image processing for segmentation of blastocysts images and extraction of potential variables for prediction of embryo fitness.DESIGNRetrospective study.SETTINGSingle reproductive medical center. IVI-RMA (Valencia, Spain) between 2017 and 2019.PATIENTSAn initial dataset including 353 images from EmbryoScope® and 474 images from Geri® incubators was acquired, of which 320 images from EmbryoScope® and 309 images from Geri® incubators were used in this study.INTERVENTION(S)None.MAIN OUTCOME MEASURE(S)Successful segmentation of images into trophectoderm (TE), blastocoel, and inner cell mass (ICM) using the proposed processing steps.RESULTSA total of 33 variables were automatically generated by digital image processing, each representing a different aspect of the embryo and describing a different characteristic of the expanding blastocyst (EX), ICM, or TE. These variables can be categorized into texture, gray level average, gray level standard deviation, modal value, relations, and light level. The automated and directed steps of the proposed processing protocol exclude spurious results, except when image quality (e.g., focus) prevents correct segmentation.CONCLUSIONSThe proposed image processing protocol that can successfully segment human blastocyst images from two distinct sources and extract 33 variables with potential utility in embryo selection for ART.