Table 3.

Comparison table of studies associated with predicting IVF outcomes

Paper	AI Architecture	Input data type	Training data size	CNN function	Prediction accuracy
Kan-Tor et al. (2020) [56]	DNN with Decision Tree classifier	Time-lapse videos captured using various incubator from various clinics	± 6200 labeled embryo divided into train-validation with 28% separated as test	Predicting blastocysts formation	(AUC) Training: ± 0.83 Test across age group: ± 0.75
Kan-Tor et al. (2020) [56]	DNN with Logistic Regression classifier	Time-lapse videos captured using various incubator from various clinics	± 5500 labeled embryo separated into train-validation with 21% separated as test	Predicting embryo implantation potential	(AUC) ± 0.75 average
Kanakasabapathy et al. (2020) [55]	Xception [57]	3469 recorded video from 543 patients	Training: 1190, Test: 748	Predicting blastocysts formation	± 71.87%