Table 2.

Results of using architectures without pre-trained weights in DeemEmbryo and SI-DeepEmbryo.

DeepEmbryo Type	AlexNet	ResNet18	ResNet34	Inception V3	DenseNet121
SI-DeepEmbryo	Not learned	Not learned	Not learned	58.33%	56.77%
DeepEmbryo	Not learned	58.33%	Overfit	66.67%	Overfit