Towards the automation of early-stage human embryo development detection

Table 4 Average classification accuracy

Fold number	VGG architecture					Avg	AlexNet architecture					Avg
	Embryo development stages						Embryo development stages
	1	2	3	4	> 4		1	2	3	4	> 4
1	0.928	0.924	0.904	0.939	0.923	0.923	0.910	0.927	0.903	0.913	0.911	0.912
2	0.940	0.956	0.947	0.979	0.938	0.952	0.924	0.951	0.944	0.949	0.916	0.936
3	0.933	0.927	0.925	0.950	0.931	0.933	0.912	0.947	0.915	0.921	0.911	0.921
4	0.951	0.961	0.955	0.958	0.925	0.950	0.933	0.935	0.932	0.920	0.937	0.931
5	0.943	0.919	0.912	0.914	0.902	0.918	0.916	0.938	0.899	0.937	0.911	0.920
6	0.949	0.947	0.910	0.944	0.927	0.935	0.929	0.931	0.921	0.934	0.938	0.930
7	0.929	0.932	0.921	0.934	0.914	0.926	0.931	0.962	0.939	0.928	0.917	0.935
8	0.935	0.958	0.952	0.959	0.933	0.947	0.921	0.921	0.930	0.956	0.932	0.932
9	0.945	0.932	0.913	0.957	0.907	0.931	0.926	0.935	0.908	0.933	0.902	0.920
10	0.926	0.956	0.953	0.968	0.924	0.945	0.918	0.953	0.909	0.937	0.944	0.932
Average	0.938	0.941	0.929	0.950	0.922	0.936	0.922	0.940	0.920	0.933	0.922	0.927
St. deviation	0.009	0.016	0.021	0.018	0.011	0.012	0.008	0.013	0.016	0.013	0.014	0.008

ISSN: 1475-925X