Table 4.

Difference in number of annotated IRFs between Grader 1 and Grader 2 in the training set scans 1 to 4 for each vendor.

Set	Spectralis	Cirrus	Topcon	Nidek	Mean diff. ± SD (IRFs)
Training 1	1	7	48	24	20 ± 21.06
Training 2	3	8	106	95	53 ± 55.07
Training 3	21	67	46	153	71.75 ± 57.34
Training 4	8	9	33	n/a	16.67 ± 14.15
Mean diff. ± SD (IRFs)	8.25 ± 8.99	22.75 ± 29.51	58.25 ± 32.52	90.67 ± 64.61