Table 3.

Evaluation of fibrous cap thickness, with absolute error (mean $\pm$ standard deviation), bias, $95\%$ limits of agreement (Lim), and Pearson coefficient (R), for the automatic method (Auto) and the two analysts (${\mathcal{A}}_1$ and ${\mathcal{A}}_2$)

Errors $(\mathrm{\mu }\text{m})$	Overall cap thickness over the entire ROI				Minimal cap thickness per frame
	Absolute	Bias	Lim	R	Absolute	Bias	Lim	R
Testing set $({\mathit{\Omega }}_2=21)$
Auto versus ${\mathcal{A}}_1$	30 $\pm$ 37	1.4	$[-92,95]$	0.85	22 $\pm$ 18	8.4	$[-46,63]$	0.73
Auto versus ${\mathcal{A}}_2$	36 $\pm$ 41	4.9	$[-101,111]$	0.81	26 $\pm$ 22	4.6	$[-61,70]$	0.62
${\mathcal{A}}_1$ versus ${\mathcal{A}}_2$	36 $\pm$ 41	3.6	$[-102,109]$	0.82	21 $\pm$ 19	$-$3.8	$[-59,52]$	0.74
Training set $({\mathit{\Omega }}_1=10)$
Auto versus ${\mathcal{A}}_1$	31 $\pm$ 39	$-$1.4	$[-99,96]$	0.86	30 $\pm$ 27	2.4	$[-77,82]$	0.82
Auto versus ${\mathcal{A}}_2$	32 $\pm$ 33	$-$1.8	$[-91,87]$	0.87	29 $\pm$ 27	2.0	$[-75,79]$	0.85
${\mathcal{A}}_1$ versus ${\mathcal{A}}_2$	35 $\pm$ 41	$-$0.4	$[-105,105]$	0.84	24 $\pm$ 25	$-$0.4	$[-68,67]$	0.89