Automated machine learning for the identification of asymptomatic COVID-19 carriers based on chest CT images

BMC Medical Imaging

Table 1 Differences in image preprocessing among the three authors using the Kruskal‒Wallis H test and ICC analysis

Texture	Features	ICC	95%CI		P _{Kruskal−Wallis H test}
Grey histogram	Grayscale mean	0.959	0.865	0.985	0.5069
	Contrast mean	0.908	0.779	0.963	0.2446
	R	0.910	0.778	0.964	0.2446
	Third moment	0.961	0.909	0.984	0.6608
	Consistency	0.975	0.921	0.991	0.5838
	Entropy	0.966	0.899	0.988	0.5375
Grey level co-occurrence matrix	Max-probability	0.974	0.913	0.991	0.5871
	Contrast	0.883	0.769	0.948	0.6691
	Correlation	0.822	0.655	0.920	0.4459
	Energy	0.976	0.924	0.991	0.5871
	Homogeneity	0.970	0.903	0.989	0.6326
	Entropy	0.968	0.902	0.988	0.5730
Gabor filter	Texture mean	0.995	0.983	0.998	0.8384
	Contrast	0.994	0.985	0.998	0.8257
	Entropy	0.994	0.980	0.998	0.8777
Gauss Markov random field	GMRF1	0.938	0.875	0.973	0.9784
	GMRF2	0.935	0.862	0.972	0.6458
	GMRF3	0.785	0.611	0.900	0.7857
	GMRF4	0.852	0.716	0.933	0.5980
	GMRF5	0.752	0.560	0.883	0.8573
	GMRF6	0.654	0.420	0.830	0.8505
	GMRF7	0.710	0.501	0.860	0.6711
	GMRF8	0.804	0.639	0.910	0.9584
	GMRF9	0.817	0.661	0.916	0.8218
	GMRF10	0.896	0.798	0.954	0.9863
	GMRF11	0.782	0.605	0.899	0.8561
	GMRF12	0.775	0.592	0.895	0.8077
Tamura	Coarseness	0.991	0.982	0.996	0.9472
	Contrast	0.959	0.895	0.984	0.5367
	Directionality	0.754	0.564	0.884	0.8448
	Line-likeness	0.900	0.788	0.958	0.6008
	Roughness	0.963	0.906	0.986	0.5375

ISSN: 1471-2342