Table 2.

Statistical analysis between v16pU-Net embedded with learning scheme A and all other configurations (U-Net (Ronneberger et al., 2015) with P, HP and A as well as v16U-Net with A) through Student’s paired t-tests using Dice, sensitivity, specificity, Jaccard, Cohen’s kappa scores as well as absolute surface error over the pathological dataset. Bold p-values (<0.05) highlight statistically significant results.

Metric	Scheme	P	HP	A
	Network	U-Net (Ronneberger et al., 2015)			v16U-Net
dice	Deltoid	3.7×10−29	8.8×10−21	9.7×10−11	7.5×10−6
	Infraspinatus	2.3×10−19	5.4×10−7	0.491	0.935
	Supraspinatus	4.3×10−6	4.8×10−5	3.2×10−6	7.0×10−5
	subscapularis	1.1×10−13	1.5×10−7	0.001	0.006
sens	Deltoid	1.2×10−23	3.7×10−23	2.2×10−9	4.4×10−4
	Infraspinatus	6.1×10−16	3.9×10−4	0.117	0.788
	Supraspinatus	8.6×10−4	8.7×10−5	0.016	0.135
	subscapularis	2.7×10−12	2.8×10−10	0.002	1.5×10−6
spec	Deltoid	7.9×10−9	2.6×10−7	0.457	0.069
	Infraspinatus	1.3×10−5	1.2×10−6	0.010	0.387
	Supraspinatus	8.2×10−6	0.118	1.0×10−5	7.8×10−9
	subscapularis	0.924	0.005	0.078	2.5×10−4
jacc	Deltoid	5.1×10−35	9.9×10−25	7.0×10−11	1.1×10−5
	Infraspinatus	1.6×10−23	1.6×10−9	0.251	0.917
	Supraspinatus	4.9×10−10	8.3×10−7	1.8×10−7	1.7×10−6
	subscapularis	1.2×10−17	1.0×10−9	2.6×10−4	0.002
kappa	Deltoid	2.6×10−29	7.9×10−21	8.9×10−11	6.8×10−6
	Infraspinatus	1.7×10−19	4.6×10−7	0.476	0.928
	Supraspinatus	3.4×10−6	4.3×10−5	2.9×10−6	5.8×10−5
	subscapularis	9.5×10−14	1.4×10−7	0.001	0.006
ASE	Deltoid	2.2×10−15	7.9×10−13	0.001	0.021
	Infraspinatus	8.4×10−10	0.021	0.308	0.813
	Supraspinatus	7.3×10−5	8.2×10−4	0.004	0.036
	subscapularis	0.009	0.009	0.005	0.011