Automated Segmentation Method for Low Field 3D Stomach MRI Using Transferred Learning Image Enhancement Network

Table 7

Comparison of segmentation results of different image enhancement methods.

Method	Metric	#patient				Mean
		1	2	3	4
Traditional_based	DC	0.643	0.853	0.621	0.544	0.665
	Haus	9	8.062	6.708	8.485	8.064
	Sen	0.494	0.781	0.475	0.403	0.538
	Spe	0.999	0.999	0.999	0.999	0.999
	OR	0.039	0.047	0.052	0.073	0.053
	UR	0.486	0.208	0.497	0.553	0.436
CycleGAN_based	DC	0.684	0.865	0.681	0.532	0.690
	Haus	9	7.874	6.708	8	7.895
	Sen	0.535	0.814	0.551	0.509	0.602
	Spe	0.999	0.999	0.999	0.999	0.999
	OR	0.029	0.064	0.064	0.026	0.046
	UR	0.451	0.174	0.419	0.478	0.381