Table 7. Overview of previously published works on the LA wall segmentation from (LGE) MRI and CT.

Study	Data	Algorithm	Evaluation	Result
Hsing et al. (2014)	55 LGE MRI	Manual	Tk	Tk = 7.0 ± 1.8 mm (before ablation) Tk = 10.7 ± 4.1 mm (after ablation)
Veni et al. (2017)	72 LGE MRI + 170 Synthetic	ShapeCut	ASD, HD, clinical evaluation	Synthetic: ASD = 0.25 ± 0.04 mm; HD = 1.95 ± 0.38 mm LGE MRI: ASD = 0.66 ± 0.14 mm LGE MRI scar segmentation: MSE = 3.07; R-square = 0.83
Zhao et al. (2017)	LGE MRI	Laplace equation	Tk	Tk = 3.7 ± 1.7 mm
Wang et al. (2019b)	154 LGE MRI + ex vivo data	Convex hull method + coupled PDE	Tk	LGE MRI: Tk = 0.4–11.7 mm and median = 3.88 mm
		Level-set		Tk = 2.16 ± 0.58 mm, Dice = 0.72
Karim et al. (2018)	10 MRI	Region growing Watershed	Tk, Dice, tissue mass	Tk = 6.04 ± 3.63 mm, Dice = 0.39 Tk = 3.45 ± 3.57 mm, Dice = 0.67
Inoue et al. (2014)	5 enhanced CT	Multi-region segmentation software + manual correction	Tk, visualization	Tk = 0.5-3.5 mm
Bishop et al. (2016)	10 CT	Morphological operations + Laplace equation	Tk	Errors ≤ 0.2 mm for Tk of 0.5–5.0 mm
Inoue and Dran-gova (2016)	10 CT	Mesh vertex normal traversal	Tk, Dice, tissue mass	Tk = 1.13 ± 1.02 mm (A), 1.26 ± 0.83 mm (P) Dice = 0.33 (A), 0.39 (P)
Tao et al. (2016b)	10 CT	Nonlinear intensity transformation + level-set	Tk, Dice, tissue mass	Tk = 1.34 ± 0.89 mm (A), 0.78 ± 0.41 mm (P) Dice = 0.43 (A), 0.21 (P)
Jia et al. (2016)	10 CT	Region growing + Marker-controlled GAC	Tk, Dice, tissue mass	Tk = 0.75 ± 0.38 mm (A), 1.46 ± 1.57 mm (P) Dice = 0.30 (A), 0.50 (P)

MSE: mean square error; A: anterior; P: posterior; Tk: thickness; GAC: geodesic active contour; PDE: partial differential equations. Note that the evaluation measures and results in Inoue and Drangova (2016), Tao et al. (2016b) and Jia et al. (2016) are from the benchmark paper (Karim et al., 2018).