Table 4.

The detail of datasets used for LV segmentation.

Study	Dataset
[63, 68, 69, 81, 95, 96, 102],	Sunnybrook
[57]	TWINS-UK
[61, 66, 77, 95, 97, 103, 104, 120],	ACDC
[38, 61, 66, 85, 86, 107, 121],	MICCAI
[92, 94, 100, 114]	CAMUS
[101, 104]	LVSC
[97, 101]	STACOM
[77]	(UKBB)
[81]	Cardiac Atlas
[113]	CETUS
[103]	TOF
[62]	Own dataset (220 subjects)
[75]	Own dataset (5008 subjects and 93,500 images)
[61]	Own dataset
[49]	Own dataset 1100 subjects
[49]	Own dataset 1100 subjects
[82]	Own dataset (69 images)
[83]	Own dataset (500 patients)
[84]	Own dataset (250 patients)
[91]	Own dataset (596 subjects)
[90]	Own dataset (94 cases)
[98]	York + own dataset (17,683 images)
[100]	Own dataset (1080 + 290) +
[99]	Own dataset (5000 Pocus + 16000 cart base)
[78]	Own dataset: 492 scans
[88]	Own dataset 25,027 scans (N = 12,984 patients)
[79]	Own dataset 23 sequences (670 images)
[93]	Four different datasets
[80]	Kaggle, 484 examinations
[89]	Own dataset 563 patients
[106]	143 postinfarction patients, LUMC
[111]	Own dataset (2900 2D short-axis cine MR images of 145 subjects)
[115]	Own dataset (2900 2D short-axis cine MR images of 145 subjects)
[108]	Own dataset
[64]	Own dataset (30 scans and 2000 images)
[112]	Own dataset (2900 short-axis cardiac MR images of 145 subjects)
[109]	Own dataset (900 cases)
[105]	Own dataset (127 videos)
[116]	York cardiac segmentation database
[117]	Own dataset (30 patients)
[118]	Own dataset (50 scans)
[122]	Cap a total of 2490 images from 83 subjects
[119]	Own dataset 137 patients