. Author manuscript; available in PMC: 2024 Apr 30.

Published in final edited form as: IEEE Trans Med Imaging. 2024 Mar 5;43(3):1259–1269. doi: 10.1109/TMI.2023.3331982

Table III. Comparison of sequence generation performance among CGAN, CVAE, CVAE-GAN and the proposed model. The clinical measures derived from each real sample are compared to those derived from 20 synthetic samples of exactly the same conditions. The mean and the minimal differences of the clinical measures are reported here.

Model	d_LVEDV (mL)		d_LVESV (mL)		d_RVEDV (mL)		d_RVESV (mL)		d_LVM (g)
Model	mean	best/min	mean	best/min	mean	best/min	mean	best/min	mean	best/min
CGAN [34]	35.58_±20.33	15.66_±16.67	20.06_±9.71	19.74_±9.72	51.47_±25.25	14.71_±17.12	17.57_±12.19	17.04_±12.18	38.26_±19.15	10.40_±11.23
CVAE [48]	35.74_±16.99	4.91 _±9.84	13.92_±6.06	1.87_±3.46	44.97_±21.58	6.46 _±12.92	19.49_±9.21	2.86_±5.74	23.07_±9.96	2.70 _±4.33
CVAE-GAN [6]	51.32_±20.40	6.33_±11.96	19.80_±6.53	1.69 _±2.57	48.94_±28.66	8.28_±17.52	25.26_±10.99	2.57 _±4.11	51.03_±11.40	8.29_±7.91
Proposed	25.93 _±17.47	6.87 _±12.09	11.74 _±8.41	3.54_±6.25	34.63 _±21.31	6.88_±12.87	15.54 _±11.33	5.12_±9.19	17.34 _±9.89	2.95_±5.62