notesum.ai

Published at December 4

EchoONE: Segmenting Multiple echocardiography Planes in One Model

cs.CV

Released Date: December 4, 2024

Authors: Jiongtong Hu¹, Wei Zhuo¹, Jun Cheng¹, Yingying Liu², Wufeng Xue¹, Dong Ni¹

Aff.: ¹Shenzhen University; ²Shenzhen People's Hospital

Arxiv: http://arxiv.org/pdf/2412.02993v1

Refer to caption

Method	2CH			3CH			4CH			PSAX
Method	mDice	mIoU	HD95	mDice	mIoU	HD95	mDice	mIoU	HD95	mDice	mIoU	HD95
U-Net[30]	86.40	77.05	7.53	85.89	76.02	4.84	86.62	77.96	6.79	82.32	71.64	6.65
DeepLabV3+[6]	87.73	78.92	6.82	86.54	77.01	4.76	88.35	80.13	6.18	84.40	74.16	5.84
SwinUNet[1]	86.13	76.86	7.76	83.68	73.19	5.53	84.69	75.67	7.47	79.94	68.04	7.69
H2Former[11]	88.01	79.50	6.66	86.45	77.05	4.58	88.26	80.16	6.00	85.27	75.21	6.07
SAM[17]	26.24	16.01	140.01	35.30	24.04	55.30	27.22	16.75	134.79	25.10	15.53	86.53
MedSAM[23]	81.76	70.46	11.71	82.57	71.22	7.03	84.14	73.75	9.56	76.99	63.98	9.01
SAMed[42]	81.95	70.83	9.61	79.74	67.24	7.18	82.49	71.64	8.66	65.10	51.15	11.33
SAM-Med2D[7]	81.59	72.90	10.07	74.33	65.88	8.47	82.22	73.79	8.44	73.40	62.75	9.8
MSA[37]	85.88	76.00	9.04	80.50	68.18	6.59	85.41	75.69	8.30	76.77	63.65	9.54
SAMUS[21]	87.51	79.38	7.57	85.54	76.17	4.92	88.37	80.35	6.54	85.92	76.52	5.57
EchoONE	89.67	81.87	6.52	87.27	78.01	4.54	90.28	82.67	5.61	88.26	79.68	4.67