Visual Computing for Industry, Biomedicine, and Art

Table 6 Quantitative comparison of segmentation performance with different methods trained on the BUSI dataset and tested on UDIAT dataset

From: Multi-task approach based on combined CNN-transformer for efficient segmentation and classification of breast tumors in ultrasound images

Model	Accuracy (%)	DC (%)	IoU (%)	Sensitivity (%)
U-Net [8]	97.08	70.02	54.45	75.57
UT-Net [55]	95.79	58.40	41.84	67.60
LinkNet [54]	96.92	73.55	59.09	89.77
TransUNet [17]	97.61	76.37	62.51	87.28
D-LinkNet [54]	97.66	77.96	64.74	88.00
Axial-DeepLab [40]	97.60	77.16	62.33	84.31
ATFE-Net [52]	97.81	78.44	65.03	85.20
Ours	97.88	81.52	70.32	90.32

Back to article page