EG-TransUNet: a transformer-based U-Net with enhanced and guided models for biomedical image segmentation

BMC Bioinformatics

Table 3 Comparisons with the state-of-the-art baselines on the 2018 data science bowl (DSB) dataset

Method	Year	mDice	mIoU	Recall	Precision
U-Net [5]	2015	0.7573	0.9103	–	–
PraNet [48]	2020	0.8103	0.7108	0.8062	0.8231
U-Net++ [7]	2018	0.8974	0.9255	–	–
DoubleU-Net [25]	2020	0.9133	0.8407	0.6407	0.9406
TransAttUnet_R [13]	2021	0.9162	0.8498	0.9185	0.9193
DS-TransUNet-B [14]	2021	0.9200	0.8589	0.9427	0.9054
DS-TransUNet-L [14]	2021	0.9219	0.8612	0.9378	0.9124
MSRF-Net [41]	2021	0.9224	0.8534	0.9402	0.9022
EG-TransUNet	–	0.9349	0.8908	0.9482	0.9336

The “–” denotes the corresponding result is not provided. For each column, the best results are highlighted

ISSN: 1471-2105