TaiChiNet : Negative-Positive Cross-Attention Network for Breast Lesion Segmentation in Ultrasound Images
Breast lesion segmentation in ultrasound images is essential for computer-aided breast-cancer diagnosis. To improve the segmentation performance, most approaches design sophisticated deep-learning models by mining the patterns of foreground lesions and normal backgrounds simultaneously or by unilaterally enhancing foreground lesions via various focal losses. However, the potential of normal backgrounds is underutilized, which could reduce false positives by compacting the feature representation of all normal backgrounds. From a novel viewpoint of bilateral enhancement, we propose a negative-positive cross-attention network to concentrate on normal backgrounds and foreground lesions, respectively. Derived from the complementing opposites of bipolarity in TaiChi, the network is denoted as TaiChiNet, which consists of the negative normal-background and positive foreground-lesion paths. To transmit the information across the two paths, a cross-attention module, a complementary MLP-head, and a complementary loss are built for deep-layer features, shallow-layer features, and mutual-learning supervision, separately. To the best of our knowledge, this is the first work to formulate breast lesion segmentation as a mutual supervision task from the foreground-lesion and normal-background views. Experimental results have demonstrated the effectiveness of TaiChiNet on two breast lesion segmentation datasets with a lightweight architecture. Furthermore, extensive experiments on the thyroid nodule segmentation and retinal optic cup/disc segmentation datasets indicate the application potential of TaiChiNet.
Medienart: |
E-Artikel |
---|
Erscheinungsjahr: |
2024 |
---|---|
Erschienen: |
2024 |
Enthalten in: |
Zur Gesamtaufnahme - volume:28 |
---|---|
Enthalten in: |
IEEE journal of biomedical and health informatics - 28(2024), 3 vom: 29. März, Seite 1516-1527 |
Sprache: |
Englisch |
---|
Beteiligte Personen: |
Wang, Jinting [VerfasserIn] |
---|
Links: |
---|
Themen: |
---|
Anmerkungen: |
Date Completed 07.03.2024 Date Revised 07.03.2024 published: Print-Electronic Citation Status MEDLINE |
---|
doi: |
10.1109/JBHI.2024.3352984 |
---|
funding: |
|
---|---|
Förderinstitution / Projekttitel: |
|
PPN (Katalog-ID): |
NLM36697372X |
---|
LEADER | 01000caa a22002652 4500 | ||
---|---|---|---|
001 | NLM36697372X | ||
003 | DE-627 | ||
005 | 20240307232205.0 | ||
007 | cr uuu---uuuuu | ||
008 | 240114s2024 xx |||||o 00| ||eng c | ||
024 | 7 | |a 10.1109/JBHI.2024.3352984 |2 doi | |
028 | 5 | 2 | |a pubmed24n1319.xml |
035 | |a (DE-627)NLM36697372X | ||
035 | |a (NLM)38206781 | ||
040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
041 | |a eng | ||
100 | 1 | |a Wang, Jinting |e verfasserin |4 aut | |
245 | 1 | 0 | |a TaiChiNet |b Negative-Positive Cross-Attention Network for Breast Lesion Segmentation in Ultrasound Images |
264 | 1 | |c 2024 | |
336 | |a Text |b txt |2 rdacontent | ||
337 | |a ƒaComputermedien |b c |2 rdamedia | ||
338 | |a ƒa Online-Ressource |b cr |2 rdacarrier | ||
500 | |a Date Completed 07.03.2024 | ||
500 | |a Date Revised 07.03.2024 | ||
500 | |a published: Print-Electronic | ||
500 | |a Citation Status MEDLINE | ||
520 | |a Breast lesion segmentation in ultrasound images is essential for computer-aided breast-cancer diagnosis. To improve the segmentation performance, most approaches design sophisticated deep-learning models by mining the patterns of foreground lesions and normal backgrounds simultaneously or by unilaterally enhancing foreground lesions via various focal losses. However, the potential of normal backgrounds is underutilized, which could reduce false positives by compacting the feature representation of all normal backgrounds. From a novel viewpoint of bilateral enhancement, we propose a negative-positive cross-attention network to concentrate on normal backgrounds and foreground lesions, respectively. Derived from the complementing opposites of bipolarity in TaiChi, the network is denoted as TaiChiNet, which consists of the negative normal-background and positive foreground-lesion paths. To transmit the information across the two paths, a cross-attention module, a complementary MLP-head, and a complementary loss are built for deep-layer features, shallow-layer features, and mutual-learning supervision, separately. To the best of our knowledge, this is the first work to formulate breast lesion segmentation as a mutual supervision task from the foreground-lesion and normal-background views. Experimental results have demonstrated the effectiveness of TaiChiNet on two breast lesion segmentation datasets with a lightweight architecture. Furthermore, extensive experiments on the thyroid nodule segmentation and retinal optic cup/disc segmentation datasets indicate the application potential of TaiChiNet | ||
650 | 4 | |a Journal Article | |
700 | 1 | |a Liang, Jiafei |e verfasserin |4 aut | |
700 | 1 | |a Xiao, Yang |e verfasserin |4 aut | |
700 | 1 | |a Zhou, Joey Tianyi |e verfasserin |4 aut | |
700 | 1 | |a Fang, Zhiwen |e verfasserin |4 aut | |
700 | 1 | |a Yang, Feng |e verfasserin |4 aut | |
773 | 0 | 8 | |i Enthalten in |t IEEE journal of biomedical and health informatics |d 2013 |g 28(2024), 3 vom: 29. März, Seite 1516-1527 |w (DE-627)NLM217081614 |x 2168-2208 |7 nnns |
773 | 1 | 8 | |g volume:28 |g year:2024 |g number:3 |g day:29 |g month:03 |g pages:1516-1527 |
856 | 4 | 0 | |u http://dx.doi.org/10.1109/JBHI.2024.3352984 |3 Volltext |
912 | |a GBV_USEFLAG_A | ||
912 | |a GBV_NLM | ||
951 | |a AR | ||
952 | |d 28 |j 2024 |e 3 |b 29 |c 03 |h 1516-1527 |