File Download

There are no files associated with this item.

  Links for fulltext
     (May Require Subscription)
Supplementary

Article: MDF-Net: A Multi-Scale Dynamic Fusion Network for Breast Tumor Segmentation of Ultrasound Images

TitleMDF-Net: A Multi-Scale Dynamic Fusion Network for Breast Tumor Segmentation of Ultrasound Images
Authors
KeywordsBreast tumor segmentation
deep supervised learning
multiscale refinement architecture
ultrasound images
Issue Date1-Jan-2023
PublisherInstitute of Electrical and Electronics Engineers
Citation
IEEE Transactions on Image Processing, 2023, v. 32, p. 4842-4855 How to Cite?
AbstractBreast tumor segmentation of ultrasound images provides valuable information of tumors for early detection and diagnosis. Accurate segmentation is challenging due to low image contrast between areas of interest; speckle noises, and large inter-subject variations in tumor shape and size. This paper proposes a novel Multi-scale Dynamic Fusion Network (MDF-Net) for breast ultrasound tumor segmentation. It employs a two-stage end-to-end architecture with a trunk sub-network for multiscale feature selection and a structurally optimized refinement sub-network for mitigating impairments such as noise and inter-subject variation via better feature exploration and fusion. The trunk network is extended from UNet++ with a simplified skip pathway structure to connect the features between adjacent scales. Moreover, deep supervision at all scales, instead of at the finest scale in UNet++, is proposed to extract more discriminative features and mitigate errors from speckle noise via a hybrid loss function. Unlike previous works, the first stage is linked to a loss function of the second stage so that both the preliminary segmentations and refinement subnetworks can be refined together at training. The refinement sub-network utilizes a structurally optimized MDF mechanism to integrate preliminary segmentation information (capturing general tumor shape and size) at coarse scales and explores inter-subject variation information at finer scales. Experimental results from two public datasets show that the proposed method achieves better Dice and other scores over state-of-the-art methods. Qualitative analysis also indicates that our proposed network is more robust to tumor size/shapes, speckle noise and heavy posterior shadows along tumor boundaries. An optional post-processing step is also proposed to facilitate users in mitigating segmentation artifacts. The efficiency of the proposed network is also illustrated on the 'Electron Microscopy neural structures segmentation dataset'. It outperforms a state-of-the-art algorithm based on UNet-2022 with simpler settings. This indicates the advantages of our MDF-Nets in other challenging image segmentation tasks with small to medium data sizes.
Persistent Identifierhttp://hdl.handle.net/10722/338275
ISSN
2023 Impact Factor: 10.8
2023 SCImago Journal Rankings: 3.556
ISI Accession Number ID

 

DC FieldValueLanguage
dc.contributor.authorQi, W-
dc.contributor.authorWu, HC-
dc.contributor.authorChan, SC-
dc.date.accessioned2024-03-11T10:27:39Z-
dc.date.available2024-03-11T10:27:39Z-
dc.date.issued2023-01-01-
dc.identifier.citationIEEE Transactions on Image Processing, 2023, v. 32, p. 4842-4855-
dc.identifier.issn1057-7149-
dc.identifier.urihttp://hdl.handle.net/10722/338275-
dc.description.abstractBreast tumor segmentation of ultrasound images provides valuable information of tumors for early detection and diagnosis. Accurate segmentation is challenging due to low image contrast between areas of interest; speckle noises, and large inter-subject variations in tumor shape and size. This paper proposes a novel Multi-scale Dynamic Fusion Network (MDF-Net) for breast ultrasound tumor segmentation. It employs a two-stage end-to-end architecture with a trunk sub-network for multiscale feature selection and a structurally optimized refinement sub-network for mitigating impairments such as noise and inter-subject variation via better feature exploration and fusion. The trunk network is extended from UNet++ with a simplified skip pathway structure to connect the features between adjacent scales. Moreover, deep supervision at all scales, instead of at the finest scale in UNet++, is proposed to extract more discriminative features and mitigate errors from speckle noise via a hybrid loss function. Unlike previous works, the first stage is linked to a loss function of the second stage so that both the preliminary segmentations and refinement subnetworks can be refined together at training. The refinement sub-network utilizes a structurally optimized MDF mechanism to integrate preliminary segmentation information (capturing general tumor shape and size) at coarse scales and explores inter-subject variation information at finer scales. Experimental results from two public datasets show that the proposed method achieves better Dice and other scores over state-of-the-art methods. Qualitative analysis also indicates that our proposed network is more robust to tumor size/shapes, speckle noise and heavy posterior shadows along tumor boundaries. An optional post-processing step is also proposed to facilitate users in mitigating segmentation artifacts. The efficiency of the proposed network is also illustrated on the 'Electron Microscopy neural structures segmentation dataset'. It outperforms a state-of-the-art algorithm based on UNet-2022 with simpler settings. This indicates the advantages of our MDF-Nets in other challenging image segmentation tasks with small to medium data sizes.-
dc.languageeng-
dc.publisherInstitute of Electrical and Electronics Engineers-
dc.relation.ispartofIEEE Transactions on Image Processing-
dc.rightsThis work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.-
dc.subjectBreast tumor segmentation-
dc.subjectdeep supervised learning-
dc.subjectmultiscale refinement architecture-
dc.subjectultrasound images-
dc.titleMDF-Net: A Multi-Scale Dynamic Fusion Network for Breast Tumor Segmentation of Ultrasound Images-
dc.typeArticle-
dc.identifier.doi10.1109/TIP.2023.3304518-
dc.identifier.scopuseid_2-s2.0-85169504197-
dc.identifier.volume32-
dc.identifier.spage4842-
dc.identifier.epage4855-
dc.identifier.eissn1941-0042-
dc.identifier.isiWOS:001060337200001-
dc.identifier.issnl1057-7149-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats