View in EDS

MSBP-Net: A multi-scale boundary prediction network for automated polyp segmentation.

Saved in:
Bibliographic Details
Title: MSBP-Net: A multi-scale boundary prediction network for automated polyp segmentation.
Authors: Pan, Xing-Liang1,2 (AUTHOR), Ding, Ju-Rong1,2 (AUTHOR) jurongding@gmail.com, Li, Xia1,2 (AUTHOR), Liu, Shuo1,2 (AUTHOR), Wang, Jie1,2 (AUTHOR), Hua, Bo1,2 (AUTHOR), Tang, Guo-Zhi1,2 (AUTHOR), Zhong, Chang-Hua1,2 (AUTHOR)
Source: Pattern Recognition. Feb2026, Vol. 170, pN.PAG-N.PAG. 1p.
Subjects: Colon polyps, Image segmentation, Machine learning, Deep learning, Image processing, Colorectal cancer, Edge detection (Image processing)
Abstract: Accurate polyp segmentation is a critical step in the early diagnosis of colorectal cancer. Predicting the polyp boundaries with powerful camouflage properties is an intricate challenge in automatic segmentation tasks. Here, we propose a multi-scale boundary prediction network (MSBP-Net) for effective polyp segmentation with low complexity. The MSBP-Net consists of a pre-trained pyramid vision transformer and a novel lightweight decoder which contains three modified receptive field blocks (mRFBs), three boundary prediction modules (BPMs) and a shallow filtering module (SFM). Firstly, the mRFBs are developed to suppress redundant information and irrelevant background. Secondly, the BMPs are built based on reverse attention and multi-scale criss-cross attention to efficiently explore boundaries and fuse multi-scale information for recovering coarse-grained polyp masks. Then, the SFM is proposed based on Laplacian edge operators, which mines boundary cues in the shallow layers as a supplement to the coarse-grained masks and finally gains fine-grained segmentation masks. We conduct extensive comparative experiments on five public datasets. The ablation results prove the effectiveness of the proposed modules, achieving a highest dice similarity coefficient of 0.940. Compared to existing state-of-the-art methods, the MSBP-Net performs close to the optimal methods on the five test sets, while reducing the complexity by at least 20.2 %. The lightweight design of the MSBP-Net achieves a speed of 41 frames per second on a 3070 GPU (8 GB memory). Therefore, the MSBP-Net is a promising polyp segmentation method with robust performance and low complexity, and has the potential for real-time segmentation tasks. [ABSTRACT FROM AUTHOR]
Copyright of Pattern Recognition is the property of Pergamon Press - An Imprint of Elsevier Science and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
FullText Text:
  Availability: 0
Header DbId: egs
DbLabel: Engineering Source
An: 187839855
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: MSBP-Net: A multi-scale boundary prediction network for automated polyp segmentation.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Pan%2C+Xing-Liang%22">Pan, Xing-Liang</searchLink><relatesTo>1,2</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Ding%2C+Ju-Rong%22">Ding, Ju-Rong</searchLink><relatesTo>1,2</relatesTo> (AUTHOR)<i> jurongding@gmail.com</i><br /><searchLink fieldCode="AR" term="%22Li%2C+Xia%22">Li, Xia</searchLink><relatesTo>1,2</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Liu%2C+Shuo%22">Liu, Shuo</searchLink><relatesTo>1,2</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Wang%2C+Jie%22">Wang, Jie</searchLink><relatesTo>1,2</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Hua%2C+Bo%22">Hua, Bo</searchLink><relatesTo>1,2</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Tang%2C+Guo-Zhi%22">Tang, Guo-Zhi</searchLink><relatesTo>1,2</relatesTo> (AUTHOR)<br /><searchLink fieldCode="AR" term="%22Zhong%2C+Chang-Hua%22">Zhong, Chang-Hua</searchLink><relatesTo>1,2</relatesTo> (AUTHOR)
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%22Pattern+Recognition%22">Pattern Recognition</searchLink>. Feb2026, Vol. 170, pN.PAG-N.PAG. 1p.
– Name: Subject
  Label: Subjects
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Colon+polyps%22">Colon polyps</searchLink><br /><searchLink fieldCode="DE" term="%22Image+segmentation%22">Image segmentation</searchLink><br /><searchLink fieldCode="DE" term="%22Machine+learning%22">Machine learning</searchLink><br /><searchLink fieldCode="DE" term="%22Deep+learning%22">Deep learning</searchLink><br /><searchLink fieldCode="DE" term="%22Image+processing%22">Image processing</searchLink><br /><searchLink fieldCode="DE" term="%22Colorectal+cancer%22">Colorectal cancer</searchLink><br /><searchLink fieldCode="DE" term="%22Edge+detection+%28Image+processing%29%22">Edge detection (Image processing)</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Accurate polyp segmentation is a critical step in the early diagnosis of colorectal cancer. Predicting the polyp boundaries with powerful camouflage properties is an intricate challenge in automatic segmentation tasks. Here, we propose a multi-scale boundary prediction network (MSBP-Net) for effective polyp segmentation with low complexity. The MSBP-Net consists of a pre-trained pyramid vision transformer and a novel lightweight decoder which contains three modified receptive field blocks (mRFBs), three boundary prediction modules (BPMs) and a shallow filtering module (SFM). Firstly, the mRFBs are developed to suppress redundant information and irrelevant background. Secondly, the BMPs are built based on reverse attention and multi-scale criss-cross attention to efficiently explore boundaries and fuse multi-scale information for recovering coarse-grained polyp masks. Then, the SFM is proposed based on Laplacian edge operators, which mines boundary cues in the shallow layers as a supplement to the coarse-grained masks and finally gains fine-grained segmentation masks. We conduct extensive comparative experiments on five public datasets. The ablation results prove the effectiveness of the proposed modules, achieving a highest dice similarity coefficient of 0.940. Compared to existing state-of-the-art methods, the MSBP-Net performs close to the optimal methods on the five test sets, while reducing the complexity by at least 20.2 %. The lightweight design of the MSBP-Net achieves a speed of 41 frames per second on a 3070 GPU (8 GB memory). Therefore, the MSBP-Net is a promising polyp segmentation method with robust performance and low complexity, and has the potential for real-time segmentation tasks. [ABSTRACT FROM AUTHOR]
– Name: AbstractSuppliedCopyright
  Label:
  Group: Ab
  Data: <i>Copyright of Pattern Recognition is the property of Pergamon Press - An Imprint of Elsevier Science and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=187839855
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1016/j.patcog.2025.112101
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 1
        StartPage: N.PAG
    Subjects:
      – SubjectFull: Colon polyps
        Type: general
      – SubjectFull: Image segmentation
        Type: general
      – SubjectFull: Machine learning
        Type: general
      – SubjectFull: Deep learning
        Type: general
      – SubjectFull: Image processing
        Type: general
      – SubjectFull: Colorectal cancer
        Type: general
      – SubjectFull: Edge detection (Image processing)
        Type: general
    Titles:
      – TitleFull: MSBP-Net: A multi-scale boundary prediction network for automated polyp segmentation.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Pan, Xing-Liang
      – PersonEntity:
          Name:
            NameFull: Ding, Ju-Rong
      – PersonEntity:
          Name:
            NameFull: Li, Xia
      – PersonEntity:
          Name:
            NameFull: Liu, Shuo
      – PersonEntity:
          Name:
            NameFull: Wang, Jie
      – PersonEntity:
          Name:
            NameFull: Hua, Bo
      – PersonEntity:
          Name:
            NameFull: Tang, Guo-Zhi
      – PersonEntity:
          Name:
            NameFull: Zhong, Chang-Hua
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 02
              Text: Feb2026
              Type: published
              Y: 2026
          Identifiers:
            – Type: issn-print
              Value: 00313203
          Numbering:
            – Type: volume
              Value: 170
          Titles:
            – TitleFull: Pattern Recognition
              Type: main
ResultId 1