Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated 47 min ago
Peer-reviewedCC BYScientific Reports

Boundary-interior ratio: a spatially decomposed evaluation framework for CNN and mamba-based breast ultrasound segmentation

  • Aymen M. Al-Hejri1
  • Wad Ghaban2
  • Sultan S. Alshamrani3
  • Ahmed S. AlGhamdi4
  • Ebrahim Mohammed Senan5
5 affiliations
  1. 1Faculty of Administrative and Computer Sciences, University of Albaydha
  2. 2Department of Computer Science, Faculty of Computers and Information Technology, University of Tabuk
  3. 3Department of Information Technology, College of Computers and Information Technology, Taif University
  4. 4Department of Computer Engineering, Collage of Computers and Information Technology, Taif University
  5. 5Department of Artificial Intelligence, Faculty of Computer Science and Information Technology, Al-Razi University
Read full textPDFDOICite

Abstract

Breast tumour segmentation in ultrasound imaging requires both fine boundary delineation and global contextual understanding. Existing evaluation metrics such as the Dice Similarity Coefficient provide aggregate performance estimates but do not distinguish segmentation behaviour between boundary and interior tumour regions. This study presents a hypothesis-driven comparative evaluation examining whether CNN, pure SSM, and hybrid SSM-CNN architectures exhibit distinct boundary-interior segmentation profiles. U-Net (ResNet-34 encoder), Vision Mamba (pure SSM), and VM-UNet (hybrid SSM-CNN) were trained on 3,333 breast ultrasound images (BUSI, BUS-BRA, BUS-UC) and externally validated on three independent sites (BrEaST, BLUI, BUS-UCLM; n = 748). We introduce the Boundary-Interior Ratio (BIR), a spatially decomposed metric quantifying relative segmentation performance between boundary and interior regions, together with a Grad-CAM-based Boundary-Interior Activation Ratio (BIAR) for mechanistic analysis. A structured four-variant ablation study examines the contributions of skip connections, VSS blocks, and hierarchical encoding. U-Net achieves the highest internal Dice (0.874) and strongest external generalisation (mean Dice = 0.771). Both SSM-based architectures exhibit higher BIR than U-Net (VM-UNet: 0.707; Vision Mamba: 0.722 vs. U-Net: 0.683), indicating distinct relative boundary-interior performance profiles. The ablation study shows that removing VSS blocks substantially improves absolute segmentation performance (Dice: 0.777→0.832; HD95: 18.75→12.61 px) while leaving BIR essentially unchanged (0.707→0.708), whereas removing skip connections produces the largest BIR change (0.707→0.652). These findings demonstrate that aggregate segmentation quality and boundary-interior performance provide complementary information, and that BIR reveals architectural differences not captured by Dice alone. The ablation results further indicate that, under the present training configuration, hierarchical encoding and skip-connected pathways have greater effects on overall segmentation performance and BIR variation, respectively, than the inclusion of VSS blocks alone. Under the present training configuration, the evaluated SSM architecture does not provide an absolute performance advantage, highlighting the need for future evaluation with pretrained SSM representations.

Related

  1. A segmentation-guided CNN–Vision transformer feature fusion framework for multi-class breast ultrasound image classification
    Meiru Wu, Jian Wang
    PLOS OnePeer-reviewedCC BY 4.0Full textMedical imaging & radiologyOncologyPLOS
  2. Development and evaluation of automated organ segmentation models for abdominal ultrasound in injured children
    Newton Addo, Kush Narang, Ritik Sinha +7
    medRxivPreprintCC BYFull textMedical imaging & radiologymedRxivPMC
  3. Spatially-adaptive dynamic dictionary learning for fine-grained remote sensing image segmentation
    Yan Zhang, Zifei Luo, Bo Yang +2
    Scientific ReportsPeer-reviewedCC BYFull text1 citedGeneral medicineDOAJ
  4. UAV image segmentation in complex mining areas by integrating spatially adaptive feature modulation and multi-scale context fusion modules
    Qing Li, Mingxing Lu, Yan Li
    Scientific ReportsPeer-reviewedCC BYFull textGeneral medicineDOAJ
  5. Comparison of Atheroma Burden Changes Using Conventional vs. End-Diastolic Intravascular Ultrasound Segmentation ― Post Hoc Analysis of the PACMAN-AMI Study ―
    Nathan Angelo Lecaros Yap, Christos V. Bourantas, Sylvain Losdat +15
    Circulation JournalPeer-reviewedCC BY-NC-NDFull text2 citedSurgery & anaesthesiaJ-STAGE

Cite

BibTeXRIS
BibTeX
@article{alhejri2026boundaryinterior,
  title = {{Boundary-interior ratio: a spatially decomposed evaluation framework for CNN and mamba-based breast ultrasound segmentation}},
  author = {Aymen M. Al-Hejri and Wad Ghaban and Sultan S. Alshamrani and Ahmed S. AlGhamdi and Ebrahim Mohammed Senan},
  journal = {Scientific Reports},
  year = {2026},
  month = sep,
  doi = {10.1038/s41598-026-68828-5},
  url = {https://doi.org/10.1038/s41598-026-68828-5},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - Boundary-interior ratio: a spatially decomposed evaluation framework for CNN and mamba-based breast ultrasound segmentation
AU  - Aymen M. Al-Hejri
AU  - Wad Ghaban
AU  - Sultan S. Alshamrani
AU  - Ahmed S. AlGhamdi
AU  - Ebrahim Mohammed Senan
PY  - 2026
DA  - 2026/09/01
JO  - Scientific Reports
DO  - 10.1038/s41598-026-68828-5
UR  - https://doi.org/10.1038/s41598-026-68828-5
AB  - Breast tumour segmentation in ultrasound imaging requires both fine boundary delineation and global contextual understanding. Existing evaluation metrics such as the Dice Similarity Coefficient provide aggregate performance estimates but do not distinguish segmentation behaviour between boundary and interior tumour regions. This study presents a hypothesis-driven comparative evaluation examining whether CNN, pure SSM, and hybrid SSM-CNN architectures exhibit distinct boundary-interior segmentation profiles. U-Net (ResNet-34 encoder), Vision Mamba (pure SSM), and VM-UNet (hybrid SSM-CNN) were trained on 3,333 breast ultrasound images (BUSI, BUS-BRA, BUS-UC) and externally validated on three independent sites (BrEaST, BLUI, BUS-UCLM; n = 748). We introduce the Boundary-Interior Ratio (BIR), a spatially decomposed metric quantifying relative segmentation performance between boundary and interior regions, together with a Grad-CAM-based Boundary-Interior Activation Ratio (BIAR) for mechanistic analysis. A structured four-variant ablation study examines the contributions of skip connections, VSS blocks, and hierarchical encoding. U-Net achieves the highest internal Dice (0.874) and strongest external generalisation (mean Dice = 0.771). Both SSM-based architectures exhibit higher BIR than U-Net (VM-UNet: 0.707; Vision Mamba: 0.722 vs. U-Net: 0.683), indicating distinct relative boundary-interior performance profiles. The ablation study shows that removing VSS blocks substantially improves absolute segmentation performance (Dice: 0.777→0.832; HD95: 18.75→12.61 px) while leaving BIR essentially unchanged (0.707→0.708), whereas removing skip connections produces the largest BIR change (0.707→0.652). These findings demonstrate that aggregate segmentation quality and boundary-interior performance provide complementary information, and that BIR reveals architectural differences not captured by Dice alone. The ablation results further indicate that, under the present training configuration, hierarchical encoding and skip-connected pathways have greater effects on overall segmentation performance and BIR variation, respectively, than the inclusion of VSS blocks alone. Under the present training configuration, the evaluated SSM architecture does not provide an absolute performance advantage, highlighting the need for future evaluation with pretrained SSM representations.
SN  - 2045-2322
ER  -