Serial Local Patterns and Irregular Dependencies Extract and Cascaded Fusion Network for Structural Crack Segmentation

Hui Liu; Chen Jia; Xu Cheng; Xiufeng Liu; Fan Shi

doi:10.1109/ICASSP49660.2025.10887799

Serial Local Patterns and Irregular Dependencies Extract and Cascaded Fusion Network for Structural Crack Segmentation

Hui Liu
, Chen Jia^*
, Xu Cheng
, Xiufeng Liu
, Fan Shi

^*Corresponding author for this work

Tianjin University of Technology

Research output: Contribution to journal › Conference article › Research › peer-review

Abstract

Achieving pixel-level crack segmentation in complex scenarios is a major challenge, as current methods have difficulty effectively integrating both local features and irregular pixel dependencies. In this paper, we introduce a Cascaded Fusion Network (LICFN) specifically designed for crack segmentation, which extracts fine local details and pixel dependencies using a hybrid feature extractor and effectively enhances and fuses them through a cascaded fusion module. To comprehensively evaluate the network, we also created a benchmark dataset, TUT, which includes various scenarios. Experimental results show that our method surpasses others, achieving F1 and mIoU scores of 0.8439 and 0.8509, respectively. The dataset is available at https://github.com/Karl1109/TUT.

Original language	English
Journal	ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
ISSN	1520-6149
DOIs	https://doi.org/10.1109/ICASSP49660.2025.10887799
Publication status	Published - 2025
Event	2025 IEEE International Conference on Acoustics, Speech, and Signal Processing - Hyderabad, India Duration: 6 Apr 2025 → 11 Apr 2025

Conference

Conference	2025 IEEE International Conference on Acoustics, Speech, and Signal Processing
Country/Territory	India
City	Hyderabad
Period	06/04/2025 → 11/04/2025
Sponsor	IEEE

Keywords

Cascaded fusion
Crack dataset
Irregular dependencies
Local patterns
Structural cracks segmentation

Access to Document

10.1109/ICASSP49660.2025.10887799

OpenUrl availability

Check availability in DTU Findit

Cite this

@inproceedings{41be562f2af544f8b53dfb3837d62410,

title = "Serial Local Patterns and Irregular Dependencies Extract and Cascaded Fusion Network for Structural Crack Segmentation",

abstract = "Achieving pixel-level crack segmentation in complex scenarios is a major challenge, as current methods have difficulty effectively integrating both local features and irregular pixel dependencies. In this paper, we introduce a Cascaded Fusion Network (LICFN) specifically designed for crack segmentation, which extracts fine local details and pixel dependencies using a hybrid feature extractor and effectively enhances and fuses them through a cascaded fusion module. To comprehensively evaluate the network, we also created a benchmark dataset, TUT, which includes various scenarios. Experimental results show that our method surpasses others, achieving F1 and mIoU scores of 0.8439 and 0.8509, respectively. The dataset is available at https://github.com/Karl1109/TUT.",

keywords = "Cascaded fusion, Crack dataset, Irregular dependencies, Local patterns, Structural cracks segmentation",

author = "Hui Liu and Chen Jia and Xu Cheng and Xiufeng Liu and Fan Shi",

note = "Publisher Copyright: {\textcopyright} 2025 IEEE.; 2025 IEEE International Conference on Acoustics, Speech, and Signal Processing , ICASSP ; Conference date: 06-04-2025 Through 11-04-2025",

year = "2025",

doi = "10.1109/ICASSP49660.2025.10887799",

language = "English",

journal = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",

issn = "1520-6149",

publisher = "IEEE Service Center",

}

TY - GEN

T1 - Serial Local Patterns and Irregular Dependencies Extract and Cascaded Fusion Network for Structural Crack Segmentation

AU - Liu, Hui

AU - Jia, Chen

AU - Cheng, Xu

AU - Liu, Xiufeng

AU - Shi, Fan

PY - 2025

Y1 - 2025

N2 - Achieving pixel-level crack segmentation in complex scenarios is a major challenge, as current methods have difficulty effectively integrating both local features and irregular pixel dependencies. In this paper, we introduce a Cascaded Fusion Network (LICFN) specifically designed for crack segmentation, which extracts fine local details and pixel dependencies using a hybrid feature extractor and effectively enhances and fuses them through a cascaded fusion module. To comprehensively evaluate the network, we also created a benchmark dataset, TUT, which includes various scenarios. Experimental results show that our method surpasses others, achieving F1 and mIoU scores of 0.8439 and 0.8509, respectively. The dataset is available at https://github.com/Karl1109/TUT.

AB - Achieving pixel-level crack segmentation in complex scenarios is a major challenge, as current methods have difficulty effectively integrating both local features and irregular pixel dependencies. In this paper, we introduce a Cascaded Fusion Network (LICFN) specifically designed for crack segmentation, which extracts fine local details and pixel dependencies using a hybrid feature extractor and effectively enhances and fuses them through a cascaded fusion module. To comprehensively evaluate the network, we also created a benchmark dataset, TUT, which includes various scenarios. Experimental results show that our method surpasses others, achieving F1 and mIoU scores of 0.8439 and 0.8509, respectively. The dataset is available at https://github.com/Karl1109/TUT.

KW - Cascaded fusion

KW - Crack dataset

KW - Irregular dependencies

KW - Local patterns

KW - Structural cracks segmentation

U2 - 10.1109/ICASSP49660.2025.10887799

DO - 10.1109/ICASSP49660.2025.10887799

M3 - Conference article

AN - SCOPUS:105003881525

SN - 1520-6149

JO - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

JF - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

T2 - 2025 IEEE International Conference on Acoustics, Speech, and Signal Processing

Y2 - 6 April 2025 through 11 April 2025

ER -

Serial Local Patterns and Irregular Dependencies Extract and Cascaded Fusion Network for Structural Crack Segmentation

Abstract

Conference

Keywords

Access to Document

OpenUrl availability

Fingerprint

Cite this