<abstract xmlns="http://www.w3.org/1999/xhtml">

<p>Automatic crack detection in construction facilities is a challenging yet crucial task. However, existing deep learning (DL)-based semantic segmentation methods for this field are based on fully supervised learning models and pixel-level manual annotation, which are time-consuming and labor-intensive. To solve this problem, this paper proposes a novel crack semantic segmentation network using weakly supervised approach and mixed-label training strategy. Firstly, an image patch-level classifier of crack is trained to generate a coarse localization map for automatic pseudo-labeling of cracks combined with a thresholding-based method. Then, we integrated the pseudo-annotated with manual-annotated samples with a ratio of 4:1 to train the crack segmentation network with a mixed-label training strategy, in which the manual labels were assigned with a higher weight value. The experimental data on two public datasets demonstrate that our proposed method achieves a comparable accuracy with the fully supervised methods, reducing over 65% of the manual annotation workload.</p>
</abstract>

Automatic crack detection in construction facilities is a challenging yet crucial task. However, existing deep learning (DL)-based semantic segmentation methods for this field are based on fully supervised learning models and pixel-level manual annotation, which are time-consuming and labor-intensive. To solve this problem, this paper proposes a novel crack semantic segmentation network using weakly supervised approach and mixed-label training strategy. Firstly, an image patch-level classifier of crack is trained to generate a coarse localization map for automatic pseudo-labeling of cracks combined with a thresholding-based method. Then, we integrated the pseudo-annotated with manual-annotated samples with a ratio of 4:1 to train the crack segmentation network with a mixed-label training strategy, in which the manual labels were assigned with a higher weight value. The experimental data on two public datasets demonstrate that our proposed method achieves a comparable accuracy with the fully supervised methods, reducing over 65% of the manual annotation workload.

Automatic Crack Detection Using Weakly Supervised Semantic Segmentation Network and Mixed-Label Training Strategy

Foundations of Computing and Decision Sciences

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Automatic crack detection in construction facilities is a challenging yet crucial task. However, existing deep learning (DL)-based semantic segmentation...

Automatic Crack Detection Using Weakly Supervised Semantic Segmentation Network and Mixed-Label Training Strategy

Publié en ligne: 16 févr. 2024

Pages: 95 - 118

Reçu: 30 août 2023

Accepté: 16 nov. 2023

DOI: https://doi.org/10.2478/fcds-2024-0007

Mots clés
weakly supervised learning, crack detection, semantic segmentation, mixed label

© 2024 Shuyuan Zhang et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Automatic Crack Detection Using Weakly Supervised Semantic Segmentation Network and Mixed-Label Training Strategy

Publié en ligne: 16 févr. 2024

Pages: 95 - 118

Reçu: 30 août 2023

Accepté: 16 nov. 2023

DOI: https://doi.org/10.2478/fcds-2024-0007

Mots clésweakly supervised learning, crack detection, semantic segmentation, mixed label

© 2024 Shuyuan Zhang et al., published by Sciendo

This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Mots clés
weakly supervised learning, crack detection, semantic segmentation, mixed label