MSP U-Net: Crack Segmentation for Low-Resolution Images Based on Multi-Scale Parallel Attention U-Net

As the expected lifespans of structures and road approaches, as well as the importance of road maintenance, increase globally, safety inspections have emerged as a crucial task. Nonetheless, the existing crack detection models focus on multi-scale feature loss and performance degradation in learning...

Full description

Saved in:
Bibliographic Details
Main Authors: Joon-Hyeok Kim, Ju-Hyeon Noh, Jun-Young Jang, Hee-Deok Yang
Format: Article
Language:English
Published: MDPI AG 2024-12-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/14/24/11541
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850240024747966464
author Joon-Hyeok Kim
Ju-Hyeon Noh
Jun-Young Jang
Hee-Deok Yang
author_facet Joon-Hyeok Kim
Ju-Hyeon Noh
Jun-Young Jang
Hee-Deok Yang
author_sort Joon-Hyeok Kim
collection DOAJ
description As the expected lifespans of structures and road approaches, as well as the importance of road maintenance, increase globally, safety inspections have emerged as a crucial task. Nonetheless, the existing crack detection models focus on multi-scale feature loss and performance degradation in learning various types of cracks. We propose the Multi-Scale Parallel Attention U-Net (MSP U-Net) as a network designed for low-resolution images that considers the irregular characteristics of cracks. MSP U-Net applies a large receptive field flock to an attention U-Net, minimizing feature loss across multiple scales. Using the Crack500 dataset, our network achieved a mean intersection of union (mIoU) of 0.7752, outperforming the existing methods on low-resolution images.
format Article
id doaj-art-9b30d64f63ef4ab8a043d42191cf32c6
institution OA Journals
issn 2076-3417
language English
publishDate 2024-12-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-9b30d64f63ef4ab8a043d42191cf32c62025-08-20T02:00:59ZengMDPI AGApplied Sciences2076-34172024-12-0114241154110.3390/app142411541MSP U-Net: Crack Segmentation for Low-Resolution Images Based on Multi-Scale Parallel Attention U-NetJoon-Hyeok Kim0Ju-Hyeon Noh1Jun-Young Jang2Hee-Deok Yang3Department of Computer Science, Chosun University, Gwangju 61452, Republic of KoreaDepartment of Computer Science, Chosun University, Gwangju 61452, Republic of KoreaDepartment of Computer Science, Chosun University, Gwangju 61452, Republic of KoreaDepartment of Computer Science, Chosun University, Gwangju 61452, Republic of KoreaAs the expected lifespans of structures and road approaches, as well as the importance of road maintenance, increase globally, safety inspections have emerged as a crucial task. Nonetheless, the existing crack detection models focus on multi-scale feature loss and performance degradation in learning various types of cracks. We propose the Multi-Scale Parallel Attention U-Net (MSP U-Net) as a network designed for low-resolution images that considers the irregular characteristics of cracks. MSP U-Net applies a large receptive field flock to an attention U-Net, minimizing feature loss across multiple scales. Using the Crack500 dataset, our network achieved a mean intersection of union (mIoU) of 0.7752, outperforming the existing methods on low-resolution images.https://www.mdpi.com/2076-3417/14/24/11541crack detectiondeep learningsemantic segmentation
spellingShingle Joon-Hyeok Kim
Ju-Hyeon Noh
Jun-Young Jang
Hee-Deok Yang
MSP U-Net: Crack Segmentation for Low-Resolution Images Based on Multi-Scale Parallel Attention U-Net
Applied Sciences
crack detection
deep learning
semantic segmentation
title MSP U-Net: Crack Segmentation for Low-Resolution Images Based on Multi-Scale Parallel Attention U-Net
title_full MSP U-Net: Crack Segmentation for Low-Resolution Images Based on Multi-Scale Parallel Attention U-Net
title_fullStr MSP U-Net: Crack Segmentation for Low-Resolution Images Based on Multi-Scale Parallel Attention U-Net
title_full_unstemmed MSP U-Net: Crack Segmentation for Low-Resolution Images Based on Multi-Scale Parallel Attention U-Net
title_short MSP U-Net: Crack Segmentation for Low-Resolution Images Based on Multi-Scale Parallel Attention U-Net
title_sort msp u net crack segmentation for low resolution images based on multi scale parallel attention u net
topic crack detection
deep learning
semantic segmentation
url https://www.mdpi.com/2076-3417/14/24/11541
work_keys_str_mv AT joonhyeokkim mspunetcracksegmentationforlowresolutionimagesbasedonmultiscaleparallelattentionunet
AT juhyeonnoh mspunetcracksegmentationforlowresolutionimagesbasedonmultiscaleparallelattentionunet
AT junyoungjang mspunetcracksegmentationforlowresolutionimagesbasedonmultiscaleparallelattentionunet
AT heedeokyang mspunetcracksegmentationforlowresolutionimagesbasedonmultiscaleparallelattentionunet