TransMF: Transformer-Based Multi-Scale Fusion Model for Crack Detection
Cracks are widespread in infrastructure that are closely related to human activity. It is very popular to use artificial intelligence to detect cracks intelligently, which is known as crack detection. The noise in the background of crack images, discontinuity of cracks and other problems make the cr...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-07-01
|
Series: | Mathematics |
Subjects: | |
Online Access: | https://www.mdpi.com/2227-7390/10/13/2354 |
_version_ | 1797434105895321600 |
---|---|
author | Xiaochen Ju Xinxin Zhao Shengsheng Qian |
author_facet | Xiaochen Ju Xinxin Zhao Shengsheng Qian |
author_sort | Xiaochen Ju |
collection | DOAJ |
description | Cracks are widespread in infrastructure that are closely related to human activity. It is very popular to use artificial intelligence to detect cracks intelligently, which is known as crack detection. The noise in the background of crack images, discontinuity of cracks and other problems make the crack detection task a huge challenge. Although many approaches have been proposed, there are still two challenges: (1) cracks are long and complex in shape, making it difficult to capture long-range continuity; (2) most of the images in the crack dataset have noise, and it is difficult to detect only the cracks and ignore the noise. In this paper, we propose a novel method called <i>Transformer-based Multi-scale Fusion Model</i> (TransMF) for crack detection, including an Encoder Module (EM), Decoder Module (DM) and Fusion Module (FM). The Encoder Module uses a hybrid of convolution blocks and Swin Transformer block to model the long-range dependencies of different parts in a crack image from a local and global perspective. The Decoder Module is designed with symmetrical structure to the Encoder Module. In the Fusion Module, the output in each layer with unique scales of Encoder Module and Decoder Module are fused in the form of convolution, which can release the effect of background noise and strengthen the correlations between relevant context in order to enhance the crack detection. Finally, the output of each layer of the Fusion Module is concatenated to achieve the purpose of crack detection. Extensive experiments on three benchmark datasets (CrackLS315, CRKWH100 and DeepCrack) demonstrate that the proposed TransMF in this paper exceeds the best performance of present baselines. |
first_indexed | 2024-03-09T10:27:30Z |
format | Article |
id | doaj.art-ba57899425034165923b4ffb3889d5d3 |
institution | Directory Open Access Journal |
issn | 2227-7390 |
language | English |
last_indexed | 2024-03-09T10:27:30Z |
publishDate | 2022-07-01 |
publisher | MDPI AG |
record_format | Article |
series | Mathematics |
spelling | doaj.art-ba57899425034165923b4ffb3889d5d32023-12-01T21:35:44ZengMDPI AGMathematics2227-73902022-07-011013235410.3390/math10132354TransMF: Transformer-Based Multi-Scale Fusion Model for Crack DetectionXiaochen Ju0Xinxin Zhao1Shengsheng Qian2Railway Engineering Research Institute, China Academy of Railway Sciences Corporation Limited, Beijing 100081, ChinaRailway Engineering Research Institute, China Academy of Railway Sciences Corporation Limited, Beijing 100081, ChinaInstitute of Automation, Chinese Academy of Sciences, Beijing 100090, ChinaCracks are widespread in infrastructure that are closely related to human activity. It is very popular to use artificial intelligence to detect cracks intelligently, which is known as crack detection. The noise in the background of crack images, discontinuity of cracks and other problems make the crack detection task a huge challenge. Although many approaches have been proposed, there are still two challenges: (1) cracks are long and complex in shape, making it difficult to capture long-range continuity; (2) most of the images in the crack dataset have noise, and it is difficult to detect only the cracks and ignore the noise. In this paper, we propose a novel method called <i>Transformer-based Multi-scale Fusion Model</i> (TransMF) for crack detection, including an Encoder Module (EM), Decoder Module (DM) and Fusion Module (FM). The Encoder Module uses a hybrid of convolution blocks and Swin Transformer block to model the long-range dependencies of different parts in a crack image from a local and global perspective. The Decoder Module is designed with symmetrical structure to the Encoder Module. In the Fusion Module, the output in each layer with unique scales of Encoder Module and Decoder Module are fused in the form of convolution, which can release the effect of background noise and strengthen the correlations between relevant context in order to enhance the crack detection. Finally, the output of each layer of the Fusion Module is concatenated to achieve the purpose of crack detection. Extensive experiments on three benchmark datasets (CrackLS315, CRKWH100 and DeepCrack) demonstrate that the proposed TransMF in this paper exceeds the best performance of present baselines.https://www.mdpi.com/2227-7390/10/13/2354crack detectionconvolutional neural networktransformermulti-scale fusion |
spellingShingle | Xiaochen Ju Xinxin Zhao Shengsheng Qian TransMF: Transformer-Based Multi-Scale Fusion Model for Crack Detection Mathematics crack detection convolutional neural network transformer multi-scale fusion |
title | TransMF: Transformer-Based Multi-Scale Fusion Model for Crack Detection |
title_full | TransMF: Transformer-Based Multi-Scale Fusion Model for Crack Detection |
title_fullStr | TransMF: Transformer-Based Multi-Scale Fusion Model for Crack Detection |
title_full_unstemmed | TransMF: Transformer-Based Multi-Scale Fusion Model for Crack Detection |
title_short | TransMF: Transformer-Based Multi-Scale Fusion Model for Crack Detection |
title_sort | transmf transformer based multi scale fusion model for crack detection |
topic | crack detection convolutional neural network transformer multi-scale fusion |
url | https://www.mdpi.com/2227-7390/10/13/2354 |
work_keys_str_mv | AT xiaochenju transmftransformerbasedmultiscalefusionmodelforcrackdetection AT xinxinzhao transmftransformerbasedmultiscalefusionmodelforcrackdetection AT shengshengqian transmftransformerbasedmultiscalefusionmodelforcrackdetection |