Two-step deep learning approach for pavement crack damage detection and segmentation

Jiang, Yongqing; Pang, Dongfang; Li, Chengdong; Yu, Yulong; Cao, Yukang

doi:10.1080/10298436.2022.2065488

Cited by 12 publications

(5 citation statements)

References 60 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Second, we introduce the attention mechanism: the attention mechanism suppresses interference and enhances crack identi cation accuracy. Jiang et al [13] introduced CBAM (Convolutional Block Attention Module) to improve segmentation accuracy; Zhou et al [2] combined channel attention and spatial attention modules to improve the effectiveness of crack feature extraction. We lter the seam linear information by introducing BRA for low-level feature selection at the end of the encoder to enhance the model's immunity to seam linear.…”

Section: Improvement Of Unet Architecture To Enhance Crack Detection ...mentioning

confidence: 99%

“…Zhang et al [14] replaced the traditional convolution with DSC in UNet, but when dealing with complex and channel-associated segmentation of cracks in the tunnel lining, it may lead to poor segmentation results due to ignoring the cross-channel correlation. Jiang et al [13] applied the Ghost module in DeepLabv3 + to reduce the number of native features by generating "Ghost" features, which improves the e ciency but may sacri ce the richness and diversity of some features. The decoder in this study cascades the DSC and Ghost modules into a DDG module, which draws on the strengths of each to complement each other and enhance the adaptation to cracks at different scales.…”

Section: Improvement Of Unet Architecture To Enhance Crack Detection ...mentioning

confidence: 99%

See 1 more Smart Citation

EFE-UNet: Efficient Detection of Tunnel Cracks under Seam Linear Interference and Multiscale Crack Differences

Guo,

Cui,

Wang

et al. 2024

Preprint

View full text Add to dashboard Cite

Aiming at the problems of poor continuity and low recognition rate of tunnel lining crack segmentation under the influence of seam linear interference and multi-scale differences, and the large number of model parameters, which makes it difficult to balance the efficiency and accuracy of the model well, a tunnel lining crack segmentation method based on the improved UNet network, EFE-UNet (Efficient Feature Enhancement UNet), is proposed. Initially, an EMO (Efficient MOdel) lightweight network is used to replace the VggNet backbone network in the original UNet network, which enhances the ability to describe the feature context information of the cracks and reduces the parameters of the network at the same time; BRA (Bi-Level Routing Attention) is added at the end of the encoder to implement the low-level feature selection and filter the crack information using the weights, to enhance the model's accuracy of crack segmentation. seam linear information to enhance the sensitivity of the model to the crack region and the anti-interference ability to the seam. A DSC (Depthwise Separable Convolution) and Ghost module cascaded with a DDG (Double DSC Ghost) module are used instead of the standard convolution of the decoder to reduce the model complexity and enhance the adaptability to cracks at different scales. On the Tunnel200 dataset, EFE-UNet outperforms other semantic segmentation methods with the best metrics. It also shows excellent performance on the CRACK500 dataset.

show abstract

Section: Improvement Of Unet Architecture To Enhance Crack Detection ...mentioning

confidence: 99%

Section: Improvement Of Unet Architecture To Enhance Crack Detection ...mentioning

confidence: 99%

EFE-UNet: Efficient Detection of Tunnel Cracks under Seam Linear Interference and Multiscale Crack Differences

Guo,

Cui,

Wang

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…In another study, Jiang et al [14] proposed a method based on deep learning called DDSNet, combines optimized YOLOv4 and deeplabv3+ models for two-stage pavement crack detection and segmentation. They stated that the accuracy is improved by 2.23% and 7.47%, respectively, and the inference speed is increased by 35.3% and 50.3%, respectively.…”

Section: Literature Reviewmentioning

confidence: 99%

Deep Learning for Physical Damage Detection in Buildings: A Comparison of Transfer Learning Methods

Ekici

USTAOĞLU²

2023

Turkish Journal of Science and Technology

View full text Add to dashboard Cite

The detection of physical damage in buildings is a critical task in ensuring the safety and integrity of structures. In this study, we investigate the effectiveness of deep learning methods for detecting physical damage in buildings, specifically focusing on cracks, defect, moisture, and undamaged classes. We use transfer learning methods, including VGG16, GoogLeNet, and ResNet50, to classify a dataset of 7200 images. We split the dataset into training, validation, and testing sets, and evaluate the performance of the models using metrics such as accuracy, precision, recall, and F1-score. Our results show that all three models achieve high accuracy on the test set, with VGG16 and ResNet50 outperforming GoogLeNet. Additionally, precision, recall, and F1-score metrics indicate strong performance across all classes, with VGG16 and ResNet50 achieving particularly high scores. Our study demonstrates the effectiveness of deep learning methods for physical damage detection in buildings and provides insights into the comparative performance of transfer learning methods.

show abstract

“…In this section, we evaluate the performance of the proposed model. For comprehensive comparison, we select CNN models including FCN [10], U-Net [12] GhostNet [13], MobileNetv3 [41], and DeepLabv3+ [42] Transformer models including ViT [15], Segmenter [16], Swin Transformer [43], SETR [17], DeiT [44], as well as two-stage models, such as ALSNet [45], TSUNet [46], DDNet [47], APCNet [48], CMMANet [49], and PCDNet [50] as benchmark methods. For the GhostNet and ViT models, we preserve their encoding parts and replace the original classification head with a 1 × 1 convolution.…”

Section: Comparisons With the State Of The Artmentioning

confidence: 99%

TTH-Net: Two-Stage Transformer–CNN Hybrid Network for Leaf Vein Segmentation

Song,

Yu,

Zhang

2023

Applied Sciences

View full text Add to dashboard Cite

Leaf vein segmentation is crucial in species classification and smart agriculture. The existing methods combine manual features and machine learning techniques to segment coarse leaf veins. However, the extraction of the intricate patterns is time consuming. To address the issues, we propose a coarse-to-fine two-stage hybrid network termed TTH-Net, which combines a transformer and CNN to accurately extract veins. Specifically, the proposed TTH-Net consists of two stages and a cross-stage semantic enhancement module. The first stage utilizes the Vision Transformer (base version) to extract globally high-level feature representations. Based on these features, the second stage identifies fine-grained vein features via CNN. To enhance the interaction between the two stages, a cross-stage semantic enhancement module is designed to integrate the strengths of the transformer and CNN, which also improves the segmentation accuracy of the decoder. Extensive experiments on the public dataset LVN are conducted, and the results prove that TTH-Net has significant advantages over other methods in leaf vein segmentation.

show abstract

Two-step deep learning approach for pavement crack damage detection and segmentation

Cited by 12 publications

References 60 publications

EFE-UNet: Efficient Detection of Tunnel Cracks under Seam Linear Interference and Multiscale Crack Differences

EFE-UNet: Efficient Detection of Tunnel Cracks under Seam Linear Interference and Multiscale Crack Differences

Deep Learning for Physical Damage Detection in Buildings: A Comparison of Transfer Learning Methods

TTH-Net: Two-Stage Transformer–CNN Hybrid Network for Leaf Vein Segmentation

Contact Info

Product

Resources

About