Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

UP-CrackNet: Unsupervised Pixel-Wise Road Crack Detection via Adversarial Image Restoration

Nachuan Ma, Rui Fan, Lihua Xie

TL;DR

This work proposes an unsupervised pixel-wise road crack detection network, known as UP-CrackNet, which outperforms other general-purpose unsupervised anomaly detection algorithms, and exhibits satisfactory performance and superior generalizability when compared with state-of-the-art supervised crack segmentation algorithms.

Abstract

Over the past decade, automated methods have been developed to detect cracks more efficiently, accurately, and objectively, with the ultimate goal of replacing conventional manual visual inspection techniques. Among these methods, semantic segmentation algorithms have demonstrated promising results in pixel-wise crack detection tasks. However, training such networks requires a large amount of human-annotated datasets with pixel-level annotations, which is a highly labor-intensive and time-consuming process. Moreover, supervised learning-based methods often struggle with poor generalizability in unseen datasets. Therefore, we propose an unsupervised pixel-wise road crack detection network, known as UP-CrackNet. Our approach first generates multi-scale square masks and randomly selects them to corrupt undamaged road images by removing certain regions. Subsequently, a generative adversarial network is trained to restore the corrupted regions by leveraging the semantic context learned from surrounding uncorrupted regions. During the testing phase, an error map is generated by calculating the difference between the input and restored images, which allows for pixel-wise crack detection. Our comprehensive experimental results demonstrate that UP-CrackNet outperforms other general-purpose unsupervised anomaly detection algorithms, and exhibits satisfactory performance and superior generalizability when compared with state-of-the-art supervised crack segmentation algorithms. Our source code is publicly available at mias.group/UP-CrackNet.

UP-CrackNet: Unsupervised Pixel-Wise Road Crack Detection via Adversarial Image Restoration

TL;DR

This work proposes an unsupervised pixel-wise road crack detection network, known as UP-CrackNet, which outperforms other general-purpose unsupervised anomaly detection algorithms, and exhibits satisfactory performance and superior generalizability when compared with state-of-the-art supervised crack segmentation algorithms.

Abstract

Over the past decade, automated methods have been developed to detect cracks more efficiently, accurately, and objectively, with the ultimate goal of replacing conventional manual visual inspection techniques. Among these methods, semantic segmentation algorithms have demonstrated promising results in pixel-wise crack detection tasks. However, training such networks requires a large amount of human-annotated datasets with pixel-level annotations, which is a highly labor-intensive and time-consuming process. Moreover, supervised learning-based methods often struggle with poor generalizability in unseen datasets. Therefore, we propose an unsupervised pixel-wise road crack detection network, known as UP-CrackNet. Our approach first generates multi-scale square masks and randomly selects them to corrupt undamaged road images by removing certain regions. Subsequently, a generative adversarial network is trained to restore the corrupted regions by leveraging the semantic context learned from surrounding uncorrupted regions. During the testing phase, an error map is generated by calculating the difference between the input and restored images, which allows for pixel-wise crack detection. Our comprehensive experimental results demonstrate that UP-CrackNet outperforms other general-purpose unsupervised anomaly detection algorithms, and exhibits satisfactory performance and superior generalizability when compared with state-of-the-art supervised crack segmentation algorithms. Our source code is publicly available at mias.group/UP-CrackNet.
Paper Structure (21 sections, 8 equations, 4 figures, 8 tables)

This paper contains 21 sections, 8 equations, 4 figures, 8 tables.

Table of Contents

  1. Introduction
  2. Literature review
  3. Traditional Road Crack Detection Methods
  4. Supervised Road Crack Detection Methods
  5. Unsupervised Anomaly Detection Methods Based on Image Restoration
  6. Methodology
  7. Architecture Overview
  8. Undamaged Road Image Random Corruption
  9. Adversarial Image Restoration
  10. Road Crack Detection
  11. Loss Functions
  12. Restoration Loss
  13. Adversarial loss
  14. Experimental results
  15. Datasets
  16. ...and 6 more sections

Figures (4)

  • Figure 1: An illustrative pipeline of our proposed UP-CrackNet.
  • Figure 2: Examples of experimental results on the Crack500 dataset: (a) DeepLabv3+ chen2018encoder; (b) ENet paszke2016enet; (c) PSPNet zhao2017pyramid; (d) UperNet xiao2018unified; (e) SegResNet badrinarayanan2017segnet; (f) UNet ronneberger2015u; (g) BiSeNetv2 yu2021bisenet; (h) DDRNet pan2022deep; (i) Lawin yan2022lawin; (j) Deepcrack19 liu2019deepcrack; (k) Deepcrack18 zou2018deepcrack; (l) SCADN yan2021learning; (m) RIAD zavrtanik2021reconstruction; (n) UP-CrackNet. The true-positive, false-positive, and false-negative pixels are shown in green, blue, and red, respectively.
  • Figure 3: Examples of experimental results on the Deepcrack dataset liu2019deepcrack:(a) DeepLabv3+ chen2018encoder; (b) ENet paszke2016enet; (c) PSPNet zhao2017pyramid; (d) UperNet xiao2018unified; (e) SegResNet badrinarayanan2017segnet; (f) UNet ronneberger2015u; (g) BiSeNetv2 yu2021bisenet; (h) DDRNet pan2022deep; (i) Lawin yan2022lawin; (j) Deepcrack19 liu2019deepcrack; (k) Deepcrack18 zou2018deepcrack; (l) SCADN yan2021learning; (m) RIAD zavrtanik2021reconstruction; (n) UP-CrackNet. The true-positive, false-positive, and false-negative pixels are shown in green, blue, and red, respectively.
  • Figure 4: Failure cases of UP-CrackNet on the Deepcrack dataset liu2019deepcrack. The true-positive, false-positive, and false-negative pixels are shown in green, blue, and red, respectively.