Fast and Accurate Road Crack Detection Based on Adaptive Cost-Sensitive Loss Function

Kai Li

doi:10.1109/TCYB.2021.3103885

Fast and Accurate Road Crack Detection Based on Adaptive Cost-Sensitive Loss Function

2023-02-01

26

OA

AI

DOI：10.1109/TCYB.2021.3103885

Original

Original request for help

Save

Abstract

En 中文

Numerous detection problems in computer vision, including road crack detection, suffer from exceedingly foreground-background imbalance. Fortunately, modification of loss function appears to solve this puzzle once and for all. In this article, we propose a pixel-based adaptive weighted cross-entropy (WCE) loss in conjunction with Jaccard distance to facilitate high-quality pixel-level road crack detection. Our work profoundly demonstrates the influence of loss functions on detection outcomes and sheds light on the sophisticated consecutive improvements in the realm of crack detection. Specifically, to verify the effectiveness of the proposed loss, we conduct extensive experiments on four public databases, that is, CrackForest, AigleRN, Crack360, and BJN260. Compared to the vanilla WCE, the proposed loss significantly speeds up the training process while retaining the performance.

Keywords:

Roads

Training

Costs

Image edge detection

Training data

Sampling methods

Adaptation models

Crack detection

Jaccard distance

U-Net

weighted cross-entropy (WCE)

Fast and Accurate Road Crack Detection Based on Adaptive Cost-Sensitive Loss Function

Abstract

Journal

Organization