A Two-Stream Context-Aware ConvNet for Pavement Distress Detection

Lõuk, Roland; Tepljakov, Aleksei; Riid, Andri

doi:10.1109/tsp49548.2020.9163538

Cited by 2 publications

(1 citation statement)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In our previous research [12], a ResNet based on transfer learning [26] architecture was found to attain satisfactory results for situations in which the patches were sampled from roads with less ambiguous defects and better lighting conditions. However, our further developments on similar datasets have shown that there is a possibility of obtaining slight performance gains by utilizing the contextual area surrounding each of the patches within the orthoframe [27]. By extracting relevant features from not only the patch but also its surroundings, we can train a classifier better equipped to deal with difficult cases in which more contextual information is needed for an accurate classification (such as when a pavement crack lies at the very edge of the patch).…”

Section: Architecturementioning

confidence: 99%

Pavement Defect Segmentation in Orthoframes with a Pipeline of Three Convolutional Neural Networks

et al. 2020

Self Cite

View full text Add to dashboard Cite

In the manuscript, the issue of detecting and segmenting out pavement defects on highway roads is addressed. Specifically, computer vision (CV) methods are developed and applied to the problem based on deep learning of convolutional neural networks (ConvNets). A novel neural network structure is considered, based on a pipeline of three ConvNets and endowed with the capacity for context awareness, which improves grid-based search for defects on orthoframes by considering the surrounding image content—an approach, which essentially draws inspiration from how humans tend to solve the task of image segmentation. Also, methods for assessing the quality of segmentation are discussed. The contribution also describes the complete procedure of working with pavement defects in an industrial setting, involving the workcycle of defect annotation, ConvNet training and validation. The results of ConvNet evaluation provided in the paper hint at a successful implementation of the proposed technique.

show abstract

Section: Architecturementioning

confidence: 99%