Weakly supervised pseudo label generation for construction vehicle segmentation

W. C. Chern, V. Asari, T. Nguyen, H. Kim

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Segmentation tasks in computer vision have been adopted in various studies in the civil engineering domain to provide accurate object locations in images. However, preparing annotation to train segmentation models is a time consuming and costly process, which hinders the use of segmentation models in vision-based applications. To address the problem, this study proposes a fusion model integrating self-supervised equivariant attention mechanism (SEAM) and sub-category exploration (SC-CAM) to generate pseudo labels in the form of polygon annotation from bounding box annotation that is relatively easy to obtain. To test the performance of the fusion model, a public data set - Advanced Infrastructure Management Group (AIM) dataset - for construction object detection was selected to generate pseudo labels; the effectiveness of pseudo labels was measured by the segmentation performance of a feature pyramid network (FPN) trained with the pseudo labels. FPN showed the mean intersection over union (mIoU) score of 86.03%, demonstrating the potential of the proposed fusion model to reduce the manual annotation efforts in preparing training data for segmentation models.

Original languageEnglish
Title of host publicationProceedings of the 39th International Symposium on Automation and Robotics in Construction, ISARC 2022
PublisherInternational Association for Automation and Robotics in Construction (IAARC)
Pages41-46
Number of pages6
ISBN (Electronic)9789526952420
Publication statusPublished - 2022
Event39th International Symposium on Automation and Robotics in Construction, ISARC 2022 - Bogota, Colombia
Duration: 2022 Jul 132022 Jul 15

Publication series

NameProceedings of the International Symposium on Automation and Robotics in Construction
Volume2022-July
ISSN (Electronic)2413-5844

Conference

Conference39th International Symposium on Automation and Robotics in Construction, ISARC 2022
Country/TerritoryColombia
CityBogota
Period22/7/1322/7/15

Bibliographical note

Publisher Copyright:
© 2022 International Association on Automation and Robotics in Construction.

All Science Journal Classification (ASJC) codes

  • Artificial Intelligence
  • Safety, Risk, Reliability and Quality
  • Control and Systems Engineering
  • Building and Construction
  • Computer Science Applications
  • Civil and Structural Engineering

Fingerprint

Dive into the research topics of 'Weakly supervised pseudo label generation for construction vehicle segmentation'. Together they form a unique fingerprint.

Cite this