Skip to main navigation Skip to search Skip to main content

Deep CNN-based pedestrian detection for intelligent infrastructure

  • Bilel Tarchoun
  • , Imen Jegham
  • , Anouar Ben Khalifa
  • , Ihsen Alouani
  • , Mohamed Ali Mahjoub

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

Abstract

Autonomous driving systems and driver assistance systems are becoming the center of attention in transport technology. Given its safety criticality, pedestrian detection is a highly important task. Transport oriented intelligent systems use embedded sensors for the detection task. However, vehicle side detection is starting to show its limitations especially when dealing with certain challenges such as occlusions. In this paper, we propose an infrastructure side perception system that has a bird's eye view. We introduce a new deep pedestrian detector that can use the detection results to warn nearby vehicles of the presence of pedestrians on the road. The results show that our proposed system is able to detect pedestrians in most conditions with 70.41% precision and 69.17% recall.

Original languageEnglish
Title of host publicationProceedings of the 5th International Conference on Advanced Technologies for Signal and Image Processing, ATSIP 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
Number of pages6
ISBN (Electronic)9781728175133
ISBN (Print)9781728175140
DOIs
Publication statusPublished - 20 Oct 2020
Externally publishedYes
Event5th International Conference on Advanced Technologies for Signal and Image Processing - Sousse, Tunisia
Duration: 02 Sept 202005 Sept 2020

Publication series

NameInternational Conference on Advanced Technologies for Signal and Image Processing: Proceedings
ISSN (Print)2641-5941
ISSN (Electronic)2687-878X

Conference

Conference5th International Conference on Advanced Technologies for Signal and Image Processing
Abbreviated titleATSIP
Country/TerritoryTunisia
CitySousse
Period02/09/202005/09/2020

Keywords

  • Faster R-CNN
  • intelligent infrastructure
  • Intelligent transportation systems
  • Pedestrian detection
  • transfer learning

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Signal Processing
  • Statistics, Probability and Uncertainty
  • Safety, Risk, Reliability and Quality
  • Statistics and Probability

Fingerprint

Dive into the research topics of 'Deep CNN-based pedestrian detection for intelligent infrastructure'. Together they form a unique fingerprint.

Cite this