Detail publikace

On Image Segmentation Techniques for Driver Inattention Systems

HORÁK, K. HONZÍK, P. KUČERA, P.

Originální název

On Image Segmentation Techniques for Driver Inattention Systems

Typ

článek ve sborníku ve WoS nebo Scopus

Jazyk

angličtina

Originální abstrakt

Visual systems for automatic monitoring of driver vigilance usually have to address two main problems. First of all, they have to acquire and process image sequence so that fatigue features can be simply extracted. Secondly, visual systems have to analyse a set of acquired features and subsequently recognize dangerous behaviour such as driver inattention or sleepiness. This paper is focused particularly on segmentation methods used for reliable eyes tracking, because of eyes features are probably most significant features for determining of a driver fatigue. Fundamentals segmentation methods as simple colour segmentation or Hough transform as well as more complex methods as Haar-like features or symmetries detection are introduced in the paper. Several of the most frequently used fatigue features are listed and described at the end of the paper. All the presented methods were tested and verified on both laboratory and real sets of images.

Klíčová slova

driver fatigue, eyes detection, inattention, image processing, segmentation, tracking

Autoři

HORÁK, K.; HONZÍK, P.; KUČERA, P.

Rok RIV

2011

Vydáno

17. 6. 2011

Nakladatel

Institute of Automation and Computer Science

Místo

Brno, Czech Republic

ISBN

978-80-214-4120-0

Kniha

The Proceedings of the 17th International Conference on Soft Computing

Strany od

1

Strany do

6

Strany počet

5

BibTex

@inproceedings{BUT74375,
  author="Karel {Horák} and Petr {Honzík} and Pavel {Kučera}",
  title="On Image Segmentation Techniques for Driver Inattention Systems",
  booktitle="The Proceedings of the 17th International Conference on Soft Computing",
  year="2011",
  pages="1--6",
  publisher="Institute of Automation and Computer Science",
  address="Brno, Czech Republic",
  isbn="978-80-214-4120-0"
}