Skip to Main content Skip to Navigation
Theses

Le système visuel humain au secours de la vision par ordinateur

Abstract : The aim of this work is to demonstrate that the human visual modelling yields to effi cient tools dedicated to image processing. The benefi ts are directly related to the human visual system properties and give the ability to solve common major problems encountered in image analysis such as noise reduction, details enhancement, back light correction. We propose a set of low level image processing tools which realize specifi c processing such as contours extraction, spectrum analysis, event detection. These tools are combined in order to create high level image analysis and we propose in this manuscript two examples: a face analyser which extracts eye blinks, yawnings, head motion, this application being applied to hypovigilance detection for drivers. The second considered application concerns general motion analysis and we propose a moving object tracker able to eliminate noise segmentation problems. Finally, we propose a bio-inspired algorithm for object identifi cation and classifi cation.
Complete list of metadatas

Cited literature [67 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00193715
Contributor : Patricia Reynier <>
Submitted on : Tuesday, December 4, 2007 - 2:29:10 PM
Last modification on : Thursday, November 19, 2020 - 12:59:51 PM
Long-term archiving on: : Monday, April 12, 2010 - 6:01:44 AM

Identifiers

  • HAL Id : tel-00193715, version 1

Collections

Citation

Alexandre Benoit. Le système visuel humain au secours de la vision par ordinateur. Traitement du signal et de l'image [eess.SP]. Institut National Polytechnique de Grenoble - INPG, 2007. Français. ⟨tel-00193715⟩

Share

Metrics

Record views

588

Files downloads

6345