Share Email Print

Proceedings Paper

A practical real-time detection visual system for driver's eye closure state tracking
Author(s): Tong Zhan; Zheng-ming Li; Jin Zhang
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

We use an onboard computer and a smart webcam built a practical real-time detection visual system. Installed in a driving cabin in a specific way, it is able to determine the degree of closure of the driver's eyelids for further study of detecting and mitigating driver drowsiness and fatigue. For a practical real-time application, both the temporal resolution and the image resolution of video processing are key performances. Instead of the traditional video processing strategy which detects face in the first frame and keep tracking in the following, a frame-by-frame detecting and locating strategy is adopted, which firstly locates face region in a size-reduced image, secondly face region in original image is located by linear mapping calculation, thirdly eye region is located in face region according to the characteristics of a facial geometric model, this model allows personal calibration intuitively in run-time according to different individuals. A novel combination of several practical methods proved effective in computer vision is proposed, including capturing and resizing frames, face and eye localization, adaptive-threshold segmentation, mathematical morphology etc. Experiments show that the proposed system achieves a performance of image resolution of 640×480 pixels, temporal resolution of 53 frames / second, average eye-blink detection accuracy of 96% under various illumination conditions.

Paper Details

Date Published: 13 January 2012
PDF: 8 pages
Proc. SPIE 8349, Fourth International Conference on Machine Vision (ICMV 2011): Machine Vision, Image Processing, and Pattern Analysis, 83491S (13 January 2012); doi: 10.1117/12.920976
Show Author Affiliations
Tong Zhan, Guangdong Polytechnic Normal Univ. (China)
Zheng-ming Li, Guangdong Polytechnic Normal Univ. (China)
Jin Zhang, Guangdong Polytechnic Normal Univ. (China)

Published in SPIE Proceedings Vol. 8349:
Fourth International Conference on Machine Vision (ICMV 2011): Machine Vision, Image Processing, and Pattern Analysis
Zhu Zeng; Yuting Li, Editor(s)

© SPIE. Terms of Use
Back to Top