Eye Blink Detection Using Variance of Motion Vectors
Abstract
A new eye blink detection algorithm is proposed. It is based on analyzing the variance of the vertical motions in the eye region. The face and eyes are detected with a Viola–Jones type algorithm. Next, a flock of KLT trackers is placed over the eye region. For each eye, region is divided into $3\times 3$ cells. For each cell an average “cell” motion is calculated. Simple state machines analyse the variances for each eye. The proposed method has lower false positive rate compared to other methods based on tracking. We introduce a new challenging dataset Eyeblink8 . Our method achieves the best reported mean accuracy 99 % on the Talking dataset and state-of-the-art results on the ZJU dataset.
Cite
Text
Drutarovsky and Fogelton. "Eye Blink Detection Using Variance of Motion Vectors." European Conference on Computer Vision, 2014. doi:10.1007/978-3-319-16199-0_31Markdown
[Drutarovsky and Fogelton. "Eye Blink Detection Using Variance of Motion Vectors." European Conference on Computer Vision, 2014.](https://mlanthology.org/eccv/2014/drutarovsky2014eccv-eye/) doi:10.1007/978-3-319-16199-0_31BibTeX
@inproceedings{drutarovsky2014eccv-eye,
title = {{Eye Blink Detection Using Variance of Motion Vectors}},
author = {Drutarovsky, Tomas and Fogelton, Andrej},
booktitle = {European Conference on Computer Vision},
year = {2014},
pages = {436-448},
doi = {10.1007/978-3-319-16199-0_31},
url = {https://mlanthology.org/eccv/2014/drutarovsky2014eccv-eye/}
}