Recognition of Human Facial Expressions Without Feature Extraction
Abstract
This paper presents a new method of recognizing facial expressions using a two dimensional physical model named Potential Net. The advantage of the method is not to need extracting facial features from an image, so that it is robust for variations of illumination and facial individualities. Potential Net is a physical model which consists of nodes connected by springs in two dimensional grid configuration. This net is set on a facial image and is deformed by image force, which moves the nodes to positions near to facial features. Recognition is executed by analyzing the similarity between model nets prepared previously and a net deformed by an input image.
Cite
Text
Matsuno et al. "Recognition of Human Facial Expressions Without Feature Extraction." European Conference on Computer Vision, 1994. doi:10.1007/3-540-57956-7_58Markdown
[Matsuno et al. "Recognition of Human Facial Expressions Without Feature Extraction." European Conference on Computer Vision, 1994.](https://mlanthology.org/eccv/1994/matsuno1994eccv-recognition/) doi:10.1007/3-540-57956-7_58BibTeX
@inproceedings{matsuno1994eccv-recognition,
title = {{Recognition of Human Facial Expressions Without Feature Extraction}},
author = {Matsuno, Katsuhiro and Lee, Chil-Woo and Tsuji, Saburo},
booktitle = {European Conference on Computer Vision},
year = {1994},
pages = {513-520},
doi = {10.1007/3-540-57956-7_58},
url = {https://mlanthology.org/eccv/1994/matsuno1994eccv-recognition/}
}