Journal
MULTIMEDIA TOOLS AND APPLICATIONS
Volume 76, Issue 2, Pages 2995-3010Publisher
SPRINGER
DOI: 10.1007/s11042-016-3282-9
Keywords
Facial expression recognition; Extended local binary pattern; Covariance matrix transform; KLT; Machine learning
Categories
Funding
- National Natural Science Foundation of China [11172342, 11372167]
- Fundamental Research Funds for the Central Universities [GK201405007]
- Interdisciplinary Incubation Project of Learning Science of Shaanxi Normal University
- Program of Key Science and Technology Innovation Team in Shaanxi Province [2014KTC-18]
Ask authors/readers for more resources
According to the deficiencies of Local Binary Pattern (LBP), the dimension of extraction is large, and it is not conducive to describe all characteristics of image texture, this paper proposes a novel facial expression recognition algorithm K-ELBP which uses uniform patterns of Extended Local Binary Pattern (ELBP), and combines with the covariance matrix transform in K-L transform (KLT). In this paper, ELBP is used for the first step to extract the feature matrix of expression images, then covariance matrix transform is applied to the ELBP matrix for reducing the dimension, which aims at extracting the main feature vectors. And the best recognition performance is obtained by using SVM for classification. A series of experiments by using different divided methods are designed to evaluate the effects of characteristics which are extracted by K-ELBP algorithm. According to the results of the experiments, the proposed K-ELBP algorithm can extract facial expression features effectively, and the rates of recognition are satisfying.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available