Call For Paper Volume:4 Issue:10 Oct'2017 |

LOCAL DIRECTIONAL NUMBER BASED CLASSIFICATION AND RECOGNITION OF EXPRESSIONS USING SUBSPACE METHODS

Publication Date : 19/05/2015



Author(s) :

Augi K Joseph , Prof. Arya Mohan.


Volume/Issue :
Volume 2
,
Issue 5
(05 - 2015)



Abstract :

Facial expression is immediate means of non-verbal behavioral communication of internal emotions and intensions by human beings. Facial expression recognition is a challenging task in Biometrics and in Forensics. This paper proposes a robust and efficient subspace method for Facial Expression Recognition (FER) at different critical conditions like illumination, pose, age variations and different partial occlusions. In this paper we implement appearance based linear subspace methods like PCA and ICA for dimensionality reduction of images and LDN for extracting the features of image. These sub-space methods extracts the local features from face textures and encodes it in a compact code using directional and sign information and provides more discriminate information as compared to other existing systems. Finally different facial expressions due to human internal emotions are classified using classifiers like SVM. And for the testing of the image we use JAFFE databases and can create our own database.


No. of Downloads :

10


Indexing

Web Design MymensinghPremium WordPress ThemesWeb Development

LOCAL DIRECTIONAL NUMBER BASED CLASSIFICATION AND RECOGNITION OF EXPRESSIONS USING SUBSPACE METHODS

May 16, 2015