Expression recognition method

An expression recognition method is disclosed, which utilizes a connection-type Adaboost method to detect a human face location. After extracting a human face image, a local derivative pattern method is used to retrieve human facial expression feature values, and then a support vector machine classi...

Full description

Saved in:
Bibliographic Details
Main Authors WU, YU-SHAN, LI, TING-WEI, LIU, HENG-SONG
Format Patent
LanguageChinese
English
Published 16.03.2013
Subjects
CALCULATING
COMPUTING
COUNTING
HANDLING RECORD CARRIERS
PHYSICS
PRESENTATION OF DATA
RECOGNITION OF DATA
RECORD CARRIERS
Online AccessGet full text

Cover

More Information
Summary:An expression recognition method is disclosed, which utilizes a connection-type Adaboost method to detect a human face location. After extracting a human face image, a local derivative pattern method is used to retrieve human facial expression feature values, and then a support vector machine classification method is used for determining an expression type, wherein seven types of expressions can be recognized, including: anger, disgust, fear, happiness, normal, sadness and surprise, and an average recognition rate as high as 83.56% can be reached.
Bibliography:Application Number: TW20110132012