| Summary: | 碩士 === 國立高雄應用科技大學 === 電機工程系博碩士班 === 105 === In This thesis,build a face recognition system via Mirosoft Kinect for external device , detecting the face and extract the mouth and eyes by Haar-Like Features,proposing shadows light affects facial features distinguish presented problems and corresponding treatments to improve recognition of different brightness and under different lighting angle. In addition proposed to judge pupil location, mouth opening and closing the eyes of recognition methods, through the results of the recognition that can be used to determine the current facial expression, the facial expression can be classified as expressionless, smiling , laughing , giggllng ,surprising, and yawned.
This thesis face recognition system recognize accuracy of experimental tests of approximately 96.5% , Experimental results show paper expression differentiation of main system can accurately and effectively distinguishing knowledge from the actual environment. Can monitor the user's daily mood and expression information statistical data should be used in the home or community care centers, and other places, as participants need timely basis for giving care and companionship.
|
|---|
