| Summary: | 碩士 === 國立勤益科技大學 === 電機工程系 === 106 === Face recognition is an identification technology using personal facial features, and it always is a popular research topic. The recognition technique using face recognition is widely applied in various control systems such as computer login authentication, access control, and customs inspection. In this paper, we propose a face recognition system based on stereo vision algorithm and principal component analysis. The system captures images using two parallel-arranged webcams simulating people binocular vision. The depth information of the image target object is calculated using stereo vision algorithm, which determines whether the target image is a three-dimensional object; and the mistake in recognizing a flat face photo as a person will be eliminated. Finally, face recognition is performed using principal component analysis algorithm (PCA), linear recognition analysis algorithm (LDA) and independent component analysis algorithm (ICA). Our research identifies faces with different people and photos, and experiments under different illumination and distance conditions. The experimental results show that the stereo vision with PCA has the best recognition rate. The people face recognition system in this study eliminates the flat face photos and realizes the real people identify.
|
|---|
