Research On Face Recognition And Tracking Technology In Instant Message

With the technology of the computer and robot development, robot will come into the family life little by little. There are lots of methods for human-robot interaction. The paper contains the filed of robotics, human-robot interaction, patterns recognition, and so on. It exploits human-robot interaction as a target. It is one of methods for face recognition and tracking. It mainly includes three aspects as fallows.1. Face RecognitionFirst, we need judge whether there is face in vision of camera for the object recognition. A face detection algorithm is proposed in this paper, which adopts Adaboost detection based on Haar feature. To import the accuracy, we can set up the YCbCr of Mix-Gaussian model. Then using the face eigenfaces of PCA, recognize face. It is proved that it can accurately and real-time detect and recognize through the algorithms.2. Face TrackingTaking into account the interference, such as background, it adapts the particle filtering algorithm with color and context feature. We use the HSV model and LBP (Local Binary Pattern) operator to extract context feature. Experimental results show that the proposed method is of higher robustness and provides more accurate results in the Instant Message Robot.3. The System of Face Recognition and Tracking Implement in Instant Message RobotThe paper induces the scheme on hardware and software. It mainly introduces some parameters of hardware, such as COMS camera, Cortex-M3 LPC1800 processor and console. It also shows procedures of implementers'software in the Visual Studio 2008.