Research On Facial Expression Recognition In Simplified Mode

In the recent years,scholars from all levels of society pay more attention to the research of emotional robots;it has become a new research hot shot of artificial intelligence field.This so-called emotional,it does not only make the robot to own human emotions but also can understand human emotions and adjust their own emotions to respond to human.The primary issue is how to make robots to recognize and understand the emotion from human facial expressions before the human-machines emotion communities are implemented.This paper is based on psychological theory,emotional state classification algorithm,image processing,and recognition research.The main content is to establish a facial expression recognition framework based on Gabor wavelet and local binary pattern algorithm and use the MFC framework of Visual C++ 6.0 to build a verification platform,so as to further study the mechanism of human emotion recognition and understanding.Thus finish the research work of the recognition of facial expression and emotion by intelligent systems such as computers and robots.The main content work of this thesis is focused on the following aspects:Firstly,introduce the background and the state-of-art of human facial expression recognition methods,and review on the techniques and methods that are currently used for human facial expression recognition,and on the understanding of these existing methods through a lot of experiments and analysis.Secondly,research on the facial expression detection and recognition methods based on Gabor wavelet and local binary pattern algorithm.The first step of the method is using the Gabor wavelet transform to extract features from normalized facial expression images.The second step is to set up a set of feature template samples based on the extracted feature vectors.The minimum energy and K nearest neighbors between the tested expression image and the template sample can be calculated by the improved K nearest neighbor classification method.Finally,we get the final expression classification result after comparative analysis.Be simulated by MATLAB,the functions of the different methods of processing are compared and analyzed,further prove the validity and reliability of the algorithm for the model established in this paper.Thirdly,designing the facial expression recognition system.The facial expression recognition system is built by the MFC framework of Visual C++6 software.Extracting and recognizing the features of facial expression and get the emotional state of facialexpression.Finally,the recognition results are compared with the sample labels to verify the reliability of the system in facial expression recognition.