Facial Expression Recognition On Local Binary Pattern And Local Phase Quantity

Facial expression is the carrier of human emotion,as an important information in human daily communication,reflecting the human psychological state.People convey to others different emotional information through the tiny changes in expression.With the continuous development of science and technology,the requirements of human-computer interaction are more diversified.How to make the computer read the facial expression information has become a hot research topic.Nowadays,facial expression recognition technology has been paid more and more attention and has been widely used in many fields.In this paper,the related algorithms used in the three steps of the face image recognition process are studied in depth.The research content and innovative work include the following:(1)For the pre-processing segment of pure facial image cutting problem,the use of integral projection direction,positioning the eye and the eye precise area.According to the traditional "three-court five-face" facial model,the pure facial image is cut out,and the normalized image is scaled to 128 × 128 by using bilinear interpolation to calculate the interpolation gray scale.The method of image equalization is used to normalize the image.Finally,the standard expression image is obtained,which lays the foundation for the following image feature extraction.(2)A feature extraction algorithm combining local binary pattern and local phase quantity is proposed.The local facial expression is transformed into the corresponding characteristic spectrum by using the local binary value and the local phase quantity.The optimal segmentation strategy is determined by the expression recognition rate experiment under different blocks.Then the two characteristic spectra are transformed into histogram feature sequences,and two histogram feature sequences are connected in series,which solves the problem that the local feature extraction facial expression information is lost.Compared with single local binary algorithm,local phase quantity algorithm and two-dimensional Gabor wavelet algorithm,a high recognition rate is obtained.(3)A face feature extraction and recognition algorithm based on depth belief network model is proposed.The depth search algorithm is used to determine the optimal parameters and the degree of belief network depth and iteration number.Compared with the traditional support vector machine algorithm,the overall recognition rate of the algorithm is improved by 4.1%.(4)Aiming at the problems of the traditional form recognition system,such as single form,poor scalability and incompatibility of mobile devices.An online facial expression recognition system based on B/S structure is designed.This system solves the PC,tablet,mobile phones and other equipment compatibility issues,to meet the response-style layout of the design concept,for different devices,as the mouth,to show different page styles.