Face expression recognition method based on LBP (Local Binary Pattern) and LDA (Linear Discriminant Analysis)

Abstract

The invention provides a face expression recognition method based on LBP (Local Binary Pattern) and LDA (Linear Discriminant Analysis). The method comprises the following steps that: 1, the Adaboost algorithm is utilized to perform face detection, and face location positioning is performed on the image of a detected face; 2, LBP feature extraction is performed on the positioned face image; 3, the LDA algorithm is utilized to perform dimension reduction on extracted features; and 4, features of a JAFFE image database which have been subjected to dimension reduction are adopted as the classification sample features of the KNN (K-nearest neighbor) classification algorithm, feature extraction is performed on an image to be detected, dimension reduction is performed, and the KNN is adopted to perform expression classification, so that face expression recognition is realized. With the face expression recognition method of the invention adopted, the influence of illumination change on face expression recognition is reduced; and the image data are subjected to dimension reduction, and therefore, calculation amount is greatly reduced, and recognition efficiency is improved.

Description

technical field [0001] The facial expression recognition method based on LBP and LDA belongs to the fields of computer vision, pattern recognition, artificial intelligence and so on. Background technique [0002] With the advent of the information age, computers have quietly changed people's way of life. In the process of frequent use of computers, people are increasingly looking forward to a more friendly human-computer interaction experience. Facial expressions, as part of human biological characteristics, reflect the complex and subtle emotional changes in people's hearts, and convey a large amount of emotional information. The research on facial expression recognition has an extremely important role and significance in improving people's quality of life and providing people with a convenient and quick way of life. [0003] After more than 40 years of efforts by many researchers and institutions, facial expression recognition has formed a relatively complete process, suc...

AI Technical Summary

Problems solved by technology

[0004] Facial expression recognition research has made great progress, and has also obtained fruitful results. The Chinese patent with the publication number CN102945361B discloses a facial expression recognition method based on feature point vectors and texture deformation energy parameters. The method considers different The difference between individuals solves the influence of personality information, small-scale head posture and other factors on facial expression recognition, but it also encounters many research difficulties, such as lighting changes, occlusion, posture changes, and image data dimensions are too high. Calculation problems such as large quantity

Images