Face feature point detection method

Abstract

The invention discloses a face feature point detection method. According to the face feature point detection method, a novel three-channel GEH (Gray-Edge-Hog) mode image with the attitude detection task acting as the constraint and multiclass feature graphs acting for fusion acts as input. Considering that face three-dimensional attitude information has considerable influence on detection of face global feature points under the condition of large attitude deflection especially and the detection complexity of the contour feature points can be effectively reduced by adding Hog feature information reflecting face image local feature presentation and edge image information used for edge detection and extracted by the Sobel operator, the novel GEH three-channel image is generated by extracting the image grayscale value, the edge information and the Hog features to act as the input, and the auxiliary task of three-dimensional attitude estimation acts as the constraint information to perform face feature point detection.

Description

technical field [0001] The invention belongs to the field of computer vision, and in particular relates to a face feature point detection method using face three-dimensional pose information as an auxiliary constraint in a new image mode, which has important applications in face recognition, face pose and expression analysis, and face synthesis . Background technique [0002] In recent years, with the development of deep learning, Convolutional Neural Networks (CNN) have achieved good results in facial feature point detection. CNN takes the original image of the face as input, and the features obtained by using the local receptive field strategy have better expressive ability; the weight sharing structure reduces the number of weights and thus reduces the complexity of the network model; at the same time, using the local correlation of the image Principle The downsampling of feature maps effectively reduces the amount of data processing while retaining useful structural inf...

AI Technical Summary

Problems solved by technology

However, this method only considers the deflection of the face posture in the horizontal dimension, and in most cases, the posture deflection of other dimensions also has a certain impact on the accuracy of feature point detection.

Images