Research On The Technology Of Gender Recognition Based On Face Image

In recent years,With the further study of face recognition,gender recognition become the hotspot of computer vision and pattern recognition.Gender recognition is that use computer technology to identify human gender information by the face image.which have important application prospects in artificial intelligence,system monitor,pattern recognition and so on.After analyzing the methods currently used by others,we present an Fusion algorithms of independent component analysis(ica) and Kernel Fisher Discriminant Analysis(KFDA),an improved KNN-SVM algorithm for gender recognition and an curvature algorithm of chin profile curve.Works are described as below:(1)A feature extraction algorithm which fuse ica algorithm and KFDA is proposed.Feature imformation which extracted by ICA algorithm can reconstruct more original data imformation and extracted by Fishe Discriminant Analysis can get more optimal classification information.Because of the influence of light,posture,expression,etc.so image data is nonlinear usually.but Fisher Discriminant Analysis only get linear information.therefore,we use KFDA algorithm and reduce time complexity of it by PCA.Images after preprocessing project to the subpace of ICA and KFDA.(2)An improvement algorithm Improved KNN-SVM that combined Support Vector Machine(SVM) with weigthted K Nearest neighbour(KNN)is presented to improve the accuracy of gender recognition nearby SVM hyperplane.The algorithm gets optimal threshold by a few of known gender samples,then compute the distances from the test samples to the optimal superplane of SVM in feathure space,recognize gender after comparing the distance to threshold.(3)A new Gender recognition method base on the curvature of face chin profile curve is presented.in order to get clear face chin profile curve.This method get the texture image by log-gabor algorithm firstly,then extract the feature of face chin profile curve through Active Appearance Model(AAM) moreover curve fitting by least squares.Get gender result by the curve's curvature lastly. (4)A prototype system of face gender recognition based on holistic face feature and chin profile curve feature.It can be used to prove the effectiveness of above algorithms.