Analysis And Research Of Face Attraction Based On Global Contour And Face Shape Features

Face attractiveness has a great value of research in genetic psychology and cognitive psychology.The research results can provide scientific basis for the evolution of human face(and human evolution),and have important scientific significance.In the research of face recognition,the feasibility of fine classification of face structure in improving the stability and efficiency of face recognition has been proved by experiments.In order to improve the evaluation performance of computer evaluation of face attraction,this paper deeply analyzes the problem of global contour point feature extraction,face shape structure classification and the evaluation of face attraction of various face types.Its main work is as follows:(1)The classification of the face structure is analyzed in detail,and the face structure is classified by the global contour feature.According to the global contour point feature model,the mandibular contour feature and the global contour feature are extracted.Considering the basic geometric properties of various face structures,the face structure is classified by K-nearest neighbor and mean Hausdorff distance.Experiments show that the global contour feature gets better results in the classification of the face structure,and the classification accuracy of the face structure is improved by 5.98% compared with the mandibular contour feature.(2)Aiming at the problem of face attraction evaluation,a method of "diversity" on the face structure of face dataset is proposed.The diversity,that is,in the construction of the face structure data set,the original data set is divided into small data sets with six types of face structures.Then,this paper analyzes the face attraction evaluation scores of various face structures.The results show that elliptical and heart-shaped face structures are more attractive than other face structures.(3)Combining the classification of face structure with the evaluation of face attraction,a method of face attraction evaluation based on global contour and face structure classification is proposed.The optimal face attraction evaluation method based on various face structure data sets was found.Then,this method is compared with the face attractiveness evaluation method based on the overall data set.The experiments show that face attraction evaluation performance based on face structure diversity can get better results.And the Pearson correlation coefficient increased by an average of 6.60%.