摘要
针对目前人脸美貌度评价方法计算复杂等不足,根据黄金比例的美学标准与人脸面部的三分之一比例,提出了一种基于图像的非监督人脸美貌度自动评价方法。基于人脸所具有的类Haar特征,定位人脸区域,提取人脸面部相关特征并确定其位置,结合人脸肤色的聚类性确定人脸边界。根据人脸面部各特征之间所具有的不同比例及其评分权重,通过计算所提取的人脸相关特征各比例与人脸美学人脸特征比例之间的差异度,对人脸美貌进行非监督自动评价。通过实验对比,验证了文中所提方法有效、可行。
Aiming at some limits existing in some literatures with complex computation,a new method of unsupervised evaluating human appearance beauty is developed based on the golden proportion distribution in aesthetic standard of the face appearance and the facial thirds in facial region.The facial region is extracted according to Haar-like features in the face at first.Some facial features are extracted and determined their locations.The facial boundary is ascertained based on skin clutter characteristics.There are some different proportions and evaluating weights in the extracted facial features.The facial appearance beauty is automatic evaluated by computing the difference between the proportions of the extracted facial features and the ones in the aesthetic standard of the face appearance.The performance of the proposed method is compared with some similar methods,which shows the proposal performs is better.The experimental results highlight that the developed method estimates beauty efficiently.
出处
《电子器件》
CAS
北大核心
2012年第2期244-248,共5页
Chinese Journal of Electron Devices
基金
国家自然科学基金项目(60872117
11176016
61001162)
关键词
美貌度评价
人脸检测
特征定位
黄金比例
facial appearance beauty evaluation
face detection
feature location
golden proportion
