基于彩色分量双重鉴别相似性的人脸特征提取

针对当前人脸彩色图像鉴别特征提取方法存在特征鉴别能力低、识别效果差等问题,提出基于彩色分量特征层双重鉴别相似性分析的人脸图像鉴别特征提取方法.通过使用人脸彩色图像中各个彩色分量数据和不同彩色分量数据集之间的相关性进行特征层鉴别相似性分析,并设计了R、G和B三个彩色分量图像数据集鉴别特征提取流程.仿真结果表明:CM-DDSA方法保留了三个彩色分量特征之间的平均相似性,同时在很大程度上提升了特征鉴别能力以及识别效果.

Face feature extraction based on double authentication similarity of color component

In order to solve the problem that the current feature extraction methods for the color face image authentication have low authentication capacity and poor recognition effect, a feature extraction method for the face image authentication based on double authentication similarity analysis of color component feature layer was proposed. The feature layer authentication similarity analysis was carried out through utilizing the correlation between each color component data and different color component data in the color face images. In addition, the feature extraction process for the authentication of R, G and B three color component image data sets was designed. The results show that the CM-DDSA method retains the average similarity between three color components, and the feature authentication ability and recognition effect get enhanced to a large extent at the same time.