test face meaning in Chinese

试验工祖
试验工作面

However , some of face recognition problems still require further development , this is the case for problems of recognition face images conveying changes in illumination , facial expression and changes due to the time delay between the acquistion of the reference and tested face images . our main work is to analysis methods of extraction face features and contraction of classifier . the work presented in this paper is to apply self - organizing feature map and minor component to extraction features from multi - view face images , then combine those features as a new combined feature set , in order to reduce redundancy data , we apply clone algorithms to reduce data through rotation in input space
我们改进了一种基于矩理论的识别方法，给出了计算公式和证明过程，可用于解决小规模人脸识别问题；我们将智能方法应用到人脸识别中，分别利用自组织特征映射和次分量方法抽取人脸的整体特征和局部特征，依据特征融合理论，重新组合为新的复合特征，为压缩特征数据，我们首次引入克隆选择算法自动进行特征优化选择，最后，利用支持矢量机构造多分类器进行分类识别，在不同规模人脸识别库上仿真结果表明，该系统自适应能力强，分类识别精度高，适用于大规模复杂人脸识别问题。