中文摘要：

给出一种基于压缩传感理论的人脸性别辨认算法，试图解决在复杂光照环境下人脸性别的鲁棒辨认问题。在简单介绍传统人脸性别辨认方法和压缩传感理论的相关内容后，给出本文主要工作，包括创建人脸性别数据库，构造人脸性别字典基和提出性别辨认算法等。最后，算法在面向复杂光照变化环境的Extended YaleB人脸数据库子集上对人脸的性别鉴别问题进行验证。实验结果表明本文提出算法的计算效率和识别率优于传统方法；且在复杂光照环境下，基于结构的压缩传感性别辨认方法的鲁棒性要优于传统基于能量的方法。

英文摘要：

A face gender classification method based on compressed sensing theory is proposed, the aim of which is to study the robust gender classification of face image in the complex lighting environment. After briefly introducing the traditional methods of gender classification and the relevant concept of the compressed sensing theory, the main work of the paper is given, including the creation of face gender image database, the construction of face gender dictionary bases, the proposition of gender classifica tion algorithm and so on. At last, all the experiments are made on the subset of Extended Yale B face database, which is made in complex illumination, for the verification of gender classification problem. Simulation results show that the calculating efficiency and the classification accuracy of the proposed algorithm are superior to that of the traditional method; and the robustness of the compressed sensing gender classification method based on structure is better than the traditional methods based on energy under complex lighting environment.