中文摘要：

复杂光照条件下的人脸识别是一个困难但需迫切解决的问题，为此提出了一种有效的光照补偿算法。该方法根据人脸光照线性变换子空间理论——商图像理论，构造了小波低维训练集，实现了对待识别图像的光照条件估计，并且通过加光和去光策略增强了光照补偿效果。与传统商图像理论相比，该方法利用小波分解，提高的算法执行效率，实验结果表明，该算法以较小的代价取得了较高的识别性能。

英文摘要：

Face recognition under complex illumination conditions is still an difficult but must deal with problem an effective illumination compensation method is proposed. Based on quotient image theory, the illumination condition is estimated on facial wavelet de-dimension illumination training set. Moreover, the aim of facial illumination compensation is implemented by two basic strategies of adding light and reducing light. Compared with traditional quotient image theory, this method enhances arithmetic efficiency by wavelet transform. The experimental results show that the improved methods get a very competitive recognition rate with low computational cost.