基于对称区块的人脸图像光照补偿算法

【目的】改善人脸光照补偿算法的效果,提高人脸图像预处理的质量。【方法】利用人脸左右对称的特点,提出一个基于对称区块的人脸光照补偿算法,利用人脸区块的平均灰度值对人脸图像的对称点进行光照补偿。【结果】该方法对于人脸图像有很好的光照补偿效果,优于直方图均衡化、灰度变换等常用光照补偿方法。【结论】基于对称区块的光照补偿算法对于具有对称特点的图像有补偿效果。

Face Illumination Compensation Algorithm Based on Symmetrical Blocks

[Objective] To improve the effect of face illumination compensation algorithm and raise the quality of image preprocessing.[Methods] Considering the symmetry of human face, an illumination compensation algorithm based on symmetric blocks is proposed. In this algorithm, the mean gray values of symmetry blocks are used to compensate the corresponding points.[Results] Experimental results show that the method is more effective than common algorithms such as the histogram equalization and gray-scale transformation, and etc..[Conclusion] Illumination compensation algorithm based on symmetric blocks is effective for the images with symmetrical features.