面向人脸识别的复杂光照下图像细节增强算法

为了提高在光照过度、不足或不均等复杂光照条件下的人脸识别率,提出一种复杂光照条件的人脸图像细节强化算法。首先采用对数和非线性变换对人脸图像动态范围进行压缩;然后利用反锐化掩模滤波算法消除图像模糊,增强人脸图像细节信息;最后采用Adaboost算法建立人脸分类器,并采用Yale B人脸图像数据进行仿真测试。仿真结果表明,该算法解决了复杂光照条件对人脸图像的不利影响,并进一步提高了人脸识别率。

Details strengthen algorithm for face image recognition in the complex illumination conditions

In order to improve the rate of face recognition in excessive, inadequate, or non-uniform complex illumination conditions, this paper proposes a Details strengthen algorithm for face image in the complex illumination conditions. Firstly, logarithmic and nonlinear transform are used to compress the dynamic range of face image; and then the unsharp mask filtering algorithm is used to remove the fuzzy image message to enhance the details of face image, finally, adaboost algorithm is used to build a classifier for face recognition, and the simulation test is carried out on Yale B face database. The simulation results show that the proposed algorithm has solved problem for face recognition in complex illumination conditions and improved the face recognition rate.