针对复杂光照变化影响人脸识别准确性的问题,提出了一种基于多尺度韦伯脸和梯度脸相结合的复杂光照下人脸识别方法。首先定义了能够有效描述人脸纹理结构的多尺度韦伯脸,一定程度上减弱了不同光照条件的影响;其次融合多尺度韦伯脸和梯度脸提取人脸光照不变量;最后利用SVM多类分类算法实现人脸识别。使用CMU PIE与Yale B人脸库进行验证,结果表明:提出的算法能够有效消除复杂光照变化对人脸识别的影响,即在光照极差情况下,单样本图像作为训练图像也可以有很好的识别效果,且识别率显著高于韦伯脸、多尺度韦伯脸和梯度脸。
Abstract
Aiming at the problem of the effect of Complex Illumination on face recognition rates, an integrated method based multi-scale Weber-face and gradient-face is proposed. The method starts with defining the multi-scale Weber-face, which can effectively describe the face texture structure and reduce the influence of different lighting to some extent. Then multiscale face and GF are fused to extract the face illumination invariant. Finally, multi-class support vector machine is employed for face authentication. The experiments are executed both on CMU-PIE and Yale B face databases. The experimental results have indicated that the proposed method can effectively eliminate the influence of face recognition under complex illumination and the recognition rates are superior to Weber-face, multiscale Weber-face and gradient-face, even single sample images with serious light as training sample images can also work well.
关键词
计量学 /
人脸识别 /
复杂光照 /
多尺度韦伯脸 /
梯度脸
Key words
metrology /
face recognition /
complex illumination /
multi-scale Weber-face /
gradient-face
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
［1］王成, 郭飞, 赖雄鸣, 等. 多尺度加权 LBP 的人脸识别［J］. 光电工程, 2014, 41(4): 82-88.
［2］Zou W, Yuen P, Chellappa R. Low-resolution face tracker robust to illumination variations［J］. IEEE Transactions on Image Processing, 2013, 22: 1726-1739.
［3］Xie X, Lam K M. Face recognition under varying illumination based on a 2D face shape model［J］. Pattern Recognition, 2005, 38(2): 221-230.
［4］Hyunjung S, Jiebo L, Tsuhan C. A subspace model-based approach to face relighting under unknown lighting and poses［J］. IEEE Trans on Pattern Analysis and Machine Intelligence, 2008, 17(8): 1331-1341.
［5］Cao X, Shen W, Yu L G. Illumination invariant extraction for face recognition using neighboring wavelet coeffi-cients［J］. Pattern Recognition, 2012, 45(4): 1299-1305.
［6］Jobson D J, Rahman Z, Woodell G A. A multiscale retinex for bridging the gap between color images and the human observation of scenes［J］. IEEE Transactions on Image Processing, 1997, 6(7): 965-976.
［7］Wang H, Li S Z, Wang Y, et al. Face recognition under varying lighting conditions using self quotient image［C］//Proceedings of IEEE International Conference on Automatic Face and Gesture Recognition, 2004, 819-824.
［8］Chen T, Zhou X S, Comaniciu D, et al. Total variation models for variable lighting face recognition［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2006, 28(9): 1519-1524.
［9］辜小花. 基于视觉感知和边缘保持的光照不变人脸识别［J］. 电子学报, 2013, 41(8): 1500-1505.
［10］Zhang T, Tang Y, Fang B, et al. Face recognition under varying illumination using gradient faces［J］. IEEE Transactions on Image Processing, 2009, 18(11): 2 599-2 606.
［11］Wang B, Li W, Yang W, et al. Illumination normalization based on webers law with application to face recognition［J］. IEEE Signal Processing Letters, 2011, 18(8): 462-465.
［12］李海滨, 张强, 刘彬. 基于多尺度小波变换的子线段匹配方法研究［J］. 计量学报, 2009, 30(1): 88-92.
［13］Xu B, Tang Y Y, Fang B, et al. Multi-scale gradient invariant for face recognition under varying illumination［J］. International Journal of Pattern Recognition and Artificial Intelligence, 2012, 26(8): 1-12.
［14］孟宗,李良良. 基于局部特征尺度分解和形态学分形维数的滚动轴承故障诊断方法［J］. 计量学报,2016,37（3）:284-288.
［15］赵振华, 郝晓弘. NSCT 域自适应人脸图像光照不变特征提取［J］. 光电子·激光, 2012, 23(9): 1815-1820.
［16］Li S, Gong D, Yuan Y. Face recognition using Weber local descriptors［J］. Neurocomputing, 2013, 122:272-283.
基金
国家自然科学青年基金 (61601400); 河北省自然科学青年基金 (F2012203031)
PDF(1686 KB)
