联合低秩和e_p稀疏约束矩阵回归的人脸识别算法

Face Recognition Based on Matrix Regression with Low-rank ande_p Sparse Constraints

针对遮挡和光照等因素影响的人脸图像,提出一种具有低秩稀疏性的矩阵回归模型。该模型采用低秩性约束回归误差,采用p范数约束回归系数使其达到稀疏最大化,然后通过广义迭代阈值算法求解p范数,最后用交替方向法求解模型参数。在AR和Extended Yale B人脸数据库上的实验表明,与当前的回归算法相比,该算法具有更高的识别率,能够更好地消除由遮挡引起的结构性噪声,且对光照变化也具有更强的鲁棒性。

This paper presented a model of matrix regression for face recognition to deal with varying illumination,as well as occlusion and disguise.To ensure low rank and sparse prosperities of the model,we used low rankness to constraint the regression error,and used thep-norm to constraint the regression coefficients in order to guarantee the sparest solution.We applied generalized iterated shrinkage algorithm forp-norm,and alternating direction method for regression coefficients.Experiment results on face database of AR and Extended Yale B show that the face recognition method proposed in this paper has a higher recognition rate than the current regression methods.And our method is more powerful for removing the structural noise caused by occlusion,and more robust for alleviating the effect of illumination.