自适应多阶段线性重构表示分类的人脸识别

针对以往基于表示的分类(RBC)方法在类别数较多的数据集上性能不佳的问题，提出了一种自适应多阶段线性重构表示的分类(MPRBC)方法。在每一阶段，首先得到L₁范数或L₂范数正则化的重构表示系数，然后将表示系数按类求和，根据和的大小来选取相似类，并保留相似类中的全部样本作为下一阶段的训练样本。该策略最终产生具有高分类置信度的稀疏类概率分布，根据类系数的大小自适应选择相似的类，提高了分类计算的效率。实验结果表明，该方法分类性能优于其他RBC方法，特别是在类别数较多的数据集上性能提升明显，并且CPU时间保持相对较低水平。

Self-adaptive multi-phase linear reconstruction representation based classification for face recognition

Representation-based classification (RBC) methods have recently been the promising pattern recognition technologies for object recognition. The representation coefficients of RBC as the linear reconstruction measure can be well used for classifying objects. But RBC methods performs very poorly on large-class-databases and in order to solve the problem of poor performance, a self-adaptive multi-phase linear reconstruction representation based classification (MPRBC) method is proposed. In this process, at first, the reconstruction coefficients regularized by L₁-norm or L₂-norm are obtained. Then the similar classes are selected according to the sum of the representation coefficients in each class, and all samples of similar classes are retained as training samples for the next stage. This strategy finally produces a sparse class probability distribution with higher classification confidence. The similar classes are selected adaptively according to the values of class coefficients, which improves the efficiency of the classification. Experimental results show that the proposed method is better than other RBC methods, especially on large-class-databases, and CPU time remains relatively low.