Traditional Support Vector Machine(SVM)when faced with a large sample training problem, the sample size is limited by hardware. Therefore, this paper proposes a face image gender classification algorithm based on hierarchy SVM and classifier fusion. Hierarchy SVM classifier can train the different samples in several levels by setting the threshold. Meanwhile, in the cascade of each layer, in order to reduce the influence of various factors which in the process of classifier recognition, it makes optimal classifier trained by different feature dimensions to fusion and through educe error in fusion to make neutral face samples have more clear classification. The experimental results under the same hardware conditions show that only 70, 000 samples can be contained one time to train by one-layer SVM, while more than 120, 000 samples are involved in four-layers SVM, the corresponding recognition rate is 96.7%vs. 99.1%.%传统的支持向量机(Support Vector Machines,SVM)在面对大样本训练问题时,其样本数量会受到内存的限制.因此,提出一种基于级联SVM和分类器融合的人脸图像性别识别方法.级联SVM分类器可以通过设定阈值将识别难易程度不同的样本分成若干层次来进行训练;同时,在级联的每一层上,为了降低分类器在识别过程中受各种因素的影响,对不同特征维数下得到的最优分类器进行融合,通过融合减小误差,使中性的人脸样本有更明确的分类.在同一硬件条件下的实验结果表明,单层SVM最多只能训练7万样本,而四层级联SVM训练样本数可达12万以上,相应的识别率也从单层融合前的96.7%上升至四层融合后的99.1%.
展开▼
