摘要
人脸检测在自动人脸鉴别工作中具有重要的意义 ,由于人脸图像特征的复杂性和多样性 ,使得人脸模式分类器的训练十分困难。本文提出了一种基于支持向量机 (SVM)的人脸检测算法 ,使用了奇异值分解对训练样本进行特征提取 ,再由SVM分类器进行分类 ,有效的降低了训练难度 ,采用二阶多项式作为SVM分类器的核函数。实验结果表明 。
Face detection is important in automatic face recognition.The difficulty in training a face pattern classifier as face detector is due to the diversity and complexity of face characteristic.In this paper,we propose a Support Vector Machine-based method for face detection.Singular value decomposition is used to extract and select the appropriate features of human faces,which greatly reduces the complexity of training SVM.The second-order polynomial is used as the kernel function.The experimental results demonstrate the feasibility of this approach.
出处
《计算机与数字工程》
2003年第1期69-72,共4页
Computer & Digital Engineering
