一种基于KL变换和镶嵌图模型的人脸检测方法

提出了一种基于KL变换和镶嵌图模型的人脸检测方法。该方法首先采用KL变换将整个图像空间划分为主元子空间和次元子空间，利用人脸图像大部分能量集中在主元子空间、小部分能量分散到次元子空间这一特点，将人脸从复杂背景中初步检测出来。在此基础上，采用符合人脸生理结构特征的人脸镶嵌图模型，并在分析足够多的人脸图像样本的基础上，针对人脸图像的灰度和边缘信息，建立了一种较为合适的知识库。实验结果表明，该方法能够较好地解决复杂背景下的正面人脸的检测问题。

A Method for Human Face Detection Based on KL Transform and Mosaic Model

An approach based on KL transform and mosaic image model for human face detection is proposed. In the method, the image space is partitioned to primary space and subordinate space by the method of KL transform, by which the human face can be detected from complex scene coarsely. Moreover, a mosaic image model is supplied, which is consistent with physical structure features of human face. Basing on this model, knowledge base, including information of edge and gray in image, is established by analyzing fairly enough human face image. The result of the experiment shows that this approach can solve the problem of multi human face detection from complex scene.