期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2013
卷号：55
期号：1
出版社：Journal of Theoretical and Applied
摘要：In the real-time face recognition systems, locating faces in the input image should be fast and accurate. `Therefore, the need to introduce a face localization method with less complexity rises. Recently, Template-matching Approach has been used by different researchers to find the region of interest (face location) with different positions, lighting and cluttered backgrounds. However, some matching errors always occur with regard to the appearance of parts of the image as face location because of the variations of image pixels values. In this paper, faster and reliable pre-processing method is proposed to reduce the effects of the noise on image pixels values by applying a filtering process. The result showed an important and significant improvement in localization accuracy or precision compared to using the traditional NCC alone. Since this method was implemented, there has been an eleven percent (11%) increase in accuracy compared to the traditional NCC (80% localization accuracy). In this study, Yale University�s database is used to evaluate the proposed method.
关键词：Image Processing; Face Localization; Normalized Cross Correlation
