Face detection is a key step in computer vision applications, such as face recognition and video surveillance. There is a
difficult question in automatically segmenting color images into skin color and background regions when using the
method of skin color segmentation. The border of skin color region has some background noise patches, especially using
edge detection to extract facial organ characteristics, which will lead to some redundant gradient features near the border.
In this paper, we introduce a new method based on the histogram to smooth face contour and reduce background noise
patches. The performance of our face detector based on the smoothing contour algorithm has been tested on reference
test sets, and clearly show some improvements in face detection which is based on knowledge-based methodology.