人脸图像中斑点自动去除系统的设计与实现

人脸图像中斑点的自动去除是高清晰人脸图像美化的重要步骤。本文提出了一种人脸图像斑点自动去除系统的设计思路和实现方法,系统首先采用AdaBoost算法从输入图像中检测人脸区域和双眼位置,然后利用人脸部件分布的先验知识检测出眼睛、鼻子、嘴巴等部件位置,最后利用迭代多级中值滤波方法对人脸非部件区域进行平滑处理。整个系统在VC++6.0环境下借助OpenCV图像处理库实现。实验结果表明,本系统能较好地去除脸部斑点,实现人脸图像美化。

Design and Implementation of Automatic Spots Removing System from Human Face Images

Automatic removing of spots is an important step for the beautification of high-definition human face image. In this paper the design and the implementation method of an automatic spots removing system from the face image are proposed. The system firstly detects face and eyes from an input image using AdaBoost algorithm, then apriori knowledge about the layout of face components, such as eyes, nose and mouth is used to roughly locate these face parts and finally an iterated multistage median filtering method is employed for the smoothing of non-components face area. The whole system is implemented using OpenCV under VC＋＋6.0. Experimental results show that it can effectively remove spots from human faces.