一种支持Java环境的快速人脸检测方法

One Rapid Face Detection Method Support under Java Environment

针对原始的AdaBoost算法搜索时间长的缺点，提出了一种结合肤色分割，形态学操作，轮廓提取的改进的快速人脸检测方法。由于该方法是在基于Opencv开源视觉库下实现的，为了克服只能在C／C＋＋环境下实现该方法的局限性，本文利用SUN公司提供的本地方法接口，将其移植到了java平台下，只要稍加改进，即可将该方法应用到Android等以java为开发语言的嵌入式操作系统下，该方法在提高速度的同时，也保持了较高的实用性。

This paper proposes a fast face detection method based on skin detection,morphological operation and contour extraction to overcome the disadvantage of the traditional Adaboost face detection way.Because of open source computer vision library,in order to overcome the limitation only under in C/C＋＋ environment,we use the Java native interface which provided by Sun and transplant this method to Java platform.As long as the slightly improved,we can take this application in Android embed operating system that is programmed by java language.This method improves the speed,and maintains the practicability.