摘要
设计并实现了一个敏感图片自动识别系统。该系统首先利用肤色模型从敏感图片中检测出肤色区域,再从图片肤色区域中,提取大量经验特征表示图像内容,然后采用Haar特征算法设计一个检测敏感部位对象的分类器,最后通过敏感部位的几何特征逐级识别正常图片与敏感图片。该检测系统具有良好的系统性能。
An automatic detection system for sexual pictures is designed and implemented.This system extracts considerable experience characteristics to show the pictures from the picture skin color region which is detected in the sexual pictures by means of the skin model, then a classifier is designed for the detection of sexual parts of the body by HAAR algorithm to carry on machine learning for the detection of sexual parts of the body, finally the normal pictures and the sexual pictures can be identified by way of geometric features of the sexual parts step by step.The system shows a great System Performance.
出处
《微计算机信息》
2010年第8期205-207,共3页
Control & Automation
基金
基金申请人:陈黎
项目名称:动态场景下的图像超分辨率重建
基金颁发部门:国家自然科学基金委会(60705012)
关键词
敏感图片
肤色模型
几何特征
分类器
sexual picture
skin color model
geometric features
classifier