研究了基于内容敏感图像过滤中的若干关键

技术, 分析比较了三种常用的肤色检测算法, 利用统计阈值法评价了各个特征分类图像的能力强弱, 并依此实现了一个决策树

敏感图像过滤器。为有效降低肖像类正常图像的误判率, 又在过滤器中使用了人脸检测机制-Of context-sensitive image filtering based on the number of key technologies, analysis and comparison of the three common skin detection algorithm, evaluated using statistical threshold of each image feature classification strengths and weaknesses, and implement a decision tree so sensitive image filter. Portrait class in order to reduce false positives rate of the normal image, but also used in the filter mechanism of face detection