基于内容的敏感图像检测方法研究

随着网络和多媒体技术的发展,互联网中色情图像的传播越来越泛滥,为有效地过滤敏感图像,文章提出一种基于Otsu自适应阈值分割的人体面积比例计算方法。整个检测系统利用人体肤色模型和面部识别模型,并结合面积比例识别算法等图像特征识别技术,实现对网络敏感图像的检测。实验结果表明,该方法具有较高的准确率和较快的实时在线检测速度,具有良好的实用性和应用价值。

A Method for Detecting Sensitive Images Online Based on Content

To detect and filter illegal pornographic images effectively,this paper proposes an identification method of area ratio which based on adaptive threshold segmentation of Otsu.The detection system aims to realize the filtering of pornographic images with combining skin color model,face recognition model and identification technology of image features,such as the identification algorithms of area ratio.The results show that this method has higher accuracy and quick speed of on-line detection,and has better practicability and application value.