| 基于C4.5算法的敏感图像检测方法 |
| |
| 引用本文： | 罗森林,辛子英,冯磊.基于C4.5算法的敏感图像检测方法[J].北京理工大学学报,2006,26(5):451-453. |
| |
| 作者姓名： | 罗森林 辛子英 冯磊 |
| |
| 作者单位： | 北京理工大学,信息科学技术学院电子工程系,北京,100081;北京理工大学,信息科学技术学院电子工程系,北京,100081;北京理工大学,信息科学技术学院电子工程系,北京,100081 |
| |
| 摘 要： | 提出一种基于2次C4.5分类的敏感图像检测方法.该方法利用Daubechies小波和灰度共生矩阵提取图片的纹理信息,在HSI空间和YCbCr空间提取颜色特征,进行第1次基于C4.5算法的训练和分类;对分类生成的0,1二值空间进行特征提取,将所有特征分量融合进行基于C4.5算法和2次训练和分类.基于2次训练生成的规则进行敏感图片的检测.实验结果表明,该方法对于敏感图片分类正确率达93.3%以上,与基于颜色和纹理特征的直接检测方法相比,进一步提高了正确识别率.
|
| 关 键 词： | 敏感图像 皮肤检测 颜色空间 C4.5算法 |
| 文章编号： | 1001-0645(2006)05-0451-03 |
| 收稿时间： | 12 8 2005 12:00AM |
| 修稿时间： | 2005年12月8日 |
Sensitive Image Detection Method Based on C4.5 Algorithm |
| |
| LUO Sen-lin,XIN Zi-ying and FENG Lei.Sensitive Image Detection Method Based on C4.5 Algorithm[J].Journal of Beijing Institute of Technology(Natural Science Edition),2006,26(5):451-453. |
| |
| Authors: | LUO Sen-lin XIN Zi-ying and FENG Lei |
| |
| Affiliation: | Department of Electronic Engineering, School of Information Science and Technology, Beijing Institute of Technology, Beijing 100081, China |
| |
| Abstract: | Presents a new method with twice classification.The shin texture features with Daubechies wavelet and gray concurrence matrix is extracted,and the color features in HSI space and YCbCr space are also extracted.These parameters are given to the C4.5 classifier to be trained and classified.The parameters are further extracted again in the 0,1 space generated.All the parameters are synthesized as inputs to the C4.5 algorithm.At last,the image is dealt with the rules in C4.5.Experimental results showed that this algorithm depicted better performance with correct rates reaching 93.3%. |
| |
| Keywords: | sensitive image skin detection color space C4 5 algorithm |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《北京理工大学学报》浏览原始摘要信息 |
|
点击此处可从《北京理工大学学报》下载全文 |
|
