Automatic Nipple Detection Using Cascaded AdaBoost Classifier

摘要：

Many non-pornographic images containing large exposure of skin area or approximate skin-color area are prone to be detected as the pornographic images. This paper proposes a novel method based on AdaBoost algorithm for nipple detection of pornographic images. The AdaBoost algorithm has excellent performance in both detection accuracy and detection speed. The method extracts extended Haar-like features, color features, texture features and shape features to train and obtain a cascaded AdaBoost classifier by using AdaBoost algorithm. And it is validated for locating nipple existence in pornographic images. The experimental results show that this method performs well for nipple detection in pornographic images, and can reduce effectively the false positive rate against the non-pornographic images.

关键词： AdaBoost learning algorithm Haar-like feature color features texture features shape features cascaded AdaBoost classifier

作者: Xin Kejun Wu Jian Ni Pengyu Huang Jie

作者单位: Nanjing Sampel Technology Co., Ltd. Nanjing, Jiangsu Province, China School of Information Science and Engineering Southeast University Nanjing, Jiangsu Province, China

会议类型: 国际会议

会议名称: 2012 Fifth International Symposium on Computational Intelligence and Design 第五届计算智能与设计国际会议 ISCID 2012

会议地点: 杭州

会议语种:英文

页码: 1005-1010

在线出版日期: 2012-10-28（万方平台首次上网日期，不代表论文的发表时间）