摘 要：随着科技时代的进步，互联网技术在计算机的广泛使用的给人提供便利的同时，也萌发了很多危害。人像检测技术开始于上世纪 60-70 时代，经过几十年的研究与应用，已经慢慢完善。但时代的进步让识别技术的要求愈来愈高，检测敏感照片是一项重要识别类工作，所以关于这种技术和理论不断地被人们所关注，渐渐成为研究热点。本文针对不良照片中的感兴趣区域的检测提出了一些新方法，改进训练弱分类器的权值方法，高效率的分辨出不良照片，督促互联网的绿色发展。本论文研究的问题是关于 Adaboost 算法的在不良照片中的检测，但关于识别与检测有许多客观的非人为的干扰因素存在，现在我们需要的是在存在众多干扰的情况下仍然有一个检测水平高，速度快的算法。这也一直是研究人员努力的方向。所以在此基础上，提出来了 Adaboost 算法下 Haar 特征用于检测色情图片。新的算法不仅可以迅速，又可以利用这种特征排除多余的干扰因素，促进了工作的顺利进展。关键词：不良照片；敏感信息识别；人脸检测；AdaBoost；新特征；特征抽取；误检率；Abstract:With the rapid development of Internet technology and the progress of society, the Internet brings convenience and speed to the people, but at the same time, it also sprouts many hazards. Portrait detection technology began in the 60-70 era of the last century. After decades of research and application, it has been gradually improved. However, with the development of the times, the requirement of recognition technology is getting higher and higher. Detecting sensitive photos is an important recognition work. Therefore, people pay more attention to this technology and theory, and gradually become a research hotspot. In this paper, we propose some new methods to detect regions of interest in bad photos, improve the weight method of training weak classifiers, distinguish bad photos efficiently, and urge the green development of the Internet.The problem of this paper is about the detection of Adaboost algorithm in bad photos, but there are many objec...
