不良视频检测系统的设计与实现
首发时间:2009-03-25
摘要:针对视频内容的特征,提出了一种适用于流媒体的不良视频检测系统。该系统利用改进的基于聚类的镜头分割算法对视频进行分割,然后进行关键帧提取,将对视频的检测转化为对关键帧的识别;在图象识别过程中,把颜色特征和人体姿态特征相结合,对关键部位进行特征识别;提出了一种有效的判决准则,实现整段视频的判别。实验表明,该算法识别率达到80%以上,鲁棒性较好,适用于视频检测环境,为检测不良视频提供了一个有效的方法。
For information in English, please click here
Design and Implementation of Objectionable Videos Detection System
Abstract:Based on the features of video content, this paper addresses the system of detecting objectionable videos. An improved algorithm for shot segmentation based on clustering was presented before extraction key frames. It transforms the detection of video to identification of key frames; we combined the color features and characteristics of body posture to identify the key parts of human body; an effective determining rule is given. The experimental results show that the method’s correct detection is up to 80% and it is fit for the video detection environment, our method provides an effective method for detecting objectionable videos.
Keywords: shot segmentation key-frame clustering skin color detection histogram
论文图表:
引用
No.3075345266112379****
同行评议
共计0人参与
勘误表
不良视频检测系统的设计与实现
评论
全部评论0/1000