流媒体视频文件相似性识别的方法

A method to identify the similarity of the contents of streaming video files

在现有CDN中,很多的空间用于存储基于影视作品的流媒体文体,而这些文件很多内容一样,但编码方式、码率等不一样,重复存储浪费了大量存储空间,增加了运营成本。针对影视内容的流媒体文体,提出一种采集相邻Ⅰ帧时间信息进行内容相似度识别的方法,可以有效解决基于不同编码方式、码率,但内容一致的流媒体文件的识别问题。

In CDN,a lot of space is wasted for storage of those stream media files who have same content,but employ different CODEC or bit rate etc.This article presents a new method which identifies the similarity of the contents between a group of streaming video files by the time information of adjacent Ⅰ frames,and it works for those stream media files just mentioned.