A real-time adaptive algorithm for detecting advertisement logo in videos

With the growing number of media advertising elements, the detection of time and length of advertisement logo has become a new demand of video monitoring sections. In response to this demand, this paper put forward an advertisement logo detecting approach in videos combing self-adaptive method and global feature. The proposed algorithm establishes the matching-template that represents shape feature of image online by introducing a self-adaptive method, so that background interference in the video frames can be eliminated. The experimental results prove that our approach can balance retrieval performance with efficiency, which meets the requirements of advertisement logo detection in videos.