Agent based video contents identification and data mining using watermark based filtering

Abstract

This chapter describes the agent based video contents identification scheme using watermark based filtering technique. To prevent a user from uploading illegal video contents into the WEB storages, two strategies are employed. First stage is the upload blocking of illegal contents including copyright ownership information as a watermark when a user tries to upload the illegal video content. Second stage is to monitor illegal video contents that are already uploaded. For this stage, the monitoring agent obtains video content link information, and then extracts the watermark from corresponding content using the Open API. For two stage video identification strategies, two types of watermark extraction schemes are employed. Gathered data obtained from agents is analyzed using data mining method, and reporting process is done. To show the effectiveness of the described system, some experimental evaluation and test are conducted. © 2009 Springer-Verlag US.