METHOD AND SYSTEM FOR PUSHING INDIVIDUAL ADVERTISEMENT BASED ON USER INTEREST LEARNING
摘要
A method and a system for pushing personalized advertisement based on user interest learning are disclosed. The said method comprises: obtaining plural user interest models with multi-task sorting learning (201); according to the said user interest models, extracting interested objects from videos (202); extracting various visual features of the said interested objects, and searching correlative advertisement information in an advertisement database according to the said visual features (203). With the method and the system, the pushed advertisement and content of the video are tight correlative, which satisfies personalized requirement of the user in a certain extent and realizes pushing personalized advertisement.