Personalized video composition method based on user intention learning

摘要

In this paper we present a personalized video composition method based on user intention learning. The proposed system generates video clips by choosing user's assets such as still images and video clips based on pre-designed templates. The search-and-suggestion technique plays an important role for providing appropriate templates. We mainly focus on the learning mechanism which will be gradually improving through repeated template recommendations. Our prototype demonstrates the feasibility of the proposed approach.