Optimizing Advertisement Placement Using Saliency Estimation in Filmmaking

As audiences spend an increasing amount of time consuming video content through online platforms, effective advertisement placement has become crucial for both engagement and revenue generation. While recommendation systems help match relevant advertisements to users, the optimal placement of film-related advertisements within visual content, remains underexplored. Traditionally, eye-tracking has been considered the most reliable method for gauging user attention, but it is both expensive and impractical in large-scale applications. This paper proposes a saliency-driven optimization approach for advertisement placement, leveraging saliency map estimations to predict areas of high visual attention. An automated solution that identifies optimal advertisement locations by focusing on the visual center of saliency maps is presented, eliminating the need for intrusive methods like eye-tracking. Additionally, a comparative analysis of two popular saliency estimation techniques is conducted, with both methods fine-tuned to enhance advertisement placement specifically in the context of film production. This approach aims to integrate advertisements seamlessly into visual media without disrupting viewer immersion. This approach will be used to insert film-related adverts in an audience engagement application.