OTT Ad Insertion Using Memorability Enforcement

With the proliferation of Over-the-Top (OTT) content consumption, programmatic advertising has become a key service, enhancing publishers’ revenue via advertisements (ads). Traditional ad insertion strategies have emphasized matching content with ads based on contextual relevance and presentation sequence to boost effectiveness. Although there is a great deal of progress in the aforementioned ad placement techniques, ad and content memorability coefficients are introduced as a factor in this paper. The proposed approach incorporates memorability into the ad insertion process by matching ads with placement regions within content by utilizing memorability coefficient scores. This method involves an optimization model, further extended by a greedy heuristic, to select an optimal set of ads for specific placement areas while maximizing memorability correlation. The memorability scores are calculated using a Machine Learning (ML) algorithm trained on the VideoMem dataset, adding a novel facet to improving viewer experience through targeted ad delivery. The proposed scheme is evaluated to show the ability of the formulation and the heuristic algorithm in aligning ads within content regions based on the memorability factor. Additionally, the impact on Quality-of-Experience (QoE) is also assessed following the evaluation’s determination of memorability coefficient scores.