Show and Recall: Learning What Makes Videos Memorable

TL;DR

This paper introduces a new model for predicting video memorability, considering complex video features, and demonstrates its effectiveness through experiments that outperform existing methods in video summarization tasks.

Contribution

It presents the first comprehensive prediction model for video memorability that accounts for content complexities and improves over current approaches.

Findings

The proposed model correlates well with existing memorability research.

It outperforms current methods in predicting sub-shot memorability.

Achieves competitive results in video summarization benchmarks.

Abstract

With the explosion of video content on the Internet, there is a need for research on methods for video analysis which take human cognition into account. One such cognitive measure is memorability, or the ability to recall visual content after watching it. Prior research has looked into image memorability and shown that it is intrinsic to visual content, but the problem of modeling video memorability has not been addressed sufficiently. In this work, we develop a prediction model for video memorability, including complexities of video content in it. Detailed feature analysis reveals that the proposed method correlates well with existing findings on memorability. We also describe a novel experiment of predicting video sub-shot memorability and show that our approach improves over current memorability methods in this task. Experiments on standard datasets demonstrate that the proposed metric can achieve results on par or better than the state-of-the art methods for video summarization.