Work on exploring where humans and models look while predicting video memorability accepted at WACV2025 ! Tweet l Arxiv