VISTA: A Visual and Textual Attention Dataset for Interpreting Multimodal Models.

Harshit, Tolga Tasdizen

Browse the full WACV paper archive.