Vision Transformers are Inherently Saliency Learners.

Yasser Abdelaziz Dahou Djilali, Kevin McGuinness, Noel E. O'Connor

Browse the full BMVC paper archive.