Salient-Imagenet

The dataset used in the paper is a large-scale dataset Salient-Imagenet, which is based on Imagenet with pixel-level annotations of irrelevant (spurious) and relevant (core) regions.

BibTex: