We provide eye fixation data for a total of 2,787 images spanning 21 indoor and outdoor scene categories. These images are split into two sets: 630 target images and over 2K filler images. For the 630 target images, we provide eye fixation data for an average of 16 observers per image, memorability scores, pre-computed fixation maps for training and testing saliency models, and complete annotations for all the objects in each image. For the 2K filler images, we provide eye fixation data for an average of 15 observers per image and pre-computed fixation maps for training and testing saliency models. These images are a subset of the FIGRIM fine-grained image memorability dataset.
@article{figrim, title={Intrinsic and Extrinsic Effects on Image Memorability}, author={Bylinskii, Zoya and Isola, Phillip and Bainbridge, Constance and Torralba, Antonio and Oliva, Aude}, journal={Vision research}, volume={116}, pages={165--178}, year={2015}, publisher={Elsevier} }