This dataset contains natural scene images extracted from 724 movie clips hosted on Storyblocks (formerly Videoblocks).
We downloaded a total of 724 movie clips and converted the movie frames to image files (in JPEG format) using ffmpeg at 1920x1080 pixels. Then we selected the images at 1Hz (i.e. each image is one second apart in the original movie clips), before picking images that were representative of each movie clip.
These images are the stimuli used by preprint: https://psyarxiv.com/d2s38/