By zooming in on high-resolution face photographs, we were able to recover images of unseen bystanders from reflections in the subjects‘ eyes.