Visual Metaphor Dataset

Overview 125 images of visual metaphors, individually collected from the web by the author, referred to by their source URLs. Manual annotations by a single annotator (the author), covering: the type of source the image comes from (origin_type) the two concepts/domains being visually mapped onto each other (img_domain_1, img_domain_2) a short textual metaphor inspired by the image (text_metaphor) the two concepts/domains that textual metaphor evokes (text_domain_1, text_domain_2) 250 AI-generated visual metaphors, corresponding with the 125 manually annotated textual metaphors, and 125 additional textual metaphors taken from Chakrabarty (2023) (https://zenodo.org/records/8011133)