Abstract
We describe an interactive web application that lets visitors navigate an online collection of visual art by the emotional content of the work. Visitors can find artwork that meets a chosen emotional profile, explore artwork that is spanned by the emotions contained in two images, or view the web of emotional connections linking images in a larger database. Visitors can also upload an image of their choice and explore its emotional connections to items in the database. This system provides a new modality for exploring online visual art collections relative to catalogues organized by period, genre, or artist. Its interactive format encourages playful interaction that engages online users, while its emotional organization creates the potential for surprise. This work is in prototype form; it operates on a database of 4000 images taken from a wiki-art dataset, but it is applicable to much larger collections through use of a machine learning pipeline (developed for this work) that can estimate the affective content of a wide variety of 2-dimensional visual art. A demo of the video can be seen here: https://youtu.be/ZaP7T_RLd34
Presenters
Sarah FrostResearcher, Computational Media, University of California, Santa Cruz, California, United States
Details
Presentation Type
Theme
KEYWORDS
Internet, Virtual Museums, Digitization, Emotion