We introduce a framework that uses Generative Adversarial Networks (GANs) to
study cognitive properties like memorability, aesthetics, and emotional
valence. These attributes are of interest because we do not have a concrete
visual definition of what they entail. What does it look like for a dog to be
more or less memorable? GANs allow us to generate a manifold of natural-looking
images with fine-grained differences in their visual attributes. By navigating
this manifold in directions that increase memorability, we can visualize what
it looks like for a particular generated image to become more or less
memorable. The resulting ``visual definitions" surface image properties (like
``object size") that may underlie memorability. Through behavioral experiments,
we verify that our method indeed discovers image manipulations that causally
affect human memory performance. We further demonstrate that the same framework
can be used to analyze image aesthetics and emotional valence. Visit the
GANalyze website at http://ganalyze.csail.mit.edu/.
Description
[1906.10112] GANalyze: Toward Visual Definitions of Cognitive Image Properties
%0 Journal Article
%1 goetschalckx2019ganalyze
%A Goetschalckx, Lore
%A Andonian, Alex
%A Oliva, Aude
%A Isola, Phillip
%D 2019
%K game-theory generative-models
%T GANalyze: Toward Visual Definitions of Cognitive Image Properties
%U http://arxiv.org/abs/1906.10112
%X We introduce a framework that uses Generative Adversarial Networks (GANs) to
study cognitive properties like memorability, aesthetics, and emotional
valence. These attributes are of interest because we do not have a concrete
visual definition of what they entail. What does it look like for a dog to be
more or less memorable? GANs allow us to generate a manifold of natural-looking
images with fine-grained differences in their visual attributes. By navigating
this manifold in directions that increase memorability, we can visualize what
it looks like for a particular generated image to become more or less
memorable. The resulting ``visual definitions" surface image properties (like
``object size") that may underlie memorability. Through behavioral experiments,
we verify that our method indeed discovers image manipulations that causally
affect human memory performance. We further demonstrate that the same framework
can be used to analyze image aesthetics and emotional valence. Visit the
GANalyze website at http://ganalyze.csail.mit.edu/.
@article{goetschalckx2019ganalyze,
abstract = {We introduce a framework that uses Generative Adversarial Networks (GANs) to
study cognitive properties like memorability, aesthetics, and emotional
valence. These attributes are of interest because we do not have a concrete
visual definition of what they entail. What does it look like for a dog to be
more or less memorable? GANs allow us to generate a manifold of natural-looking
images with fine-grained differences in their visual attributes. By navigating
this manifold in directions that increase memorability, we can visualize what
it looks like for a particular generated image to become more or less
memorable. The resulting ``visual definitions" surface image properties (like
``object size") that may underlie memorability. Through behavioral experiments,
we verify that our method indeed discovers image manipulations that causally
affect human memory performance. We further demonstrate that the same framework
can be used to analyze image aesthetics and emotional valence. Visit the
GANalyze website at http://ganalyze.csail.mit.edu/.},
added-at = {2020-01-06T03:20:29.000+0100},
author = {Goetschalckx, Lore and Andonian, Alex and Oliva, Aude and Isola, Phillip},
biburl = {https://www.bibsonomy.org/bibtex/2b9182725553a8fe5a82cbbf874da4b86/kirk86},
description = {[1906.10112] GANalyze: Toward Visual Definitions of Cognitive Image Properties},
interhash = {6cb7803b73621b01205545883b08284c},
intrahash = {b9182725553a8fe5a82cbbf874da4b86},
keywords = {game-theory generative-models},
note = {cite arxiv:1906.10112Comment: 17 pages, 15 figures},
timestamp = {2020-01-06T03:20:29.000+0100},
title = {GANalyze: Toward Visual Definitions of Cognitive Image Properties},
url = {http://arxiv.org/abs/1906.10112},
year = 2019
}