Toward quantifying ambiguities in artistic images

Xi Wang; Zoya Bylinskii; Aaron Hertzmann; Robert Pepperell

doi:10.1145/3418054

Toward quantifying ambiguities in artistic images

Xi Wang, Zoya Bylinskii, Aaron Hertzmann, Robert Pepperell

Research output: Contribution to journal › Article › peer-review

7 Citations (Scopus)

Abstract

It has long been hypothesized that perceptual ambiguities play an important role in aesthetic experience: A work with some ambiguity engages a viewer more than one that does not. However, current frameworks for testing this theory are limited by the availability of stimuli and data collection methods. This article presents an approach to measuring the perceptual ambiguity of a collection of images. Crowdworkers are asked to describe image content, after different viewing durations. Experiments are performed using images created with Generative Adversarial Networks, using the Artbreeder website. We show that text processing of viewer responses can provide a fine-grained way to measure and describe image ambiguities.

Original language	English
Article number	13
Journal	ACM Transactions on Applied Perception
Volume	17
Issue number	4
DOIs	https://doi.org/10.1145/3418054
Publication status	Published - 6 Nov 2020

Keywords

Aesthetics
Datasets
Generative adversarial networks (GAN)
Image descriptions
Text tagging

Access to Document

10.1145/3418054

Cite this

@article{735a256e1c9041a2a2f724bbf24fa24c,

title = "Toward quantifying ambiguities in artistic images",

abstract = "It has long been hypothesized that perceptual ambiguities play an important role in aesthetic experience: A work with some ambiguity engages a viewer more than one that does not. However, current frameworks for testing this theory are limited by the availability of stimuli and data collection methods. This article presents an approach to measuring the perceptual ambiguity of a collection of images. Crowdworkers are asked to describe image content, after different viewing durations. Experiments are performed using images created with Generative Adversarial Networks, using the Artbreeder website. We show that text processing of viewer responses can provide a fine-grained way to measure and describe image ambiguities.",

keywords = "Aesthetics, Datasets, Generative adversarial networks (GAN), Image descriptions, Text tagging",

author = "Xi Wang and Zoya Bylinskii and Aaron Hertzmann and Robert Pepperell",

note = "Publisher Copyright: {\textcopyright} 2020 Copyright held by the owner/author(s). Publication rights licensed to ACM.",

year = "2020",

month = nov,

day = "6",

doi = "10.1145/3418054",

language = "English",

volume = "17",

journal = "ACM Transactions on Applied Perception",

issn = "1544-3558",

number = "4",

}

TY - JOUR

T1 - Toward quantifying ambiguities in artistic images

AU - Wang, Xi

AU - Bylinskii, Zoya

AU - Hertzmann, Aaron

AU - Pepperell, Robert

PY - 2020/11/6

Y1 - 2020/11/6

N2 - It has long been hypothesized that perceptual ambiguities play an important role in aesthetic experience: A work with some ambiguity engages a viewer more than one that does not. However, current frameworks for testing this theory are limited by the availability of stimuli and data collection methods. This article presents an approach to measuring the perceptual ambiguity of a collection of images. Crowdworkers are asked to describe image content, after different viewing durations. Experiments are performed using images created with Generative Adversarial Networks, using the Artbreeder website. We show that text processing of viewer responses can provide a fine-grained way to measure and describe image ambiguities.

AB - It has long been hypothesized that perceptual ambiguities play an important role in aesthetic experience: A work with some ambiguity engages a viewer more than one that does not. However, current frameworks for testing this theory are limited by the availability of stimuli and data collection methods. This article presents an approach to measuring the perceptual ambiguity of a collection of images. Crowdworkers are asked to describe image content, after different viewing durations. Experiments are performed using images created with Generative Adversarial Networks, using the Artbreeder website. We show that text processing of viewer responses can provide a fine-grained way to measure and describe image ambiguities.

KW - Aesthetics

KW - Datasets

KW - Generative adversarial networks (GAN)

KW - Image descriptions

KW - Text tagging

UR - http://www.scopus.com/inward/record.url?scp=85098327228&partnerID=8YFLogxK

U2 - 10.1145/3418054

DO - 10.1145/3418054

M3 - Article

AN - SCOPUS:85098327228

SN - 1544-3558

VL - 17

JO - ACM Transactions on Applied Perception

JF - ACM Transactions on Applied Perception

IS - 4

M1 - 13

ER -

Toward quantifying ambiguities in artistic images

Abstract

Keywords

Access to Document

Other files and links

Cite this