PAC-Net: Pairwise aesthetic comparison network for image aesthetic assessment

Keunsoo Ko, Jun Tae Lee, Chang-Su Kim

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Image aesthetic assessment is important for finding well taken and appealing photographs but is challenging due to the ambiguity and subjectivity of aesthetic criteria. We develop the pairwise aesthetic comparison network (PAC-Net), which consists of two parts: aesthetic feature extraction and pairwise feature comparison. To alleviate the ambiguity and subjectivity, we train PAC-Net to learn the relative aesthetic ranks of two images by employing a novel loss function, called aesthetic-adaptive cross entropy loss. Then, we develop simple schemes for using PAC-Net in the tasks of aesthetic ranking and aesthetic classification, respectively. Experimental results demonstrate that PAC-Net achieves the state-of-the-art performances in both the ranking and classification applications.

Original languageEnglish
Title of host publication2018 IEEE International Conference on Image Processing, ICIP 2018 - Proceedings
PublisherIEEE Computer Society
Pages2491-2495
Number of pages5
ISBN (Electronic)9781479970612
DOIs
Publication statusPublished - 2018 Aug 29
Event25th IEEE International Conference on Image Processing, ICIP 2018 - Athens, Greece
Duration: 2018 Oct 72018 Oct 10

Publication series

NameProceedings - International Conference on Image Processing, ICIP
ISSN (Print)1522-4880

Conference

Conference25th IEEE International Conference on Image Processing, ICIP 2018
CountryGreece
CityAthens
Period18/10/718/10/10

Fingerprint

Feature extraction
Entropy

Keywords

  • Aesthetic ranking
  • And pairwise comparison
  • Convolutional neural networks
  • Image aesthetic assessment

ASJC Scopus subject areas

  • Software
  • Computer Vision and Pattern Recognition
  • Signal Processing

Cite this

Ko, K., Lee, J. T., & Kim, C-S. (2018). PAC-Net: Pairwise aesthetic comparison network for image aesthetic assessment. In 2018 IEEE International Conference on Image Processing, ICIP 2018 - Proceedings (pp. 2491-2495). [8451621] (Proceedings - International Conference on Image Processing, ICIP). IEEE Computer Society. https://doi.org/10.1109/ICIP.2018.8451621

PAC-Net : Pairwise aesthetic comparison network for image aesthetic assessment. / Ko, Keunsoo; Lee, Jun Tae; Kim, Chang-Su.

2018 IEEE International Conference on Image Processing, ICIP 2018 - Proceedings. IEEE Computer Society, 2018. p. 2491-2495 8451621 (Proceedings - International Conference on Image Processing, ICIP).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ko, K, Lee, JT & Kim, C-S 2018, PAC-Net: Pairwise aesthetic comparison network for image aesthetic assessment. in 2018 IEEE International Conference on Image Processing, ICIP 2018 - Proceedings., 8451621, Proceedings - International Conference on Image Processing, ICIP, IEEE Computer Society, pp. 2491-2495, 25th IEEE International Conference on Image Processing, ICIP 2018, Athens, Greece, 18/10/7. https://doi.org/10.1109/ICIP.2018.8451621
Ko K, Lee JT, Kim C-S. PAC-Net: Pairwise aesthetic comparison network for image aesthetic assessment. In 2018 IEEE International Conference on Image Processing, ICIP 2018 - Proceedings. IEEE Computer Society. 2018. p. 2491-2495. 8451621. (Proceedings - International Conference on Image Processing, ICIP). https://doi.org/10.1109/ICIP.2018.8451621
Ko, Keunsoo ; Lee, Jun Tae ; Kim, Chang-Su. / PAC-Net : Pairwise aesthetic comparison network for image aesthetic assessment. 2018 IEEE International Conference on Image Processing, ICIP 2018 - Proceedings. IEEE Computer Society, 2018. pp. 2491-2495 (Proceedings - International Conference on Image Processing, ICIP).
@inproceedings{c606bc40121c41b5a04b19188540050c,
title = "PAC-Net: Pairwise aesthetic comparison network for image aesthetic assessment",
abstract = "Image aesthetic assessment is important for finding well taken and appealing photographs but is challenging due to the ambiguity and subjectivity of aesthetic criteria. We develop the pairwise aesthetic comparison network (PAC-Net), which consists of two parts: aesthetic feature extraction and pairwise feature comparison. To alleviate the ambiguity and subjectivity, we train PAC-Net to learn the relative aesthetic ranks of two images by employing a novel loss function, called aesthetic-adaptive cross entropy loss. Then, we develop simple schemes for using PAC-Net in the tasks of aesthetic ranking and aesthetic classification, respectively. Experimental results demonstrate that PAC-Net achieves the state-of-the-art performances in both the ranking and classification applications.",
keywords = "Aesthetic ranking, And pairwise comparison, Convolutional neural networks, Image aesthetic assessment",
author = "Keunsoo Ko and Lee, {Jun Tae} and Chang-Su Kim",
year = "2018",
month = "8",
day = "29",
doi = "10.1109/ICIP.2018.8451621",
language = "English",
series = "Proceedings - International Conference on Image Processing, ICIP",
publisher = "IEEE Computer Society",
pages = "2491--2495",
booktitle = "2018 IEEE International Conference on Image Processing, ICIP 2018 - Proceedings",

}

TY - GEN

T1 - PAC-Net

T2 - Pairwise aesthetic comparison network for image aesthetic assessment

AU - Ko, Keunsoo

AU - Lee, Jun Tae

AU - Kim, Chang-Su

PY - 2018/8/29

Y1 - 2018/8/29

N2 - Image aesthetic assessment is important for finding well taken and appealing photographs but is challenging due to the ambiguity and subjectivity of aesthetic criteria. We develop the pairwise aesthetic comparison network (PAC-Net), which consists of two parts: aesthetic feature extraction and pairwise feature comparison. To alleviate the ambiguity and subjectivity, we train PAC-Net to learn the relative aesthetic ranks of two images by employing a novel loss function, called aesthetic-adaptive cross entropy loss. Then, we develop simple schemes for using PAC-Net in the tasks of aesthetic ranking and aesthetic classification, respectively. Experimental results demonstrate that PAC-Net achieves the state-of-the-art performances in both the ranking and classification applications.

AB - Image aesthetic assessment is important for finding well taken and appealing photographs but is challenging due to the ambiguity and subjectivity of aesthetic criteria. We develop the pairwise aesthetic comparison network (PAC-Net), which consists of two parts: aesthetic feature extraction and pairwise feature comparison. To alleviate the ambiguity and subjectivity, we train PAC-Net to learn the relative aesthetic ranks of two images by employing a novel loss function, called aesthetic-adaptive cross entropy loss. Then, we develop simple schemes for using PAC-Net in the tasks of aesthetic ranking and aesthetic classification, respectively. Experimental results demonstrate that PAC-Net achieves the state-of-the-art performances in both the ranking and classification applications.

KW - Aesthetic ranking

KW - And pairwise comparison

KW - Convolutional neural networks

KW - Image aesthetic assessment

UR - http://www.scopus.com/inward/record.url?scp=85062908586&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062908586&partnerID=8YFLogxK

U2 - 10.1109/ICIP.2018.8451621

DO - 10.1109/ICIP.2018.8451621

M3 - Conference contribution

AN - SCOPUS:85062908586

T3 - Proceedings - International Conference on Image Processing, ICIP

SP - 2491

EP - 2495

BT - 2018 IEEE International Conference on Image Processing, ICIP 2018 - Proceedings

PB - IEEE Computer Society

ER -