Relative learning from web images for content-adaptive enhancement

Parag Shridhar Chandakkar, Qiongjie Tian, Baoxin Li

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

2 Citations (Scopus)

Abstract

Personalized and content-adaptive image enhancement can find many applications in the age of social media and mobile computing. This paper presents a relative-learning-based approach, which, unlike previous methods, does not require matching original and enhanced images for training. This allows the use of massive online photo collections to train a ranking model for improved enhancement. We first propose a multi-level ranking model, which is learned from only relatively-labeled inputs that are automatically crawled. Then we design a novel parameter sampling scheme under this model to generate the desired enhancement parameters for a new image. For evaluation, we first verify the effectiveness and the generalization abilities of our approach, using images that have been enhanced/labeled by experts. Then we carry out subjective tests, which show that users prefer images enhanced by our approach over other existing methods.

Original languageEnglish (US)
Title of host publicationProceedings - IEEE International Conference on Multimedia and Expo
PublisherIEEE Computer Society
Volume2015-August
ISBN (Print)9781479970827
DOIs
StatePublished - Aug 4 2015
EventIEEE International Conference on Multimedia and Expo, ICME 2015 - Turin, Italy
Duration: Jun 29 2015Jul 3 2015

Other

OtherIEEE International Conference on Multimedia and Expo, ICME 2015
CountryItaly
CityTurin
Period6/29/157/3/15

Fingerprint

Mobile computing
Image enhancement
Sampling

Keywords

  • Content-adaptive image enhancement
  • learning-to-rank
  • subjective evaluation testing

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Computer Science Applications

Cite this

Chandakkar, P. S., Tian, Q., & Li, B. (2015). Relative learning from web images for content-adaptive enhancement. In Proceedings - IEEE International Conference on Multimedia and Expo (Vol. 2015-August). [7177502] IEEE Computer Society. https://doi.org/10.1109/ICME.2015.7177502

Relative learning from web images for content-adaptive enhancement. / Chandakkar, Parag Shridhar; Tian, Qiongjie; Li, Baoxin.

Proceedings - IEEE International Conference on Multimedia and Expo. Vol. 2015-August IEEE Computer Society, 2015. 7177502.

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

Chandakkar, PS, Tian, Q & Li, B 2015, Relative learning from web images for content-adaptive enhancement. in Proceedings - IEEE International Conference on Multimedia and Expo. vol. 2015-August, 7177502, IEEE Computer Society, IEEE International Conference on Multimedia and Expo, ICME 2015, Turin, Italy, 6/29/15. https://doi.org/10.1109/ICME.2015.7177502
Chandakkar PS, Tian Q, Li B. Relative learning from web images for content-adaptive enhancement. In Proceedings - IEEE International Conference on Multimedia and Expo. Vol. 2015-August. IEEE Computer Society. 2015. 7177502 https://doi.org/10.1109/ICME.2015.7177502
Chandakkar, Parag Shridhar ; Tian, Qiongjie ; Li, Baoxin. / Relative learning from web images for content-adaptive enhancement. Proceedings - IEEE International Conference on Multimedia and Expo. Vol. 2015-August IEEE Computer Society, 2015.
@inproceedings{57e9f9a9f29e4158af01fef63466742c,
title = "Relative learning from web images for content-adaptive enhancement",
abstract = "Personalized and content-adaptive image enhancement can find many applications in the age of social media and mobile computing. This paper presents a relative-learning-based approach, which, unlike previous methods, does not require matching original and enhanced images for training. This allows the use of massive online photo collections to train a ranking model for improved enhancement. We first propose a multi-level ranking model, which is learned from only relatively-labeled inputs that are automatically crawled. Then we design a novel parameter sampling scheme under this model to generate the desired enhancement parameters for a new image. For evaluation, we first verify the effectiveness and the generalization abilities of our approach, using images that have been enhanced/labeled by experts. Then we carry out subjective tests, which show that users prefer images enhanced by our approach over other existing methods.",
keywords = "Content-adaptive image enhancement, learning-to-rank, subjective evaluation testing",
author = "Chandakkar, {Parag Shridhar} and Qiongjie Tian and Baoxin Li",
year = "2015",
month = "8",
day = "4",
doi = "10.1109/ICME.2015.7177502",
language = "English (US)",
isbn = "9781479970827",
volume = "2015-August",
booktitle = "Proceedings - IEEE International Conference on Multimedia and Expo",
publisher = "IEEE Computer Society",

}

TY - GEN

T1 - Relative learning from web images for content-adaptive enhancement

AU - Chandakkar, Parag Shridhar

AU - Tian, Qiongjie

AU - Li, Baoxin

PY - 2015/8/4

Y1 - 2015/8/4

N2 - Personalized and content-adaptive image enhancement can find many applications in the age of social media and mobile computing. This paper presents a relative-learning-based approach, which, unlike previous methods, does not require matching original and enhanced images for training. This allows the use of massive online photo collections to train a ranking model for improved enhancement. We first propose a multi-level ranking model, which is learned from only relatively-labeled inputs that are automatically crawled. Then we design a novel parameter sampling scheme under this model to generate the desired enhancement parameters for a new image. For evaluation, we first verify the effectiveness and the generalization abilities of our approach, using images that have been enhanced/labeled by experts. Then we carry out subjective tests, which show that users prefer images enhanced by our approach over other existing methods.

AB - Personalized and content-adaptive image enhancement can find many applications in the age of social media and mobile computing. This paper presents a relative-learning-based approach, which, unlike previous methods, does not require matching original and enhanced images for training. This allows the use of massive online photo collections to train a ranking model for improved enhancement. We first propose a multi-level ranking model, which is learned from only relatively-labeled inputs that are automatically crawled. Then we design a novel parameter sampling scheme under this model to generate the desired enhancement parameters for a new image. For evaluation, we first verify the effectiveness and the generalization abilities of our approach, using images that have been enhanced/labeled by experts. Then we carry out subjective tests, which show that users prefer images enhanced by our approach over other existing methods.

KW - Content-adaptive image enhancement

KW - learning-to-rank

KW - subjective evaluation testing

UR - http://www.scopus.com/inward/record.url?scp=84946090926&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84946090926&partnerID=8YFLogxK

U2 - 10.1109/ICME.2015.7177502

DO - 10.1109/ICME.2015.7177502

M3 - Conference contribution

SN - 9781479970827

VL - 2015-August

BT - Proceedings - IEEE International Conference on Multimedia and Expo

PB - IEEE Computer Society

ER -