VRank: Voting system on Ranking model for human age estimation

Tekoing Lim, Kai Lung Hua, Hong Cyuan Wang, Kai Wen Zhao, Min Chun Hu, Wen Huang Cheng

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

2 Citations (Scopus)

Abstract

Ranking algorithms have proven the potential for human age estimation. Currently, a common paradigm is to compare the input face with reference faces of known age to generate a ranking relation whereby the first-rank reference is exploited for labeling the input face. In this paper, we proposed a framework to improve upon the typical ranking model, called Voting system on Ranking model (VRank), by leveraging relational information (comparative relations, i.e. if the input face is younger or older than each of the references) to make a more robust estimation. Our approach has several advantages: firstly, comparative relations can be explicitly involved to benefit the estimation task; secondly, few incorrect comparisons will not influence much the accuracy of the result, making this approach more robust than the conventional approach; finally, we propose to incorporate the deep learning architecture for training, which extracts robust facial features for increasing the effectiveness of classification. In comparison to the best results from the state-of-the-art methods, the VRank showed a significant outperformance on all the benchmarks, with a relative improvement of 5.74% ∼ 69.45% (FG-NET), 19.09% ∼ 68.71% (MORPH), and 0.55% ∼ 17.73% (IoG).

Original languageEnglish
Title of host publication2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781467374781
DOIs
Publication statusPublished - 2015 Nov 30
Event17th IEEE International Workshop on Multimedia Signal Processing, MMSP 2015 - Xiamen, China
Duration: 2015 Oct 192015 Oct 21

Publication series

Name2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015

Other

Other17th IEEE International Workshop on Multimedia Signal Processing, MMSP 2015
CountryChina
CityXiamen
Period15-10-1915-10-21

Fingerprint

Labeling
Deep learning

All Science Journal Classification (ASJC) codes

  • Signal Processing
  • Media Technology

Cite this

Lim, T., Hua, K. L., Wang, H. C., Zhao, K. W., Hu, M. C., & Cheng, W. H. (2015). VRank: Voting system on Ranking model for human age estimation. In 2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015 [7340789] (2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/MMSP.2015.7340789
Lim, Tekoing ; Hua, Kai Lung ; Wang, Hong Cyuan ; Zhao, Kai Wen ; Hu, Min Chun ; Cheng, Wen Huang. / VRank : Voting system on Ranking model for human age estimation. 2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015. Institute of Electrical and Electronics Engineers Inc., 2015. (2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015).
@inproceedings{31ddf94059cc4cabb27549d2330682d5,
title = "VRank: Voting system on Ranking model for human age estimation",
abstract = "Ranking algorithms have proven the potential for human age estimation. Currently, a common paradigm is to compare the input face with reference faces of known age to generate a ranking relation whereby the first-rank reference is exploited for labeling the input face. In this paper, we proposed a framework to improve upon the typical ranking model, called Voting system on Ranking model (VRank), by leveraging relational information (comparative relations, i.e. if the input face is younger or older than each of the references) to make a more robust estimation. Our approach has several advantages: firstly, comparative relations can be explicitly involved to benefit the estimation task; secondly, few incorrect comparisons will not influence much the accuracy of the result, making this approach more robust than the conventional approach; finally, we propose to incorporate the deep learning architecture for training, which extracts robust facial features for increasing the effectiveness of classification. In comparison to the best results from the state-of-the-art methods, the VRank showed a significant outperformance on all the benchmarks, with a relative improvement of 5.74{\%} ∼ 69.45{\%} (FG-NET), 19.09{\%} ∼ 68.71{\%} (MORPH), and 0.55{\%} ∼ 17.73{\%} (IoG).",
author = "Tekoing Lim and Hua, {Kai Lung} and Wang, {Hong Cyuan} and Zhao, {Kai Wen} and Hu, {Min Chun} and Cheng, {Wen Huang}",
year = "2015",
month = "11",
day = "30",
doi = "10.1109/MMSP.2015.7340789",
language = "English",
series = "2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
booktitle = "2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015",
address = "United States",

}

Lim, T, Hua, KL, Wang, HC, Zhao, KW, Hu, MC & Cheng, WH 2015, VRank: Voting system on Ranking model for human age estimation. in 2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015., 7340789, 2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015, Institute of Electrical and Electronics Engineers Inc., 17th IEEE International Workshop on Multimedia Signal Processing, MMSP 2015, Xiamen, China, 15-10-19. https://doi.org/10.1109/MMSP.2015.7340789

VRank : Voting system on Ranking model for human age estimation. / Lim, Tekoing; Hua, Kai Lung; Wang, Hong Cyuan; Zhao, Kai Wen; Hu, Min Chun; Cheng, Wen Huang.

2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015. Institute of Electrical and Electronics Engineers Inc., 2015. 7340789 (2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015).

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

TY - GEN

T1 - VRank

T2 - Voting system on Ranking model for human age estimation

AU - Lim, Tekoing

AU - Hua, Kai Lung

AU - Wang, Hong Cyuan

AU - Zhao, Kai Wen

AU - Hu, Min Chun

AU - Cheng, Wen Huang

PY - 2015/11/30

Y1 - 2015/11/30

N2 - Ranking algorithms have proven the potential for human age estimation. Currently, a common paradigm is to compare the input face with reference faces of known age to generate a ranking relation whereby the first-rank reference is exploited for labeling the input face. In this paper, we proposed a framework to improve upon the typical ranking model, called Voting system on Ranking model (VRank), by leveraging relational information (comparative relations, i.e. if the input face is younger or older than each of the references) to make a more robust estimation. Our approach has several advantages: firstly, comparative relations can be explicitly involved to benefit the estimation task; secondly, few incorrect comparisons will not influence much the accuracy of the result, making this approach more robust than the conventional approach; finally, we propose to incorporate the deep learning architecture for training, which extracts robust facial features for increasing the effectiveness of classification. In comparison to the best results from the state-of-the-art methods, the VRank showed a significant outperformance on all the benchmarks, with a relative improvement of 5.74% ∼ 69.45% (FG-NET), 19.09% ∼ 68.71% (MORPH), and 0.55% ∼ 17.73% (IoG).

AB - Ranking algorithms have proven the potential for human age estimation. Currently, a common paradigm is to compare the input face with reference faces of known age to generate a ranking relation whereby the first-rank reference is exploited for labeling the input face. In this paper, we proposed a framework to improve upon the typical ranking model, called Voting system on Ranking model (VRank), by leveraging relational information (comparative relations, i.e. if the input face is younger or older than each of the references) to make a more robust estimation. Our approach has several advantages: firstly, comparative relations can be explicitly involved to benefit the estimation task; secondly, few incorrect comparisons will not influence much the accuracy of the result, making this approach more robust than the conventional approach; finally, we propose to incorporate the deep learning architecture for training, which extracts robust facial features for increasing the effectiveness of classification. In comparison to the best results from the state-of-the-art methods, the VRank showed a significant outperformance on all the benchmarks, with a relative improvement of 5.74% ∼ 69.45% (FG-NET), 19.09% ∼ 68.71% (MORPH), and 0.55% ∼ 17.73% (IoG).

UR - http://www.scopus.com/inward/record.url?scp=84960345403&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84960345403&partnerID=8YFLogxK

U2 - 10.1109/MMSP.2015.7340789

DO - 10.1109/MMSP.2015.7340789

M3 - Conference contribution

AN - SCOPUS:84960345403

T3 - 2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015

BT - 2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Lim T, Hua KL, Wang HC, Zhao KW, Hu MC, Cheng WH. VRank: Voting system on Ranking model for human age estimation. In 2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015. Institute of Electrical and Electronics Engineers Inc. 2015. 7340789. (2015 IEEE 17th International Workshop on Multimedia Signal Processing, MMSP 2015). https://doi.org/10.1109/MMSP.2015.7340789