Facial attractiveness computation by label distribution learning with deep CNN and geometric features

Shu Liu, Bo Li, Yang Yu Fan, Zhe Quo, Ashok K Samal

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

6 Citations (Scopus)

Abstract

Facial attractiveness computation is a challenging task because of the lack of labeled data and discriminative features. In this paper, an end-to-end label distribution learning (LDL) framework with deep convolutional neural network (CNN) and geometric features is proposed to meet these two challenges. Different from the previous work, we recast this task as an LDL problem. Compared with the single label regression, the LDL could improve the generalization ability of our model significantly. In addition, we propose some kinds of geometric features as well as an incremental feature selection method, which could select hundred-dimensional discriminative geometric features from an exhaustive pool of raw features. More importantly, we find these selected geometric features are complementary to CNN features. Extensive experiments are carried out on the SCUT-FBP dataset, where our approach achieves superior performance in comparison to the state-of-the-arts.

Original languageEnglish (US)
Title of host publication2017 IEEE International Conference on Multimedia and Expo, ICME 2017
PublisherIEEE Computer Society
Pages1344-1349
Number of pages6
ISBN (Electronic)9781509060672
DOIs
StatePublished - Aug 28 2017
Event2017 IEEE International Conference on Multimedia and Expo, ICME 2017 - Hong Kong, Hong Kong
Duration: Jul 10 2017Jul 14 2017

Publication series

NameProceedings - IEEE International Conference on Multimedia and Expo
ISSN (Print)1945-7871
ISSN (Electronic)1945-788X

Other

Other2017 IEEE International Conference on Multimedia and Expo, ICME 2017
CountryHong Kong
CityHong Kong
Period7/10/177/14/17

Fingerprint

Labels
Neural networks
Feature extraction
Experiments

Keywords

  • Deep CNN
  • Facial attractiveness computation
  • Geometric features
  • LDL

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Computer Science Applications

Cite this

Liu, S., Li, B., Fan, Y. Y., Quo, Z., & Samal, A. K. (2017). Facial attractiveness computation by label distribution learning with deep CNN and geometric features. In 2017 IEEE International Conference on Multimedia and Expo, ICME 2017 (pp. 1344-1349). [8019454] (Proceedings - IEEE International Conference on Multimedia and Expo). IEEE Computer Society. https://doi.org/10.1109/ICME.2017.8019454

Facial attractiveness computation by label distribution learning with deep CNN and geometric features. / Liu, Shu; Li, Bo; Fan, Yang Yu; Quo, Zhe; Samal, Ashok K.

2017 IEEE International Conference on Multimedia and Expo, ICME 2017. IEEE Computer Society, 2017. p. 1344-1349 8019454 (Proceedings - IEEE International Conference on Multimedia and Expo).

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

Liu, S, Li, B, Fan, YY, Quo, Z & Samal, AK 2017, Facial attractiveness computation by label distribution learning with deep CNN and geometric features. in 2017 IEEE International Conference on Multimedia and Expo, ICME 2017., 8019454, Proceedings - IEEE International Conference on Multimedia and Expo, IEEE Computer Society, pp. 1344-1349, 2017 IEEE International Conference on Multimedia and Expo, ICME 2017, Hong Kong, Hong Kong, 7/10/17. https://doi.org/10.1109/ICME.2017.8019454
Liu S, Li B, Fan YY, Quo Z, Samal AK. Facial attractiveness computation by label distribution learning with deep CNN and geometric features. In 2017 IEEE International Conference on Multimedia and Expo, ICME 2017. IEEE Computer Society. 2017. p. 1344-1349. 8019454. (Proceedings - IEEE International Conference on Multimedia and Expo). https://doi.org/10.1109/ICME.2017.8019454
Liu, Shu ; Li, Bo ; Fan, Yang Yu ; Quo, Zhe ; Samal, Ashok K. / Facial attractiveness computation by label distribution learning with deep CNN and geometric features. 2017 IEEE International Conference on Multimedia and Expo, ICME 2017. IEEE Computer Society, 2017. pp. 1344-1349 (Proceedings - IEEE International Conference on Multimedia and Expo).
@inproceedings{f6cfa764adaa481ca7c71360f76326b9,
title = "Facial attractiveness computation by label distribution learning with deep CNN and geometric features",
abstract = "Facial attractiveness computation is a challenging task because of the lack of labeled data and discriminative features. In this paper, an end-to-end label distribution learning (LDL) framework with deep convolutional neural network (CNN) and geometric features is proposed to meet these two challenges. Different from the previous work, we recast this task as an LDL problem. Compared with the single label regression, the LDL could improve the generalization ability of our model significantly. In addition, we propose some kinds of geometric features as well as an incremental feature selection method, which could select hundred-dimensional discriminative geometric features from an exhaustive pool of raw features. More importantly, we find these selected geometric features are complementary to CNN features. Extensive experiments are carried out on the SCUT-FBP dataset, where our approach achieves superior performance in comparison to the state-of-the-arts.",
keywords = "Deep CNN, Facial attractiveness computation, Geometric features, LDL",
author = "Shu Liu and Bo Li and Fan, {Yang Yu} and Zhe Quo and Samal, {Ashok K}",
year = "2017",
month = "8",
day = "28",
doi = "10.1109/ICME.2017.8019454",
language = "English (US)",
series = "Proceedings - IEEE International Conference on Multimedia and Expo",
publisher = "IEEE Computer Society",
pages = "1344--1349",
booktitle = "2017 IEEE International Conference on Multimedia and Expo, ICME 2017",

}

TY - GEN

T1 - Facial attractiveness computation by label distribution learning with deep CNN and geometric features

AU - Liu, Shu

AU - Li, Bo

AU - Fan, Yang Yu

AU - Quo, Zhe

AU - Samal, Ashok K

PY - 2017/8/28

Y1 - 2017/8/28

N2 - Facial attractiveness computation is a challenging task because of the lack of labeled data and discriminative features. In this paper, an end-to-end label distribution learning (LDL) framework with deep convolutional neural network (CNN) and geometric features is proposed to meet these two challenges. Different from the previous work, we recast this task as an LDL problem. Compared with the single label regression, the LDL could improve the generalization ability of our model significantly. In addition, we propose some kinds of geometric features as well as an incremental feature selection method, which could select hundred-dimensional discriminative geometric features from an exhaustive pool of raw features. More importantly, we find these selected geometric features are complementary to CNN features. Extensive experiments are carried out on the SCUT-FBP dataset, where our approach achieves superior performance in comparison to the state-of-the-arts.

AB - Facial attractiveness computation is a challenging task because of the lack of labeled data and discriminative features. In this paper, an end-to-end label distribution learning (LDL) framework with deep convolutional neural network (CNN) and geometric features is proposed to meet these two challenges. Different from the previous work, we recast this task as an LDL problem. Compared with the single label regression, the LDL could improve the generalization ability of our model significantly. In addition, we propose some kinds of geometric features as well as an incremental feature selection method, which could select hundred-dimensional discriminative geometric features from an exhaustive pool of raw features. More importantly, we find these selected geometric features are complementary to CNN features. Extensive experiments are carried out on the SCUT-FBP dataset, where our approach achieves superior performance in comparison to the state-of-the-arts.

KW - Deep CNN

KW - Facial attractiveness computation

KW - Geometric features

KW - LDL

UR - http://www.scopus.com/inward/record.url?scp=85030212231&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85030212231&partnerID=8YFLogxK

U2 - 10.1109/ICME.2017.8019454

DO - 10.1109/ICME.2017.8019454

M3 - Conference contribution

AN - SCOPUS:85030212231

T3 - Proceedings - IEEE International Conference on Multimedia and Expo

SP - 1344

EP - 1349

BT - 2017 IEEE International Conference on Multimedia and Expo, ICME 2017

PB - IEEE Computer Society

ER -