Abstract
We address the problem of perceived age estimation from face images and propose a new semi-supervised age prediction method that involves two novel aspects. The first novelty is an efficient active learning strategy for reducing the cost of labeling face samples. Given a large number of unlabeled face samples, we reveal the cluster structure of the data and propose to label cluster representative samples for covering as many clusters as possible. This simple sampling strategy allows us to boost the performance of a manifold-based semi-supervised learning method only with a relatively small number of labeled samples. The second contribution is to take the heterogeneous characteristics of human age perception into account. It is rare to misregard the age of a 5-year-old child as 15 years old, but the age of a 35-year-old person is often misregarded as 45 years old. Thus, magnitude of the error is different depending on subjects' age. We carried out a large-scale questionnaire survey for quantifying human age perception characteristics and propose to encode the quantified characteristics by weighted regression. Consequently, our proposed method is expressed in the form of weighted least-squares with a manifold regularize, which is scalable to massive datasets. Through real-world age estimation experiments, we demonstrate the usefulness of the proposed method.
| Original language | English |
|---|---|
| Title of host publication | VISAPP 2010 - Proceedings of the International Conference on Computer Vision Theory and Applications |
| Pages | 319-324 |
| Number of pages | 6 |
| Volume | 2 |
| Publication status | Published - 2010 |
| Externally published | Yes |
| Event | 5th International Conference on Computer Vision Theory and Applications, VISAPP 2010 - Angers, France Duration: 2010 May 17 → 2010 May 21 |
Other
| Other | 5th International Conference on Computer Vision Theory and Applications, VISAPP 2010 |
|---|---|
| Country | France |
| City | Angers |
| Period | 10/5/17 → 10/5/21 |
Fingerprint
Keywords
- Active sample selection
- Human age perception
- Manifold regularization
- Perceived age estimation
- Semi-supervised learning
- Weighted regression
ASJC Scopus subject areas
- Computational Theory and Mathematics
- Computer Science Applications
- Computer Vision and Pattern Recognition
Cite this
Semi-supervised estimation of perceived age from face images. / Ueki, Kazuya; Sugiyama, Masashi; Ihara, Yasuyuki.
VISAPP 2010 - Proceedings of the International Conference on Computer Vision Theory and Applications. Vol. 2 2010. p. 319-324.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Semi-supervised estimation of perceived age from face images
AU - Ueki, Kazuya
AU - Sugiyama, Masashi
AU - Ihara, Yasuyuki
PY - 2010
Y1 - 2010
N2 - We address the problem of perceived age estimation from face images and propose a new semi-supervised age prediction method that involves two novel aspects. The first novelty is an efficient active learning strategy for reducing the cost of labeling face samples. Given a large number of unlabeled face samples, we reveal the cluster structure of the data and propose to label cluster representative samples for covering as many clusters as possible. This simple sampling strategy allows us to boost the performance of a manifold-based semi-supervised learning method only with a relatively small number of labeled samples. The second contribution is to take the heterogeneous characteristics of human age perception into account. It is rare to misregard the age of a 5-year-old child as 15 years old, but the age of a 35-year-old person is often misregarded as 45 years old. Thus, magnitude of the error is different depending on subjects' age. We carried out a large-scale questionnaire survey for quantifying human age perception characteristics and propose to encode the quantified characteristics by weighted regression. Consequently, our proposed method is expressed in the form of weighted least-squares with a manifold regularize, which is scalable to massive datasets. Through real-world age estimation experiments, we demonstrate the usefulness of the proposed method.
AB - We address the problem of perceived age estimation from face images and propose a new semi-supervised age prediction method that involves two novel aspects. The first novelty is an efficient active learning strategy for reducing the cost of labeling face samples. Given a large number of unlabeled face samples, we reveal the cluster structure of the data and propose to label cluster representative samples for covering as many clusters as possible. This simple sampling strategy allows us to boost the performance of a manifold-based semi-supervised learning method only with a relatively small number of labeled samples. The second contribution is to take the heterogeneous characteristics of human age perception into account. It is rare to misregard the age of a 5-year-old child as 15 years old, but the age of a 35-year-old person is often misregarded as 45 years old. Thus, magnitude of the error is different depending on subjects' age. We carried out a large-scale questionnaire survey for quantifying human age perception characteristics and propose to encode the quantified characteristics by weighted regression. Consequently, our proposed method is expressed in the form of weighted least-squares with a manifold regularize, which is scalable to massive datasets. Through real-world age estimation experiments, we demonstrate the usefulness of the proposed method.
KW - Active sample selection
KW - Human age perception
KW - Manifold regularization
KW - Perceived age estimation
KW - Semi-supervised learning
KW - Weighted regression
UR - http://www.scopus.com/inward/record.url?scp=77956309492&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77956309492&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:77956309492
SN - 9789896740283
VL - 2
SP - 319
EP - 324
BT - VISAPP 2010 - Proceedings of the International Conference on Computer Vision Theory and Applications
ER -