A regression network age estimation method with ordered anchor replacement

DOI

10.2991/978-94-6463-419-8_25How to use a DOI?

Keywords

Age estimation; Anchored network; Regression network; Convolutional neural network

Abstract

Face age estimation is often regarded as a regression problem, and anchored regression network is a representative regression algorithm. It selects a facial feature from each age group as an anchor point for linear regression, and then combines all the weighted linear regression results to obtain the predicted age. During training, to ensure the comprehensiveness of anchor points, the regression results of all anchor points are included in the calculation. However, some selected anchor points may exhibit significant differences in appearance from the samples, leading to inaccurate age predictions. To address this issue, an improved anchor point selection method is proposed. This method initially employs k-means clustering to output initial anchor points, calculates the Euclidean distance between the initial anchor points and samples to construct a distance vector. Subsequently, based on the distance vector, it identifies several nearest anchor points and replaces the initial anchor points, termed as ordered anchor point replacement.This method enhances the acquisition of anchor points, and experiments conducted on multiple datasets show promising results. The experimental outcomes yield a lower average absolute error (MAE) of 2.54, demonstrating comparability with similar algorithms. Thus, this validates the effectiveness of the proposed method in improving the accuracy of age estimation.

Copyright

© 2024 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Tian Tan
AU  - Chunlong Hu
PY  - 2024
DA  - 2024/05/07
TI  - A regression network age estimation method with ordered anchor replacement
BT  - Proceedings of the 3rd International Academic Conference on Blockchain, Information Technology and Smart Finance (ICBIS 2024)
PB  - Atlantis Press
SP  - 195
EP  - 207
SN  - 2589-4900
UR  - https://doi.org/10.2991/978-94-6463-419-8_25
DO  - 10.2991/978-94-6463-419-8_25
ID  - Tan2024
ER  -