DOI Number:10.19651/j.cnki.emt.2107612

Journal:Electronics Measurement Technology

Key Words:age estimation;dep learning;adversarial dropout regularization;convolutional neural network

Abstract:Facial appearances change slowly for adults， and the age estimation of adults in adjacent age groups is still challenging. Aiming at this problem， this paper introduced the adversarial training method into age estimation and proposed an age estimation method based on adversarial dropout regularization(ADR). The age feature learner and the discriminator are trained via the adversarial training method， and then the ability of age feature learning(especially the adjacent age group features) improves. Experimental results on three classic datasets (UTKFace， MORPH and Adience)show that the proposed model improves the accuracy of UTKFace from 42.8% to 81.6% and improves the accuracy of MORPH from 39.8% to 69.8% . Moreover， the accuracy of Adience is 63.3% . Being compared with other 4clasic models， the model using the neural networks of 5layers achieves better results than other deep neural networks and outperforms other methods with an average 17.5% higher accuracy for the young and middle-aged(15~53years old)， which shows that our model improves the performance significantly on age estimation task， and has the practical value.

Indexed by:Journal paper

Discipline:Engineering

Document Type:J

Volume:45

Issue:1

Translation or Not:no

Date of Publication:2022-01-28