Segmentation And Representation Of Data Dependent Label Distribution Learning For Age Estimation Using CNN

Abstract

We propose an approach for age estimation from  uncon- strained images based on deep convolutional neural net- works (DCNN). Our method consists of four steps: face detection, face alignment, DCNN-based feature extraction and neural network regression for age estimation. The pro- posed approach exploits two insights: (1) Features obtained from DCNN trained for face-identification task can be used for age estimation. (2) The three-layer neural network re- gression method trained on Gaussian loss performs better than traditional regression methods for apparent age esti- mation. Our method is evaluated on the apparent age esti- mation challenge developed  which it achieves the error of 0.373.