A Face Age Estimation Method Based on Convolutional Neural Network for Distribution Learning

Abstract

The invention discloses a face age estimation method based on convolutional neural network for distributed learning. The overall steps are: extracting data to form an age data set; dividing the age data set into a training set and a verification set; using a deep neural network The last fully connected layer is followed by a softmax layer; conduct age estimation network training; use softmax loss and mean‑variance loss together as supervisory signals to adjust the network; evaluate the trained network model and select the best performance model; age predictions are made based on the resulting model. By introducing a new supervisory signal mean-variance loss, the present invention effectively utilizes the interrelated nature of ages, avoids operations such as manual introduction of variance, and realizes a technology that does not require any manual intervention except preprocessing.

Description

Technical field [0001] The present invention relates to an estimation method, and more particularly to a human face age estimation method for distribution based on convolutional neural networks. Background technique [0002] Techniques for age estimation through faces are mainly divided into two categories, one is to use traditional machine learning methods for feature extraction, and then optimize the target function for extracted feature design, thereby obtaining age. The current traditional machine learning method is more to explore different features of age estimation, thereby achieving better feature extraction; another method is to use deep learning technology, end-to-end implementation of feature extraction and goals The entire task optimized. In terms of deep learning, it is to train different network structures. The essence is still different from different network structures, and optimized using end-to-end advantages. [0003] In addition to most of the methods based on...

AI Technical Summary

Problems solved by technology

By using the order of age, the idea based on ranking (sequence) is proposed to solve the regression problem into a classification problem, but this method ignores the order of age and the relationship between ages, making it difficult for features to contain these Very useful information for estimating age

There is also the use of the nature of the age distribution to optimize the distribution by assuming a variance to generate a distribution. This method needs to assume a variance to carry out follow-up work, resulting in human intervention, so the age distribution is still not ideal.

Images