Face age estimation method based on inverted residual network

Abstract

The invention discloses a face age estimation method based on an inverted residual network. The face age estimation method comprises the following steps: 1, performing preprocessing such as face detection and face alignment on a face data set; 2, dividing the data set into a training set and a test set; 3, performing data enhancement operation on the training set to serve as input of a training sample; 4, establishing a network model based on the inverted residual error; 5, taking the training sample after data enhancement as the input of the model, and training to obtain a final target training model based on the inverted residual network by using a back propagation minimum loss function; and 6, testing the target training model obtained in the step 5 by using a test set to obtain age estimation of the tested face image. According to the method, a traditional deep learning network model is abandoned, the network model based on the inverted residual error is adopted for face age estimation, on the premise that the age estimation precision is not reduced, parameters of the network model are greatly reduced, and the performance of the network model is remarkably improved.

Description

technical field [0001] The invention relates to the technical fields of image recognition and deep learning, in particular to a face age estimation method based on an inverted residual network. Background technique [0002] Face images contain a lot of information, such as identity, expression, posture, gender, and age. Age is an important biological feature of a person and can be applied to various scenarios: such as an age-based human-computer interaction system, which provides Different human-computer interaction interfaces to better serve users; age-based access control, such as prohibiting minors from accessing pornographic websites, purchasing tobacco and alcohol, etc.; personalized marketing in e-commerce, using different age groups for users marketing techniques. [0003] There are two main types of current face age classification methods: one is based on traditional machine learning methods; the other is based on deep learning methods. [0004] In recent years, wi...

AI Technical Summary

Problems solved by technology

The face age estimation method based on the residual network can solve the degradation problem well and improve the network performance, but there are still too many network parameters caused by the complexity of the network model, the real-time performance of the model is poor, and the device is overly dependent on the device, which is not effective on the mobile terminal. poor problem

Images