Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Age Estimation Method Based on Multi-Output Convolutional Neural Network and Ordinal Regression

A convolutional neural network and neural network technology, which is applied in the field of age estimation based on multi-output convolutional neural networks and ordered regression, which can solve the problems of lack of age estimation in large-scale age data sets and the inability to make further progress.

Active Publication Date: 2019-10-11
陕西慧眸一方智能科技有限公司
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0008] (1) Establish the Asian Face Age Dataset (AFAD), which contains 160,000 Asian face images, and each face image has an age label. This is by far the largest public age data set, which solves the problem Lack of large-scale age datasets prevents greater progress on the age estimation problem

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Age Estimation Method Based on Multi-Output Convolutional Neural Network and Ordinal Regression
  • Age Estimation Method Based on Multi-Output Convolutional Neural Network and Ordinal Regression
  • Age Estimation Method Based on Multi-Output Convolutional Neural Network and Ordinal Regression

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0026] Human age estimation from face images has a wide range of real-life applications. For example, in terms of security monitoring, by estimating the age of face images in surveillance cameras, minors can be warned or prevented from entering Internet cafes or bars to drink; in terms of commercial user management, advertisers can provide different potential customers Specific advertisements; in terms of biometric identification, face recognition can be performed over a long period of time. However, there are many deficiencies in existing age estimation techniques, such as feature extraction and regression or multi-classification separation, insufficient training data and uneven distribution. These deficiencies seriously affect the predictive accuracy of existing age estimation techniques.

[0027] The present invention has been exploring and researching in the field of computer vision. Aiming at the shortcomings of the existing age estimation technology, a method of age est...

Embodiment 2

[0036] The age estimation method based on multi-output convolutional neural network and ordered regression is the same as in embodiment 1, wherein the Asian face age data set is established in step (1), see figure 2 , including the following steps:

[0037] (1a) Build this dataset by collecting face images from Renren, a specific social network. Renren is a social network where students can connect with others, upload photos, post comments, etc., including middle school students, Wide application among many Asian students including high school, undergraduate and graduate students. Even after graduation, some still log into their Renren account to connect with old classmates. Therefore, the age of Renren users spans a long range from 15 years old to over 40 years old, which is conducive to building a dataset with a wide age span. Renren has a special photo album for each user to upload their own photos, so the images in the Asian face age dataset come from the avatar albums ...

Embodiment 3

[0044] The age estimation method based on multi-output convolutional neural network and ordered regression is the same as embodiment 1-2, wherein the training data for two classifications described in step (2) includes the following steps:

[0045] (2a) Convert age estimation into age ranking and process it with a series of binary classifiers, construct corresponding training data for each binary classifier of Asian face age dataset images, given ordered training data where x i ∈χ is the input space of the i-th image, y i ∈γ={r 1 ,r 2 ,...,r K} is the output space of the ordered sequence, r K > r K-1 >…>r 1 , K is the total number of levels, the symbol > indicates the sorting between different levels, and N is the total number of training data.

[0046] (2b) For the kth binary classifier, its binary classifier Indicates the serial number y of the i-th sample i Is it better than r k Large, defined as follows:

[0047]

[0048] equivalent to y i >r k , the dich...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an age estimation method based on a multi-output convolutional neural network and an ordered regression, which is realized by: 1. establishing an Asian face age data set (AFAD); 2. establishing training data for binary classification; 3. Train the deep convolutional neural network; 4. Input the test sample into the trained convolutional neural network; 5. Get the age estimate of the test sample. The present invention proposes a method for sorting ages, which combines ordered regression and deep learning methods to significantly improve the accuracy of age prediction performance. The present invention solves the problem that feature extraction and regression modeling are independently performed and optimized in existing age estimation methods, and can make full use of the sequence relationship of age labels to regress age estimation in an orderly manner, improving the accuracy of age estimation, and is also suitable for Asians. Face age estimation has established a large-scale database, which provides a database basis for face age estimation research. It can be widely used in age estimation of face images.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and mainly relates to an age estimation method of a face image, in particular to an age estimation method based on a multi-output convolutional neural network and ordered regression, which can be used for age estimation of a face image. Background technique [0002] Human age estimation from face images is a relatively new research direction with wide applications in real life. For example, in terms of security monitoring, through the input images of surveillance cameras, the age estimation system can warn or prevent minors from entering Internet cafes or drinking in bars; in terms of commercial user management, advertisers can provide different potential customers with specific Advertising; In terms of biometric identification, face recognition can be performed over a long period of time. [0003] G.Guo, G.Mu, Y.Fu and T.Huang introduced bio-inspired methods into the field of age estima...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00
CPCG06V40/178G06V40/172
Inventor 牛振兴魏雪周默袁博高新波华刚
Owner 陕西慧眸一方智能科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com