A Self-step Learning Face Age Estimation Method Based on Noise Removal

A face and noise technology, applied in the field of face age estimation, to achieve strong representation ability and improve learning robustness

Active Publication Date: 2022-03-15
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Obviously, during the training process of the age estimation task, the impact of noisy face images (with changes in pose, lighting, expression, occlusion, and misalignment) on the entire model is huge, but so far it has not yet emerged how to mitigate this. method of influence, the present invention will work around this angle

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
  • A Self-step Learning Face Age Estimation Method Based on Noise Removal
  • A Self-step Learning Face Age Estimation Method Based on Noise Removal
  • A Self-step Learning Face Age Estimation Method Based on Noise Removal

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0129] The present invention is based on the self-paced learning face age estimation method of noise elimination, and its realization comprises the following steps:

[0130] Step 1: Preprocess the dataset;

[0131] For Moprh II ( http: / / www.faceaginggroup.com / morph / ) face database uses MTCNN to detect facial feature points, and obtains 5 facial feature points; according to the obtained 5 facial feature point positioning results, the image is normalized to a 224*224*3 RGB image; Processed 55,130 face images with age labels.

[0132] Step 2: Build a deep regression forest;

[0133] image 3 Represents the general structure of the deep regression forest, where the circle represents the feature value output by the last fully connected layer of the convolutional neural network, the rectangular box represents the separation node of each tree, and the diamond box represents the leaf node of each tree;

[0134] with respectively represent the input and output spaces of the d...

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 a method for estimating the age of a human face through self-paced learning based on noise elimination, which belongs to the fields of computer vision and machine learning. Since face images often have changes in posture, lighting, expression, occlusion, and misalignment, face images are divided into simple images (the absolute error between the predicted age and the actual age is small) and difficult images (the difference between the predicted age and the actual age is small). The absolute error between them is large), and under the framework of self-paced learning, the strategy of "from simple pictures to difficult pictures" is adopted to train the deep regression network to establish the nonlinear mapping relationship between facial features and target age. At the same time, the present invention proposes The cap() function will eliminate the noise image in the training sample, thereby removing the influence of the noise image on the model. This method makes full use of the cap() function, self-paced learning and deep regression forest, which ensures that the extracted facial features have Powerful representation capabilities improve the accuracy and robustness of existing methods. This method can be applied to human-computer interaction, age-based security control, social network entertainment, and age-differentiated advertising.

Description

technical field [0001] The invention belongs to the technical field of computer vision, relates to face age estimation technology, and is mainly applied to aspects such as human-computer interaction, age-based security control, social network entertainment, and age-differentiated advertisement. Background technique [0002] Face age estimation technology refers to the technology of automatically estimating the age of faces after analyzing the features of faces through computer algorithms. Since this technology can be widely used in human-computer interaction, age-based security control, social network entertainment, and age-differentiated advertising, it is a hot research topic in the fields of computer vision and machine learning in recent years. The existing face age estimation algorithms are mainly divided into methods based on shallow models and methods based on deep learning. [0003] The rationale behind the shallow model-based approach is to decompose the task into t...

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): G06V40/16G06N3/04G06N3/08
CPCG06N3/08G06V40/161G06V40/178G06N3/045
Inventor 潘力立艾仕杰
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap