Facial image normalization method based on self-adaptive multi-column depth model

A deep model and face image technology, applied in the field of face recognition, can solve problems such as difficult to achieve practical effects

Active Publication Date: 2015-02-18
CHONGQING ZHONGKE YUNCONG TECH CO LTD
View PDF1 Cites 40 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

For example, obtaining 3D data requires additional computation and resources; and deriving 3D models from 2D data is an ill-posed problem; statistical lighting models are usually obtained from constrained environments and cannot be well extended to practical applications
More importantly, the above methods can only remove the influence of a few unfavorable factors
In actual situations, various influencing factors such as illumination, angle, expression, and occlusion usually interact with each other, and the effect of one variable factor is usually interrelated with other variable factors. Dealing with one or two variable factors will be difficult to achieve a real practical effect due to the influence of other factors

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
  • Facial image normalization method based on self-adaptive multi-column depth model
  • Facial image normalization method based on self-adaptive multi-column depth model
  • Facial image normalization method based on self-adaptive multi-column depth model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0039] The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0040] A kind of face image normalization method based on self-adaptive multi-column depth model provided by the present invention comprises the following steps:

[0041] Step 1, building an adaptive multi-column depth model. The adaptive multi-column depth model is composed of multiple deep neural networks (Deep Neural Network, DNN) linearly combined, such as figure 1 As shown, each DNN ( figure 1 A column in ) is used to remove the influence of a certain type of factor (one or a combination of factors such as lighting, angle, occlusion, expression, etc.), and its parameters are also trained using data affected by this factor. Input an uncorrected image, each DNN corrects it with certain factors and outputs a restored image, and the weight prediction module predicts the optimal The weights are combined (i.e., the weights are adaptively ...

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 relates to a facial image normalization method based on a self-adaptive multi-column depth model. The method includes the following steps of S1, self-adaptive multi-column depth model building; S2, self-adaptive multi-column depth model training and S3, target facial image normalization. According to facial image normalization method, linear combination is conducted on the multi-column depth model to achieve united correction of various negative factors affecting facial recognition, and the optimized weight value of each column of depth models is computed in a self-adaptive mode through a nonlinear optimization method, that is, the correction factor of each factors is adjusted in the self-adaptive mode according to an input image. Compared with a traditional method that facial correction is conducted through a single depth neural network model, the facial image normalization method based on the self-adaptive multi-column depth model has higher robustness for various changed factors.

Description

technical field [0001] The invention belongs to the technical field of face recognition, and relates to a face image normalization method based on an adaptive multi-column depth model. Background technique [0002] As a typical biometric identification technology, face recognition has been favored by people for its naturalness, high acceptability, and ease of concealment. Broad application prospects. However, the best face recognition systems in the world can only basically meet the requirements of general applications when the users are more cooperative and the acquisition conditions are ideal. However, in an unconstrained environment (users do not cooperate, non-ideal acquisition conditions), due to the poor stability of face features and the great influence of various external conditions (such as different lighting conditions and occlusion and other factors), the face There are great difficulties in identification. Face image normalization refers to correcting face ima...

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 Applications(China)
IPC IPC(8): G06K9/00G06K9/62
CPCG06N3/088G06V40/161
Inventor 刘艳飞周祥东周曦
Owner CHONGQING ZHONGKE YUNCONG TECH CO LTD
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