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

Expression database enhancement method, training method, computing device and storage medium

A database and expression technology, applied in the computer field, can solve the problems of sample expression diversity and limited generalization ability, poor expression recognition effect, insufficient number of training samples, etc., to preserve detailed features, improve feature generalization ability, The effect of increasing diversity

Active Publication Date: 2019-06-28
SHENZHEN UNIV
View PDF3 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide an enhancement method, a training method, a computing device and a storage medium of an expression database, aiming to solve the problems existing in the prior art, such as insufficient number of training samples, limited sample expression diversity and generalization ability, etc. The problem of poor expression recognition effect caused by the loss of detailed features

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
  • Expression database enhancement method, training method, computing device and storage medium
  • Expression database enhancement method, training method, computing device and storage medium
  • Expression database enhancement method, training method, computing device and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0046] figure 1 It shows the implementation process of the expression database enhancement method provided by Embodiment 1 of the present invention. For the convenience of description, only the parts related to the embodiment of the present invention are shown, and the details are as follows:

[0047] In step S101, the original image to be processed containing the first object, the reference original image containing the second object and the reference expression image containing the third object are obtained, the first object and the second object have the same original expression, and the third object Has a target emote different from the original emote.

[0048] In this embodiment, the object refers to a human face, however, the application examples are not limited thereto, for example: animals or other human body parts that can make expressions. In particular, the object can also refer to the part covered by the human face in the image during processing.

[0049]Facial e...

Embodiment 2

[0057] On the basis of Embodiment 1, this embodiment further provides the following content:

[0058] Such as figure 2 As shown, step S102 specifically includes:

[0059] In step S201, an intermediate image is synthesized from the original image to be processed by using the shape change of the target expression of the third object relative to the original expression of the second object.

[0060] In step S202, the target expression image is synthesized from the intermediate image by utilizing the texture change of the target expression of the third object relative to the original expression of the second object.

[0061] In this embodiment, the main implementation is to change the corresponding shape of the original image to be processed to obtain an intermediate image, and then apply texture changes to the intermediate image to obtain the target expression image, and the shape change and texture change correspond to: in the reference expression image Compared with the obje...

Embodiment 3

[0067] On the basis of Embodiment 2, this embodiment further provides the following content:

[0068] Such as image 3 As shown, step S201 specifically includes:

[0069] In step S301, the original image to be processed, the reference original image and the reference expression image are subjected to unified standardization processing on the object, and there are defined between the specified feature points and the boundary points in the original image to be processed, the reference original image and the reference expression image grid relationship.

[0070] In this embodiment, the unified and standardized processing is the processing basis of step S302, which can realize the unification of the original image to be processed, the reference original image and the reference facial expression image into the set coordinate system. The unified standardization process may specifically involve: first, the location of the specified feature points on the original image to be process...

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 method is suitable for the technical field of computers, and provides an expression database enhancement method, a training method, a computing device and a storage medium. The method comprises the steps: firstly obtaining an original image to be processed, a reference original image and a reference expression image, wherein objects in the original image to be processed and in the reference original image have the same original expression, and the object in the reference expression image has different target expressions; And synthesizing a target expression image from the to-be-processedoriginal image by utilizing the change from the original expression to the target expression, so that the object in the target expression image has the target expression. Therefore, the expression change reflected by the reference image is migrated to the to-be-processed image, so that the target image is synthesized, and the number of training samples is efficiently expanded while the diversity of sample expression is remarkably increased, the feature generalization capability is improved and the facial expression detail features are reserved.

Description

technical field [0001] The invention belongs to the technical field of computers, and in particular relates to an expression database enhancement method, a training method, a computing device and a storage medium. Background technique [0002] With the development of human-computer interaction, facial expression recognition has become a hot topic in recent decades. Today, neural networks use complex structures or multiple processing layers composed of multiple nonlinear transformations to perform high-level abstraction of data and apply them to image recognition and analysis. The expression recognition technology based on neural network has surpassed various traditional methods, and through various methods such as network design, expression database enhancement, metric learning and network composition, the generalization recognition ability of neural network has been improved. [0003] In the neural network-based facial expression recognition algorithm, a large number of tr...

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): G06K9/00G06T5/50G06F16/583
Inventor 解为成沈琳琳田怡
Owner SHENZHEN UNIV
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