Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Expression recognition method based on progressive classification

A facial expression recognition, progressive technology, applied in the field of image recognition, can solve the problems of accurate improvement, the accuracy of facial expression recognition that cannot be easily confused, and achieve the effect of reducing impact, reducing false recognition rate and improving accuracy.

Pending Publication Date: 2021-11-02
SOUTHWEST PETROLEUM UNIV
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although Document 2, Document 3 and Document 4 enrich the features extracted by convolutional neural network, they cannot accurately improve the accuracy of facial expression recognition for confusing categories

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 recognition method based on progressive classification
  • Expression recognition method based on progressive classification
  • Expression recognition method based on progressive classification

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0023] Such as figure 1 Shown, a kind of method based on the facial expression recognition method of progressive classification, comprises the following steps:

[0024] 1. Input the image to be recognized and perform enhanced preprocessing on it;

[0025] Specifically, the specific operation is: perform face detection on the expression image, and obtain the face image in the expression image; perform image enhancement by random horizontal flipping and random cropping into an image with a size of 44×44 pixels; Image expansion is achieved by cutting the four corners of the image and the center point.

[0026] 2. Construct two-level image tags. The first-level image tags divide expressions into confusing expressions and less-confusing expressions, and the second-level image tags are specific expressions;

[0027] Specifically, after constructing the two-level tags, they are named after 0 and 1 respectively, among which, 0 is the unconfusing expression label set, and 1 is the ea...

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 provides an expression recognition method based on progressive classification. The expression recognition method comprises the following steps: firstly, carrying out image enhancement by utilizing an expression image; reprocessing the labels of the original expression data, and processing the original labels into a two-stage label form of coarse labels and fine labels; modifying the network to a certain degree by combining an expression data set on the basis of the ResNet18 network, so as to construct a coarse classification network; realizing shared parameter migration of the two fine classification networks and the coarse classification network, and the whole network being of a parallel structure from the aspect of structure. According to the method, the common problem that several types of expressions are easy to confuse in an expression recognition task can be accurately solved, the recognition accuracy of the expressions easy to confuse is obviously improved, and in addition, the recognition accuracy of the expressions not easy to confuse is also improved to a certain extent.

Description

technical field [0001] The invention belongs to the technical field of image recognition, in particular to an expression recognition method based on progressive classification. Background technique [0002] Facial expression recognition is to recognize the emotion that human beings want to express by detecting the face and face, which plays an important role in education, medical treatment, entertainment and so on. With the rise of deep learning research in recent years, the improvement of computer computing capabilities, and the introduction of rich expression data sets, expression recognition has become a research hotspot in the field of computer vision. At present, facial expression recognition methods are mainly divided into two categories: expression recognition based on traditional methods and expression recognition based on deep learning. Most of the research on expression recognition based on deep learning is based on convolutional neural network. [0003] Document ...

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): G06K9/00G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/214
Inventor 程吉祥何虹斌吴丹李志丹刘家伟
Owner SOUTHWEST PETROLEUM UNIV
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
Eureka Blog
Learn More
PatSnap group products