Facial expression recognition method based on separation mixed attention mechanism

A technology of facial expression recognition and attention, which is applied in the field of computer vision, can solve the problems of insignificant improvement in recognition rate and different feature extraction capabilities, and achieve the effect of improving recognition accuracy, reducing the amount of parameters, and good convergence

Inactive Publication Date: 2022-01-04
NANJING VOCATIONAL UNIV OF IND TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In the existing attention mechanism, channel attention and spatial attention are serially superimposed, and the feature extraction capabilities of the two kinds of attention are not considered. Experiments show that the existing attention mechanism superimposes the backbone network in facial expression recognition. The recognition rate is not significantly improved during application

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 expression recognition method based on separation mixed attention mechanism
  • Facial expression recognition method based on separation mixed attention mechanism
  • Facial expression recognition method based on separation mixed attention mechanism

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0033] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0034] As shown in the figure, a facial expression recognition method based on the separation and hybrid attention mechanism includes: In order to enhance the channel feature learning ability of the network, the present invention places the channel attention extraction operation after the channel feature extraction of depth separable convolution , the internal structure of the channel attention mechanism proposed by the present invention is as follows figure 1 Shown:

[0035] The input feature map performs a 3×3 convolution operation on each channel, and then concatenates the convolution outputs on all channels through the concat operation, and processes it through the channel attention soft threshold (CA-ST, Channel Attention SoftThresholding) mechanism , and then output after 1×1 convolution. The channel attention soft threshold processing structure...

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 a facial expression recognition method based on a separation mixed attention mechanism, and the method integrates a channel attention mechanism and a space attention mechanism, and the two attention mechanisms are respectively disposed in a depth separable convolution and after a common convolution. The channel features and the cross-channel correlation learning ability of the convolutional neural network are enhanced; soft threshold mechanisms are respectively added after the channel attention mechanism and the space attention mechanism, and noise information in attention features is inhibited. The facial expression recognition accuracy can be effectively improved, the parameter quantity of the expression recognition network is reduced, and the convergence of the expression recognition network is better.

Description

technical field [0001] The invention relates to the field of computer vision, in particular to a human facial expression recognition method based on a separation and hybrid attention mechanism. Background technique [0002] Convolutional neural network is one of the representative algorithms of deep learning, and is currently widely used in computer vision, natural language processing and other fields. With the increase of computer computing power, the structure of convolutional neural network has become more and more complex, and network structures such as LeNet, AlexNet, ResNet, DenseNet, and Inception have appeared successively. In order to make the machine learning effect better under the same network structure, an attention mechanism that simulates the human visual mechanism has emerged. At present, the commonly used attention mechanism is a hybrid attention mechanism that combines spatial attention and channel attention. Experiments show that this The mechanism combin...

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): G06K9/00G06N3/04
CPCG06N3/048G06N3/045
Inventor 余久方
Owner NANJING VOCATIONAL UNIV OF IND TECH
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