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

Network model and method for face shielding detection based on graph reasoning

A network model and face occlusion technology, applied in the network model field of face occlusion detection, can solve the problems of final result impact, loss of original image information, image size cannot always meet network input requirements, etc., to improve training speed, The effect of increasing network depth and good learning performance

Active Publication Date: 2021-09-07
SHANXI UNIV
View PDF3 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

But in practice, the image size cannot always meet the input requirements of the network
Generally, cropping and stretching methods are used for preprocessing, but part of the original image information will be lost, which will affect the final result.

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
  • Network model and method for face shielding detection based on graph reasoning
  • Network model and method for face shielding detection based on graph reasoning
  • Network model and method for face shielding detection based on graph reasoning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0051] Network Model for Face Occlusion Detection Based on Graph Reasoning

[0052] like figure 1 As shown, the network model of face occlusion detection based on graph reasoning includes three parts: feature extraction network, graph attention reasoning module, and decoding; among them, the feature extraction network part includes residual network ResNet101 model and spatial pyramid pooling network, which are used to extract Face features; the graph attention reasoning module includes four sub-modules: graph projection, graph convolution, graph reasoning, and graph reprojection, which are used to obtain face feature vectors containing occluded parts; the decoding part is used to output the final mask containing occluded parts. Membrane face images and detect occluded parts.

Embodiment 2

[0054] A method for face occlusion detection based on graph reasoning, comprising the following steps

[0055] Step 1, extraction of face features (low-level features, high-level features containing occlusion information, edge features); specifically: use the residual network ResNet101 model to extract preliminary features to obtain low-level features; use spatial pyramid pooling to The output of the ResNet101 model is mapped to obtain high-level features containing occlusion information; the features output by the first, second, and fourth convolutional layers of the residual network ResNet101 model are obtained, and the edge features are obtained through the operation of the edge operator.

[0056] Step 2, obtain the face feature vector containing the occluded part; that is, use non-local operations in the graph projection sub-module to perform projection calculation on the high-level features and edge features obtained by the feature extraction network, and map the high-leve...

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 the technical field of artificial intelligence, in particular to a network model and method for face shielding detection based on graph reasoning. In order to solve the problem that the recognition precision of shielded faces is affected by the framework limitation and calculation complexity of an existing convolutional neural network, the invention provides a network model for face shielding detection based on graph reasoning. The network model comprises a feature extraction network, a graph attention reasoning module (GARM) and a decoding (deconvolution) part. Meanwhile, low-layer, high-layer and edge features of the face are extracted through a residual network and spatial pyramid pooling, similar pixel features are projected to graph nodes through graph projection, a projection data relation between the nodes is calculated to infer and analyze a possibly-shielded area, pixels are distributed to the area for detection, and finally a face shielded area is detected. Data sets such as Helen are adopted for model training and testing, and experiments prove that the detection precision and the segmentation precision of the method are superior to those of other current neural network face occlusion detection methods.

Description

technical field [0001] The invention relates to the technical field of artificial intelligence, in particular to a network model and method for human face occlusion detection based on graph reasoning. Background technique [0002] Face recognition has the advantages of simple image acquisition, low cost, and no need to touch the target during the identification process, so the application range of face recognition is becoming wider and wider. However, in the image acquisition process of the actual face recognition system, there are often uncertain factors such as illumination, posture, and occlusion, among which the occlusion factor accounts for a large proportion, and the traditional face recognition method is not good at recognizing it. How to effectively deal with it Occlusion problems and improving recognition efficiency are still one of the difficulties in face recognition systems. [0003] In order to solve the problem of face occlusion, Wu et al. proposed an occluded...

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): G06K9/00G06N3/08G06N5/04
CPCG06N5/04G06N3/08
Inventor 张丽红司春晖
Owner SHANXI 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