Unlock instant, AI-driven research and patent intelligence for your innovation.

Cross-modal infrared polarization face recognition method based on codec

A technology of infrared polarization and codec, applied in the field of image recognition, to achieve the effect of fast and effective identification

Pending Publication Date: 2022-07-01
NORTHWESTERN POLYTECHNICAL UNIV
View PDF0 Cites 0 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 a codec-based cross-modal infrared polarization face recognition method to solve the face recognition problem under insufficient illumination conditions

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
  • Cross-modal infrared polarization face recognition method based on codec
  • Cross-modal infrared polarization face recognition method based on codec
  • Cross-modal infrared polarization face recognition method based on codec

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0028] The following specific embodiments will describe the present invention in detail.

[0029] The invention provides a cross-modal infrared polarization face recognition method based on a codec, which is processed according to the following steps:

[0030] S1. In the training phase, use each face image in the visible light face database and the infrared polarized face database to train the encoder-decoder structure:

[0031] The encoder encodes each face image correspondingly into a set of identity features and modal features, and the decoder reorganizes and decodes each set of identity features and modal features to regenerate the corresponding face image;

[0032] S2. Construct feature similarity loss function Loss similar and the reconstruction loss function Loss reconstruction Constrain the codec network.

[0033] S3. In the testing stage, use the codec trained in steps S1 and S2 to encode the visible light face images in the visible light face database one by one a...

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 discloses a cross-modal infrared polarization face recognition method based on a codec, and the method comprises the following steps: S1, training a codec structure through employing each face image in a visible light face database and an infrared polarization face database in a training stage; s2, constructing a feature similarity loss function and a reconstruction loss function, and constraining the codec structure; step S3, in a test stage, using the codec trained in the steps S1 and S2 to encode visible light face images in the visible light face database one by one and generate a database identity feature idFbase, encoding a to-be-recognized infrared polarization face image to generate a to-be-recognized identity feature idFx, and using the to-be-recognized identity feature idFx to generate a to-be-recognized identity feature idFx; and calculating the cosine similarity between the to-be-recognized identity feature idFx and the database identity feature idFbase, and matching the maximum value of the similarity with the corresponding face identity. The problem of face recognition under the condition of insufficient illumination is solved.

Description

technical field [0001] The invention belongs to the technical field of image recognition, in particular to a cross-modal infrared polarization face recognition method based on a codec. Background technique [0002] The task of face recognition is to give computers the ability to identify people based on face data. Through the processing and analysis of face data, the essential characteristics that can reflect the identity of the person are extracted, so as to establish the connection between the characteristics and the identity. After years of development, many face recognition technologies have made great progress, but they are still affected by many factors in practical applications, such as changes in lighting and posture. Therefore, cross-modal face recognition will become a research hotspot in the field of face recognition. [0003] In the case of insufficient light at night, the thermal radiation image of the face can be obtained by using infrared. Since the polariz...

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): G06V40/16G06V10/74G06T9/00
CPCG06T9/00G06F18/22
Inventor 赵永强王憬诣吴荣源
Owner NORTHWESTERN POLYTECHNICAL UNIV