Near-infrared face image super-resolution reconstruction method based on deep learning

A technology of super-resolution reconstruction and face image, which is applied in the field of computer image super-resolution, can solve the problem of not being able to capture the detailed information of the face and face, and achieve the effect of improving the reconstruction effect and improving the effect of face reconstruction

Active Publication Date: 2017-09-01
福建帝视信息科技有限公司
View PDF9 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although infrared images have good adaptability to illumination changes, they cannot capture facial details

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
  • Near-infrared face image super-resolution reconstruction method based on deep learning
  • Near-infrared face image super-resolution reconstruction method based on deep learning
  • Near-infrared face image super-resolution reconstruction method based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0056] Such as Figure 1-5 As shown in one of them, the present invention discloses a method for super-resolution reconstruction of near-infrared face images based on deep learning, which includes the following steps:

[0057] Step 1, using the relative positions of the two eyes in the face image to align all the face images in the original near-infrared face image; further, the original near-infrared face image in the step 1 uses a near-infrared supplementary light device capture.

[0058] Step 2, convert the face training set obtained after alignment to a fixed ratio to obtain a training image containing K M×N super-resolution face images Where M and N are the width and height of the face image respectively, and the index i=1, 2, ..., K; further, the fixed ratio in step 2 is converted into M×N where M×N is 128×128.

[0059] Step 3, the super-resolution face image training graph Each image in S is used to generate its one-to-one corresponding K low-resolution face image ...

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 near-infrared face image super-resolution reconstruction method based on deep learning. The method comprises the following steps: 1) carrying out position alignment on all face images in an original near-infrared face image by utilizing relative position of two eyes in a face image; 2) carrying out conversion on a face training set obtained after alignment in a fixed proportion to obtain a training map I<H>i comprising K super-resolution face images; 3) carrying out processing on the super-resolution face image training map in a scaling ratio S to generate a training map I<L>i comprising K low-resolution face images in one-to-one correspondence with the K super-resolution face images; 4) obtaining a reconstructed super-resolution face image Fl"(Y) by utilizing the low-resolution face image training map I<L>i; 5) calculating Euclidean distance between the reconstructed super-resolution face image Fl"(Y) and a corresponding image in the super-resolution face image training map I<H>i; and 6) carrying out optimization based on the Euclidean distance to obtain an optimal convolution weight parameter and offset parameter. The method improves face image reconstruction effect greatly.

Description

technical field [0001] The present invention relates to the field of computer image super-resolution, in particular to a near-infrared face image super-resolution reconstruction method based on deep learning. Background technique [0002] Face image super-resolution reconstruction is a technique to reconstruct a corresponding high-resolution face image from a low-resolution face image. This technology has broad application prospects in the fields of intelligent video surveillance, face detection and recognition, facial expression recognition, face recognition and age measurement. [0003] In the actual application environment, pedestrians are often far away from the surveillance camera, or the optical resolution of the camera is not enough, and the captured face often has a low resolution and lacks a lot of facial detail feature information, so face image restoration, magnification and Identification is severely disturbed. Therefore, without the need for higher hardware eq...

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
IPC IPC(8): G06T3/40
CPCG06T3/4076
Inventor 李根童同高钦泉
Owner 福建帝视信息科技有限公司
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