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

Image fusion method of a filter preset deep learning neural network

A neural network and deep learning technology, applied in the field of digital image processing, can solve the problems of fusion image information entropy and low average gradient, and achieve the effect of clear details, good fusion and rich information

Pending Publication Date: 2019-06-21
XIDIAN UNIV
View PDF3 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to propose a filter preset deep learning neural network image fusion method to solve the problem of low fusion image information entropy and average gradient in the existing multi-scale wavelet transform image fusion method

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
  • Image fusion method of a filter preset deep learning neural network
  • Image fusion method of a filter preset deep learning neural network
  • Image fusion method of a filter preset deep learning neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0026] The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0027] refer to figure 1 , the implementation steps of the example of the present invention are as follows:

[0028] Step 1, preset the filter of the single-layer deconvolution network.

[0029] The filter of the existing single-layer deconvolution network is randomly initialized. The filter of the preset deconvolution network of the present invention is a fourth-order Butterworth filter. The specific steps are as follows:

[0030] (1a) Set the number of single-layer deconvolution network filters to 4, and mark these 4 filters as {f 1 , f 2 , f 3 , f 4};

[0031] (1b) Preset one of the four filters as a 7×7 fourth-order low-pass Butterworth filter, and design a 7×7 fourth-order low-pass Butterworth filter H in the frequency domain (u,v):

[0032]

[0033] Among them, u and v are the row and column coordinates in the frequency domai...

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 image fusion method of a filter preset deep learning neural network. The image fusion method mainly solves the problem that in an existing multi-scale wavelet transform imagefusion method, the fusion image information entropy and the average gradient are low. According to the implementation scheme, the method comprises the following steps of presetting asingle-layer deconvolution network to be a four-order Butterworth filter; obtaining a sample image; using the sample image to train a single-layer deconvolution network filter; Inferring a feature map of the to-be-fused image by using the single-layer deconvolution network trained by the filter; fusing the feature map of the to-be-fused image; and performing convolution summation on the fusion feature map and a filter of the single-layer deconvolution network to generate a fusion image. According to the method, the fused image with richer information and clearer details can be stably and effectively obtained,the information entropy and the average gradient of the multi-focus fused image are improved, and the method can be used for shooting scenes of a multi-camera mobile phone and a digital camera.

Description

technical field [0001] The invention belongs to the technical field of digital image processing, and in particular relates to an image fusion method, which can be used in shooting scenes of multi-camera mobile phones and digital cameras. Background technique [0002] There are many kinds of deep learning neural networks. The deconvolution network is a network model in the deep learning neural network. Through unsupervised learning and training, it can use the learned and trained network to infer multiple feature maps of the input image. And use these inferred feature maps and deconvolution network filters to restore and reconstruct the original image. [0003] Image fusion is the process of merging multiple original images containing different information acquired by multiple sensors for the same target according to certain fusion rules to obtain an image containing important information obtained by each sensor. Image fusion technology can take into account the image inform...

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): G06T5/50G06T5/20
Inventor 那彦王强军刘赫高兴鹏刘强强
Owner XIDIAN 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

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com