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

Recurrent neural network motion blur restoration method

A technology of cyclic neural network and motion blur, applied in biological neural network models, neural learning methods, neural architectures, etc., can solve problems such as motion blur

Pending Publication Date: 2020-10-27
HANGZHOU DIANZI UNIV
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The present invention solves the problem of motion blur caused by the relative motion between the camera and the scene, and proposes a motion blur recovery method of the cyclic neural network, adopts the coding exposure imaging mode, and first searches for the optimal code word according to the coding requirements, and adopts Long short-term memory network for network training

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
  • Recurrent neural network motion blur restoration method
  • Recurrent neural network motion blur restoration method
  • Recurrent neural network motion blur restoration method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0044] This embodiment proposes a recurrent neural network motion blur restoration method, refer to Figure 5 , including the following steps:

[0045] S1, setting the codeword objective function;

[0046] Step S1 specifically includes: the objective function formula is as follows:

[0047]

[0048] Among them, S represents the binary codeword sequence, F() represents the discrete Fourier transform operation, min() represents the calculation of the minimum value operation, var() represents the calculation of the variance operation, n represents the length of the binary codeword, α 1 , α 2 represent their respective weight coefficients. In the experiment, their respective full weight coefficients are a 1 = 0.4, a 2 = 0.6. Coded exposure imaging is the latest imaging mode, which changes the low-pass filter in traditional imaging to a band-pass filter, enabling it to retain more image detail information during the imaging process.

[0049] S2, obtaining the optimal codewo...

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 recurrent neural network motion blur restoration method. The method comprises the following steps: setting a codeword target function; obtaining an optimal code word through agenetic algorithm; obtaining a blurred image; preprocessing the data to generate a data set; training a long-term and short-term memory network by utilizing the data set; and inputting the test imageinto the trained long-term and short-term memory network to obtain a restored image. According to the method, a coding exposure imaging mode is adopted, optimal code word search is carried out according to coding requirements, a long-term and short-term memory network is adopted for network training, and finally a high-resolution image is obtained through training.

Description

technical field [0001] The invention relates to the field of computational optical imaging and the field of deep learning technology, in particular to a motion blur restoration method of a cyclic neural network. Background technique [0002] Motion blur image restoration has always been an important research topic in image restoration. The main cause of motion blur is that when shooting a moving target, the relative motion between the moving target and the imaging system often occurs within the exposure time of the camera, resulting in motion blur in imaging and a decrease in image resolution. In the traditional imaging process, the camera shutter is always open, and there must be many zeros in the frequency domain on the convolution filter, which makes the filter irreversible and makes motion blur a pathological problem. [0003] In order to solve the pathological problem of motion blurred images, researchers conduct research from two perspectives: image imaging mode and d...

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): G06T5/00G06N3/04G06N3/08
CPCG06N3/049G06N3/084G06T2207/10016G06T2207/20024G06T2207/20081G06T2207/20084G06T2207/20056G06T2207/20201G06N3/048G06N3/045G06T5/73
Inventor 崔光茫叶晓杰赵巨峰朱礼尧
Owner HANGZHOU DIANZI 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