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

Light field image saliency feature extraction, information fusion and prediction loss evaluation method

A light field image and fusion method technology, applied in image enhancement, image analysis, image data processing, etc., can solve the problems of inconsistency between optimization direction and evaluation direction, lack of overall consideration of salient objects, etc., and achieve good results

Pending Publication Date: 2020-11-20
PEKING UNIV SHENZHEN GRADUATE SCHOOL
View PDF0 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The main purpose of the present invention is to provide a light field image information fusion method, a prediction loss evaluation method, a light field image saliency feature extraction method and a computer storage medium, aiming at solving the problem of how to effectively fuse information of multiple images in the prior art
Aims to solve the problem of lack of overall consideration of salient objects in the existing technology
It aims to solve the problem of inconsistency between the optimization direction and the evaluation direction in the existing technology

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
  • Light field image saliency feature extraction, information fusion and prediction loss evaluation method
  • Light field image saliency feature extraction, information fusion and prediction loss evaluation method
  • Light field image saliency feature extraction, information fusion and prediction loss evaluation method

Examples

Experimental program
Comparison scheme
Effect test

no. 1 example

[0074] refer to figure 2 , figure 2 It is the first embodiment of the light field image information fusion method of the present invention, and the light field image information fusion method includes the following steps:

[0075] Step S110, performing feature detection on each light field image to obtain an initial feature map;

[0076] Step S120, assigning a relatively larger weight to the area related to the salient feature in each initial feature map, and obtaining a corresponding number of first feature images;

[0077] Step S130, assigning a relatively larger weight to the channel related to the salient feature in each first feature image to obtain a corresponding number of second feature images;

[0078] Step S140, selecting images related to salient features from all the second feature images and assigning relatively greater weights to obtain a corresponding number of third feature images;

[0079] Step S150, adding all the third feature images to obtain a fused l...

no. 2 example

[0087] refer to image 3 , image 3 It is the second embodiment of the light field image information fusion method of the present invention, and the light field image information fusion method includes the following steps:

[0088] Step S210, performing feature detection on each light field image to obtain an initial feature map;

[0089] Step S220, performing convolution on the initial feature map;

[0090] Step S230, performing the probability estimation of the binary classification function on the convolution result;

[0091] Step S240, attaching the probability estimation result to the initial feature map to obtain the corresponding first feature image;

[0092] Step S250, assigning a relatively larger weight to the channel related to the salient feature in each first feature image to obtain a corresponding number of second feature images;

[0093] Step S260, selecting images related to salient features from all the second feature images and assigning relatively greate...

no. 3 example

[0099] refer to Figure 4 , Figure 4 It is the third embodiment of the light field image information fusion method of the present invention, and the light field image information fusion method includes the following steps:

[0100] Step S310, performing feature detection on each light field image to obtain an initial feature map;

[0101] Step S320, assigning a relatively larger weight to the area related to the salient feature in each initial feature map to obtain a corresponding number of first feature images;

[0102] Step S330, taking the global average value of the first feature image;

[0103] Step S340, convolving the image after taking the global average value;

[0104] Step S350, performing multi-classification function probability estimation on the convolution result;

[0105] Step S360, attaching the probability estimation result to the first feature image to obtain a corresponding second feature image;

[0106] Step S370, selecting images related to salient f...

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 light field image information fusion method, which comprises the following steps of: carrying out feature detection on each light field image to obtain an initial feature map; assigning a relatively larger weight to the region related to the saliency feature in each initial feature map to obtain a corresponding number of first feature images; endowing a channel related tothe saliency feature in each first feature image with a relatively greater weight to obtain a corresponding number of second feature images; selecting an image related to the saliency feature from all the second feature images, and endowing the image with a relatively larger weight to obtain a corresponding number of third feature images; and adding all the third feature images to obtain a fusedlight field feature image. The invention further discloses a device and a computer readable storage medium, and the problem of how to effectively fuse information of multiple images is solved.

Description

technical field [0001] The invention relates to the technical field of computer vision and image processing, and in particular to a light field image information fusion method, a prediction loss evaluation method, a light field image salient feature extraction method and a computer storage medium. Background technique [0002] Salient object detection is to detect the most attractive object in the image and separate it as the foreground. [0003] In the traditional light field saliency target detection method, the light field data contains multiple images, and the multiple images have complementary information, but simple splicing or addition cannot make good use of the information of each image in the light field. The complementarity between them also retains a large amount of redundant information. Therefore, there is still a problem of how to effectively fuse information of multiple images in the prior art. [0004] Most of the traditional deep learning methods use the ...

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): G06K9/46G06K9/62G06N3/04G06T5/50
CPCG06T5/50G06T2207/20221G06V10/462G06V2201/07G06N3/045G06F18/214
Inventor 高伟范松林
Owner PEKING UNIV SHENZHEN GRADUATE SCHOOL
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