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

Image segmentation quality evaluation network system, method and system based on sorting constraint

A quality evaluation and image segmentation technology, applied in image analysis, image data processing, instruments, etc., can solve the problem of inability to correctly predict the quality ranking relationship, and achieve the effect of reducing the amount of network training

Active Publication Date: 2021-04-06
NARI INFORMATION & COMM TECH
View PDF2 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although the predicted results f(A) and f(B) are close to the real quality scores Sa and Sb, the relationship between f(A) and f(B) is obviously opposite to the ordering relationship between Sa and Sb, and the two cannot be correctly predicted. Quality Ranking Relationship Between Segmentation Results

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 segmentation quality evaluation network system, method and system based on sorting constraint
  • Image segmentation quality evaluation network system, method and system based on sorting constraint
  • Image segmentation quality evaluation network system, method and system based on sorting constraint

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0028]Such asfigure 1 As shown, an image segmentation quality evaluation network system based on sorting constraints, including two quality evaluation sub-networks Q1 and Q2 shared by two volume nuclear parameters, and mass prediction loss L1 is the loss function of quality evaluation subnet Q1, quality prediction loss L2 is the loss function of the quality evaluation subnet Q2, the sorting loss LR is used to constrain the size of the quality evaluation subnet Q1 and the quality evaluation subnet Q2 predicted size sorting relationship; the quality evaluation subnet Q1 is a twin network, including two convolutions The characteristics of nuclear parameters shared and the tributary C1 and C2, one feature conversion module, and a quality prediction module; the feature conversion module fuses the first feature extracted by the buses C1 and C2, converted into a second feature; the mass prediction module will Twenty-feature mapped into quality predictive values; quality evaluation subnet Q...

Embodiment 2

[0040]According to an embodiment, a sort-based image segmentation quality evaluation network system is provided, this embodiment provides an image-based segmentation quality evaluation method based on sorting constraints, such asimage 3 As shown, including: collecting the image in the existing public image split database and its artificial indication, for any image, generates a m-split spectrum using the existing M image segmentation method to form a database sample; according to artificial indicators , Use the segmentation evaluation standard IOU (InterSection overunion) to generate the quality score of each split spectrum, obtain the quality tag of the image split spectrum, form a sample tag; divide the database sample into training set, verification set and test set; use high sort sample selection Methods The training samples and corresponding sample labels are selected from the training set, and the sorting-based image segmentation quality evaluation network system RSQAN is used...

Embodiment 3

[0050]Based on the alternative image segmentation quality evaluation network system and the first embodiment-based image segmentation quality evaluation method according to the first embodiment, this embodiment provides an image segmentation quality evaluation system based on sort-based constraints. Including: The first module, for training completion based order-based image-based image segmentation quality evaluation network system RSQAN, constructs an image division quality prediction network system RSQPN based on the sorting constraint. The second module is used to evaluate the quality of the current split spectrum with the predicted value of the RSQPN output by image and its split spectrum as an input of the RSQPN.

[0051]Those skilled in the art will appreciate that embodiments of the present application can be provided as a method, system, or computer program product. Accordingly, the present application may employ a full hardware embodiment, a full software embodiment, or in co...

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 an image segmentation quality evaluation network system, method and system based on sorting constraints, belongs to the technical field of image segmentation, and achieves the precise prediction of the quality of segmented spectrums, especially the precise prediction of the quality relation between the segmented spectrums. The network system comprises a parameter-shared quality evaluation sub-network Q1 and a parameter-shared quality evaluation sub-network Q2, and the quality evaluation sub-network Q1 is a twin network and comprises two parameter-shared feature extraction branches C1 and C2, a feature conversion module and a quality prediction module; the feature conversion module fuses the first features extracted by the branches C1 and C2 and converts the first features into second features; the quality prediction module maps the second feature into a quality prediction value; and the quality evaluation sub-network Q2 and the quality evaluation sub-network Q1 have the same structure.

Description

Technical field[0001]The present invention belongs to the field of image segmentation, and in particular to an image division quality evaluation network system, method, and system based on sorting constraints.Background technique[0002]Image split is one of the main research in the field of computer vision and image processing, aimed at extracting the semantic object area in the image, providing simple, effective content information for the computer, is a key to solving many high-level visual tasks such as image analysis and content understanding. step. The researchers have studied a lot of research in image segmentation, and many methods are proposed for different fields. However, these methods usually solve some specific problems, only for partial images, and therefore, the existing single segmentation method cannot be applied to all image implementations. Accurate segmentation.[0003]The existing image segmentation quality evaluation method is optimized by calculating the average e...

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): G06K9/62G06T7/11G06T7/194
CPCG06T7/11G06T7/194G06F18/214
Inventor 谭凯罗旺俞弦姚一杨王小康
Owner NARI INFORMATION & COMM TECH
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