Multi-target image joint segmentation method based on multi-tag multi-sample learning

A multi-instance learning and multi-label technology, applied in the field of image processing, can solve the problems of inability to obtain consistent information, over-segmentation of segmentation results, and low segmentation accuracy, so as to overcome ambiguity and uncertainty, improve segmentation accuracy, and improve segmentation accuracy. The effect of precision

Active Publication Date: 2017-03-15
NANJING UNIV
View PDF8 Cites 19 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] (1) Since the existing methods are often unable to obtain accurate consistency information, resulting in the phenomenon of over-segmentation in the segmentation results;
[0007] (2) Since the existing methods mainly use unsupervised methods to use consistent information to guide the final segmentation, resulting in low segmentation accuracy

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
  • Multi-target image joint segmentation method based on multi-tag multi-sample learning
  • Multi-target image joint segmentation method based on multi-tag multi-sample learning
  • Multi-target image joint segmentation method based on multi-tag multi-sample learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0086] In this example, if figure 2 Shown is the input image, through the multi-target image joint segmentation method based on multi-label multi-instance learning described in the present invention, the figure 2 Each image is divided into image 3 The different common objects described in are identified by different colors, and the marks 1 to 3 represent three types of common objects respectively. In the figure, the number 1 represents the girl in the blue dress, the number 2 represents the girl in the red dress, and the number 3 represents the apple and the basket. The specific implementation process is as follows:

[0087] In step 1, object detection is used to obtain candidate object sets of the input image set, and the candidate object sets are clustered to obtain corresponding object labels. As shown in Figure 4 figure 2 In the second image, the result of the object detection of the single image in step 1.1., wherein Figure 4a is the object area obtained by step...

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 multi-target joint segmentation method based on multi-tag multi-sample learning, comprising the following steps: first, performing object detection on an input image set to get a candidate object set, and performing segmented object clustering on the candidate object set to get accurate object tags; the, performing salience detection and binary segmentation on the input image set to get a salient region, transferring the object tags in the candidate object set to the salient region, and performing super pixel segmentation on the salient region to get a salient region containing tags; and finally, using an integrated multi-sample multi-tag learning method based on feature random selection to get the object tag of each super pixel in the salient region, thus completing joint segmentation of the input image set.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to a multi-target joint segmentation processing method for digital images. Background technique [0002] In the field of digital image processing, image segmentation is a fundamental problem. Image segmentation has a great influence on many image processing problems such as image retrieval, image editing and image classification. Traditional image segmentation methods, such as document 1: Wen Ying. A robust FCM image segmentation method. Chinese patent: CN105654453A, 2016, document 2: Hu Haifeng. Image segmentation method based on visual saliency model. Chinese patent: CN105678797A, 2016, etc., are all aimed at the segmentation of a single image. This method is either difficult to accurately segment the image, or requires a lot of manual interaction, and is not suitable for the segmentation of large-scale image sets. [0003] In order to solve these problems, joint segment...

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): G06T7/11
CPCG06T2207/20081
Inventor 孙正兴杨炜辰李博胡佳高
Owner NANJING UNIV
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