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

Image saliency detection method and terminal based on multi-feature diffusion

A detection method and multi-feature technology, applied in the field of image processing, can solve the problems that the background area cannot be well suppressed and the detection of salient objects is not prominent, so as to achieve the effect of improving the accuracy

Active Publication Date: 2018-07-24
FUJIAN NORMAL UNIV
View PDF6 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is to provide an image saliency detection method and terminal based on multi-feature diffusion, which can solve the problems in the prior art that the detection of salient objects is not prominent and the background area cannot be well suppressed

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 saliency detection method and terminal based on multi-feature diffusion
  • Image saliency detection method and terminal based on multi-feature diffusion

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0095] The invention provides an image saliency detection method based on multi-feature diffusion, comprising the following steps:

[0096] S1: dividing the image into several superpixels by SLIC (simple linear iterative cluster) algorithm, and constructing a first graph structure according to the several superpixels;

[0097] The step S1 is specifically:

[0098] S101: Segment the original image of the SLIC algorithm into N superpixels;

[0099] Step S101 is specifically the process of converting the color image into a 5-dimensional feature vector in CIELAB color space and XY coordinates by using the SLIC algorithm, then constructing a distance metric for the 5-dimensional feature vector, and performing local clustering of image pixels. The SLIC algorithm can generate compact and approximately uniform superpixels, and has a high comprehensive evaluation in terms of computing speed, object contour preservation, and superpixel shape, which is more in line with the expected seg...

Embodiment 2

[0162] This embodiment is a specific application scenario of the first embodiment above.

[0163] First, when constructing the graph, a two-layer connection graph is used, and the edges of the graph are determined by the difference of features. When selecting background nodes, set the parameter c=3, and select nodes whose significance value is less than 0.5 as background nodes, and connect them to each other. Through the connected graph, calculate the inverse matrix of Laplacian that discards the eigenvectors with less information And the seed nodes obtained through high-level priori constitute a new diffusion method. Then, the saliency value of the image obtained by using the lab color space feature of the image as the bottom layer feature is used as the middle layer feature of the image again to construct a diffusion matrix, and the middle layer saliency map is obtained through the thirteenth formula. Similarly, the high-level prior composed of the background prior and th...

Embodiment 3

[0164] Please refer to figure 2 , Embodiment three of the present invention is:

[0165] The present invention provides an image saliency detection terminal based on multi-feature diffusion, which includes a memory, a processor, and a computer program stored on the memory and operable on the processor, and the processor implements the following steps when executing the program :

[0166] S1: dividing the image into several superpixels by SLIC (simple linear iterative cluster) algorithm, and constructing a first graph structure according to the several superpixels;

[0167] The step S1 is specifically:

[0168] S101: Segment the original image of the SLIC algorithm into N superpixels;

[0169] Step S101 is specifically the process of converting the color image into a 5-dimensional feature vector in CIELAB color space and XY coordinates by using the SLIC algorithm, then constructing a distance metric for the 5-dimensional feature vector, and performing local clustering of im...

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 saliency detection method and terminal based on multi-feature diffusion. When a seed node is selected, not all nodes at the edge of an image are defaulted to backgroundnodes. On the whole, three priori algorithms of background priori, color prior and position prior are fused as the high-level prior features of the image to select the seed node, so that not all saliency objects at the edge are detected as background. Finally, the multi-layer features of the image are extracted to construct different diffusion maps and diffusion matrices. The similarity of nodesis reflected from multiple angles, and corresponding saliency maps acquired from the middle and high-level features of the image are nonlinearly fused to acquire the final saliency map. Compared withsome classical saliency object detection algorithms in the prior art, the method and terminal, which are provided by the invention, have the advantages that the accuracy of image saliency detection isimproved in the aspect of a common data set, and the problems that saliency object detection is not prominent and the background area cannot be well restrained in the prior art are solved.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to an image saliency detection method and terminal based on multi-feature diffusion. Background technique [0002] Visual saliency refers to the part of the information in the image that can most attract human visual attention. Its goal is to identify the most visually prominent object or area in the image. The detection results are mainly manifested in the foreground area and background of the binary segmented image. area. It can effectively extract the foreground object of the image and reduce the complexity of scene analysis. Due to the limitation of computing resources, the human visual system can quickly and effectively locate the most interesting region from the currently seen picture, and prepare for further processing. Similarly, in order to improve computing efficiency, when performing image retrieval, target detection, image transmission, etc., some impo...

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): G06T7/00G06T7/12G06T7/90G06K9/46
CPCG06T7/0002G06T7/12G06T7/90G06T2207/10024G06V10/462
Inventor 叶锋洪斯婷陈家祯郑子华许力林晖李婉茹
Owner FUJIAN NORMAL 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