Unlock instant, AI-driven research and patent intelligence for your innovation.

A Foreground Target Extraction Method Based on Pixel Genetics

A target extraction and foreground technology, applied in the field of foreground target extraction based on pixel genetics, which can solve problems such as accuracy and precision defects

Inactive Publication Date: 2019-12-06
EAST CHINA UNIV OF SCI & TECH
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Among such methods of fully automatic foreground extraction, Sungheum proposed a method of automatically extracting foreground objects using multi-view technology in 2011; Zhang Yupeng et al. proposed an automatic foreground and background segmentation technology based on the coded aperture depth in 2012; Xie Changting et al. proposed an automatic trimap generation technology in 2013; Yuan Yubo et al. proposed a foreground extraction technology based on Gaussian variation in 2014; although the above method has achieved good results, there are varying degrees of accuracy and precision. defect

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
  • A Foreground Target Extraction Method Based on Pixel Genetics
  • A Foreground Target Extraction Method Based on Pixel Genetics
  • A Foreground Target Extraction Method Based on Pixel Genetics

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0032] Step 1, use the Ncut technology and the Gaussian variation model to obtain the key foreground area of ​​the image.

[0033] Step 2, use the Canny operator to extract the edge information from the key foreground area obtained in step 1 and the original image respectively.

[0034] Step 3, take the original image information e obtained in step 2 I and the edge information e of the key foreground of the image F the intersection of e s , as shown in the following formula (1); use the connectivity of the image edge and the color similarity ρ between pixels to obtain a relatively complete edge information b(x, y) of the image foreground, as shown in the following formula (2) , where the initial value of b(x,y) is e s (x,y), ρ((x,y),(x 0 ,y 0 )) represents the similarity of adjacent pixels in the image, as shown in the following formula (3), δ is the threshold, after a lot of experiments, we choose 5 in the experiment. Define an initial profile C(x,y)=b(x,y).

[0035] e...

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

This patent proposes a new method for extracting foreground objects from digital images based on pixel inheritance. The method includes the following steps: (1) using Ncut technology and Gaussian variation model to extract the key foreground from the image to form the initial population; (2) using the edge detection method to extract the edge of the original image and the key foreground, take out the common part, Use connectivity and pixel similarity to determine the genetic boundary (target contour); (3) construct an adaptive function, select pixels with strong adaptability as parents, and reproduce new foreground targets in the image through genetic processes such as crossover and mutation. The whole process of the method is fully automatic without any human interaction. A large number of experiments have proved that this method is very effective and can get more accurate experimental results.

Description

technical field [0001] The invention mainly relates to image processing technology, in particular to a foreground object extraction method based on pixel inheritance. Background technique [0002] Digital image foreground target extraction is an important basic problem in computer vision, and its task is to separate the target of interest from the background from a single image or a sequence of images for subsequent processing. The effective extraction of the foreground area is very important for medium and high-level tasks such as target classification, identity recognition, and behavior understanding, because the subsequent processing usually only considers the pixels corresponding to the foreground area in the image, and will greatly promote video coding, retrieval, and so on. , human-computer interaction, motion capture and other topics. However, due to many uncontrollable factors in the actual acquisition environment, such as the influence of camera movement, illuminat...

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 Patents(China)
IPC IPC(8): G06K9/46
CPCG06V10/44
Inventor 袁玉波陈凯星刘赟王金海赵婷婷溥冬梅张静陈志华
Owner EAST CHINA UNIV OF SCI & TECH