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

A Video Segmentation Method Based on Object Multipart Learning

A video segmentation, multi-component technology, applied in the field of image processing, can solve problems such as inaccurate segmentation results

Active Publication Date: 2017-08-25
INST OF AUTOMATION CHINESE ACAD OF SCI
View PDF9 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In existing methods, the algorithm generally builds a global representation model without local constraints, which may lead to inaccurate segmentation results, especially in complex backgrounds or scenes with severe motion.

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 Video Segmentation Method Based on Object Multipart Learning
  • A Video Segmentation Method Based on Object Multipart Learning
  • A Video Segmentation Method Based on Object Multipart Learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0036] The following describes the present invention in detail through specific examples, which are not intended to limit the present invention. The whole process of the present invention is as follows: firstly input the target frame to be segmented in the first frame, and then use the interactive segmentation method to extract the target to be segmented from the background. The initial target multi-part model is then generated using the SLIC algorithm. Order M 0 ={M 1 ,...,M k} is the k components of the target, and the corresponding label is {l 1 ,...,l k}, M i =(A i ,P i ,Θ i ) is the i-th model, where A i is the HSV histogram of the model, P i is the center position of the model, Θ i is the set of pixel positions belonging to the model. Simultaneously build the background model M 0 ={M 0,1 ,...,M 0,n} to describe the complex background, that is, the outliers relative to the foreground target, where the label is l 0 , n is the number of parts in the backgrou...

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 a video segmentation algorithm based on multiple-target-component learning. According to the algorithm, multi-component tracking and segmentation information is unified in one energy function. Online multiple-target-component tracking provides effective timing sequence motion and structure constraint information for segmentation, and meanwhile, multiple-target-component segmentation produces accurate local representation appearance and position information to improve the multiple-target-component tracking precision. Further, iterative optimization is performed on multiple-target-component tracking and multiple-target-component segmentation steps with an RANSAC-style algorithm, and accurate video segmentation results are acquired.

Description

technical field [0001] The invention relates to the fields of image processing, pattern recognition and computer vision, in particular to a video segmentation method based on multi-component joint segmentation and tracking. Background technique [0002] Recently, the field of visual tracking focuses on accurately segmenting the edge of the target from the background, which can have many subsequent high-level applications, such as behavior recognition, scene understanding, and depth and occlusion reasoning. However, it is still difficult to design a relatively robust video segmenter, because a variety of internal and external factors, such as the deformation of the target itself, the influence of the surrounding complex environment, occlusion, etc., will cause the segmenter to fail. Accurately segment the target. [0003] Generally speaking, most of the current video segmentation methods use offline batch processing to accurately extract the target edge from the background. ...

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): G06T7/215G06T7/194
CPCG06T7/215G06T7/251G06T2207/10016
Inventor 雷震文珑银李子青
Owner INST OF AUTOMATION CHINESE ACAD OF SCI