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

An increment type characteristic background modeling algorithm of self-adapting weight selection

A background modeling and self-adaptive technology, applied in computing, computer components, character and pattern recognition, etc., can solve problems such as interpretation, algorithm accuracy is greatly affected, and weight calculation methods are not provided

Inactive Publication Date: 2008-07-09
ZHEJIANG UNIV
View PDF0 Cites 23 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, there are still two main problems in this feature background modeling method: first, this method needs to prepare a set of sample background images in advance, which images to choose as samples have a great impact on the accuracy of the algorithm, and the original feature background modeling The method does not discuss how to update the model quickly; secondly, because the characteristic root decomposition is carried out on the entire image, this leads to the "absorption" of large foreground moving objects into the background model, so that the ideal background image cannot be generated
However, the author did not explain these weights, did not give physical meaning to these weights, and did not provide a quantitative and reasonable weight calculation method
Obviously, the method of determining weights only by experience is very unstable, and it is difficult to establish an optimized background model and generate an ideal background image
In addition, the weight is applied to the entire image, without considering the different contributions of different regions of the image to the background model

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
  • An increment type characteristic background modeling algorithm of self-adapting weight selection
  • An increment type characteristic background modeling algorithm of self-adapting weight selection
  • An increment type characteristic background modeling algorithm of self-adapting weight selection

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0055] Example of background modeling for speedboats on the West Lake:

[0056] The first scene is the West Lake reflected in the setting sun, and a speedboat speeds across the lake in the distance. There are certain difficulties in motion detection in this scene: First, the foreground object we want to detect, that is, the speedboat, occupies too small an area in the entire picture, and the movement of the speedboat is easily absorbed by the background model as background noise; Secondly, the background changes in the video screen are more significant, including the leaves blown by the wind near the camera and the sunlight reflected by the lake waves. as shown in picture 2.

[0057] The motion detection results are shown in Figure 3. (a) and (b) are the background image and the detected foreground area produced by the traditional feature background modeling method, (c) and (d) are the background image and the foreground area obtained by the method of the present invention. ...

Embodiment 2

[0059] Example of background modeling for the moving human body by the West Lake:

[0060] The second scene is a person walking along the Su Causeway by the lake, and the background is the West Lake with wind and wave movement, as shown in Figure 4(a). The difficulty here is that the background lake has a large area, and the background wave motion is very complicated. At the same time, the nearby willow branches are dancing with the wind, which can easily be regarded as foreground objects. Based on this scene, we tested the traditional feature background modeling method and the method of the present invention respectively, and the generated background images are shown in Fig. 4(b) and Fig. 4(c) respectively. Figure 4(b) shows that there are obvious "ghosting" phenomena in the area where people pass by in the background image. This is because the traditional method treats the video frame as a whole without considering the motion of different areas in the image. Therefore, the ...

Embodiment 3

[0062] Example of background modeling for large-scale complex scenes:

[0063]The last scenario is challenging for all background modeling based motion detection methods. The test video was shot on the lawn of the Yuquan Campus of Zhejiang University. Someone ran quickly across the entire scene. The scene contained very complex background changes, such as irregularly moving crowds in the distance and bushes swaying in the wind. The traditional method and the method of the present invention are also tested based on this scenario, and Fig. 5 shows the effect of the algorithm. In Fig. 5, the first row is the 1373rd, 1410th, 1450th, 1634th and 1660th frame of the original video, the second row is the corresponding background image generated by the traditional method, and the third row is generated by the method of the present invention Corresponds to the background image. In the second row of images, the image area in the circle has a very obvious "ghosting" effect, which is als...

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 a self-adaptive weight selected incremental characteristic background modeling method. The invention can conduct incremental real-time updating of a background model according to the movement contained in each video frame in the movement detection, and appoint a weight for every frame in the updating process to improve the expression and description abilities of the background model. The method comprises the following steps of: roughly detecting the movement area of the present frame by an un-updated background model, constructing a weight for the movement area based on the reconstruction error of the model, updating the background model by right of the incremental principal component analysis based on the weighted present frame and generating background images. With well expressing the dynamic changes of the complex scenes and being sensitive to the objects with obvious movement prospect, the inventive technique has great application value in the fields of video surveillance, etc.

Description

technical field [0001] The invention relates to video motion detection, in particular to an incremental feature background modeling method for adaptively selecting weights. Background technique [0002] Motion detection and motion tracking are relatively low-level problems in the field of vision, and motion detection is the premise of tracking. As a class of methods to solve the problem of motion detection, background modeling has attracted the attention of many researchers in recent years. Some typical background modeling methods include: Gaussian model, mixed Gaussian model, kernel density estimation, and Eigen-background Modeling. The first three methods are pixel-based methods, that is, a separate model is established for each pixel in the time domain for description. The computational complexity of pixel-based background modeling is high, and it is not conducive to capturing complex background content, such as weather conditions such as rain and snow, and wind blowing...

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
IPC IPC(8): G06T7/00G06T7/20G06K9/00G06K9/46
Inventor 庄越挺张剑肖俊吴飞
Owner ZHEJIANG 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