VIBE的几篇文章B哥的背景建模运动目标检测大作
5星 · 超过95%的资源 需积分: 9 199 浏览量
2012-11-29
11:37:17
上传
评论 2
收藏 750KB PDF 举报 
VIBE: A POWERFUL RANDOM TECHNIQUE TO ESTIMATE THE BACKGROUND IN
VIDEO SEQUENCES
Olivier Barnich and Marc Van Droogenbroeck
University of Liège
Montefiore Institute, INTELSIG Group
Liège, Belgium
ABSTRACT
Background subtraction is a crucial step in many automatic
video content analysis applications. While numerous accept-
able techniques have been proposed so far for background
extraction, there is still a need to produce more efficient al-
gorithms in terms of adaptability to multiple environments,
noise resilience, and computation efficiency. In this paper, we
present a powerful method for background extraction that im-
proves in accuracy and reduces the computational load. The
main innovation concerns the use of a random policy to select
values to build a samples-based estimation of the background.
To our knowledge, it is the first time that a random aggrega-
tion is used in the field of background extraction. In addition
we propose a novel policy that propagates information be-
tween neighboring pixels of an image. Experiment detailed
in this paper show how our method improves on other widely
used techniques, and how it outperforms these techniques for
noisy images.
Index Terms— Surveillance, Pattern recognition, Signal
analysis, Video signal processing
1. INTRODUCTION
Background subtraction is one of the most widely used tool
in automatic video content analysis, especially in video-
surveillance. Numerous methods for background subtraction
techniques have been proposed over the years (see [1, 2] for
surveys). In most of them, a model of the recent history is
built for each pixel location. The classification of new pixel
values is achieved by comparing each of them to the corre-
sponding pixel models. These techniques can be divided in
two categories: (i) parametric techniques that use a paramet-
ric model for each pixel location and (ii) samples-based tech-
niques that build their model by aggregating previously ob-
served values for each pixel location.
The Gaussian Mixture Model [3] is probably the most pop-
ular parametric technique. It is adaptive and able to deal with
the multi-modal appearance of the background of a dynamic
The first author has a grant funded by the FRIA (Belgium).
environment (changing time of day, clouds, tree leafs,...).
However since its sensitivity cannot be accurately tuned,
its ability to successfully handle high- and low-frequency
changes in the background is debatable, as detailed in [4].
Furthermore the estimation of the parameters of the model
(especially the variance) can become problematic for noisy
images.
Samples-based techniques [4, 5, 6] circumvent a part of the
parameters estimation step by building their models from ob-
served pixel values. This enhances their robustness to noise.
They provide fast responses to high-frequency events in the
background by directly including newly observed values in
their pixel models. However, their ability to successfully han-
dle concomitant events evolving at various speeds is limited
since they update their pixel models in a first-in first-out man-
ner. As a matter of fact, some of them use two sub-models for
each pixel [4, 5]: a short term model and a long term model.
While this can be a convenient solution, it is artificial and re-
quires fine tuning to work properly in for any given situation.
This paper presents a samples-based algorithm for back-
ground subtraction. The first contribution is a novel a random
selection policy that ensures a smooth exponentially decaying
lifespan for the sample values that constitute the pixel models.
It makes it able to successfully deal with concomitant events
with a single model of a reasonable size for each pixel. The
second contribution is related to the post-processing on which
all the abovementioned methods rely to give some degree of
spatial consistency to their results. For that purpose, we use
an innovative, fast, and simple spatial information propaga-
tion method that randomly diffuses pixel values across neigh-
boring pixels. Accordingly, our method is able to produce
spatially coherent results directly. As a third contribution, we
provide and instantaneous initialization technique that makes
our algorithm usable starting from the second frame of a se-
quence.
Section 2 describes our new background subtraction algo-
rithm. Experimental results are detailed in Section 3. Sec-
tion 4 concludes the paper.
945978-1-4244-2354-5/09/$25.00 ©2009 IEEE ICASSP 2009
夏天7788
- 粉丝: 541
- 资源: 1
最新资源
- MyBatis动态SQL是一种强大的特性，它允许我们在SQL语句中根据条件动态地添加或删除某些部分，从而实现更加灵活和高效的数据
- MyBatis动态SQL是一种强大的特性，它允许我们在SQL语句中根据条件动态地添加或删除某些部分，从而实现更加灵活和高效的数据
- 【ERP标准流程-标准流程-进货管理】（DOC 17页）.doc
- MyBatis动态SQL是一种强大的特性，它允许我们在SQL语句中根据条件动态地添加或删除某些部分，从而实现更加灵活和高效的数据
- MyBatis动态SQL是一种强大的特性，它允许我们在SQL语句中根据条件动态地添加或删除某些部分，从而实现更加灵活和高效的数据
- Java爬虫项目【项目开发计划】（共12页）.docx
- 11111111111
- MyBatis动态SQL是一种强大的特性，它允许我们在SQL语句中根据条件动态地添加或删除某些部分，从而实现更加灵活和高效的数据
- MyBatis动态SQL是一种强大的特性，它允许我们在SQL语句中根据条件动态地添加或删除某些部分，从而实现更加灵活和高效的数据
- MyBatis动态SQL是一种强大的特性，它允许我们在SQL语句中根据条件动态地添加或删除某些部分，从而实现更加灵活和高效的数据
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
- 1
- 2
- 3
前往页