توضیحات
ABSTRACT
Background modeling is a key step of background subtraction methods used in the context of static camera. The goal is to obtain a clean background and then detect moving objects by comparing it with the current frame. Mixture of Gaussians Model [1] is the most popular technique and presents some limitations when dynamic changes occur in the scene like camera jitter, illumination changes and movement in the background. Furthermore, the MGM is initialized using a training sequence which may be noisy and/or insufficient to model correctly the background. All these critical situations generate false classification in the foreground detection mask due to the related uncertainty. To take into account this uncertainty, we propose to use a Type-2 Fuzzy Mixture of Gaussians Model. Results show the relevance of the proposed approach in presence of camera jitter, waving trees and water rippling.
INTRODUCTION
The common approach for discriminating moving objects from the background is the background subtraction which is used in the field of video surveillance [2], optical motion capture [3–5] and multimedia applications [6]. In this context, background modeling is the first key step to obtain a clean background. The simplest way to model the background is to acquire a background image which doesn’t include any moving object. In some environments, the background isn’t available and can always be changed under critical situations like camera jitter, illumination changes, objects being introduced or removed from the scene. To take into account these problems of robustness and adaptation, many background modeling methods have been developed and the most recent surveys can be found in [2, 7, 8]. These background modeling methods can be classified in the following categories: Basic Background Modeling [9–11], Statistical Background Modeling [12, 1, 13], Fuzzy Background Modeling [14, 15] and Background Estimation [16–18]. The models the most used are the statistical ones: The first way to represent statistically the background is to assume that the history over time of intensity values of a pixel can be modeled by a single Gaussian [9]. However, a unimodal model cannot handle dynamic backgrounds when there are waving trees, water rippling or moving algae. To solve this problem, the Mixture of Gaussians Models (MGM) has been used to model dynamic backgrounds [1]. This model has some disadvantages. Background having fast variations cannot be accurately modeled with just a few Gaussians (usually 3 to 5), causing problems for sensitive detection. So, a non-parametric technique was developed for estimating background probabilities at each pixel from many recent samples over time using Kernel density estimation [13] but it is time consuming. Finally, due to a good compromise between robustness and time/memory requirements MGM are the most used.
Year: 2008
Publisher : Springer
By : Fida El Baf, Thierry Bouwmans, Bertrand Vachon
File Information: English Language/ 11 Page / size: 454 KB
Download: click
سال : 2008
ناشر : Springer
کاری از : Fida El Baf, Thierry Bouwmans, Bertrand Vachon
اطلاعات فایل : زبان انگلیسی / 11 صفحه / حجم : KB 454
لینک دانلود : روی همین لینک کلیک کنید
نقد و بررسیها
هنوز بررسیای ثبت نشده است.