Efficient method for moving object detection in cluttered background using Gaussian Mixture Model

Abstract

Foreground object detection in video is a f undamental step for automated video surveillance system and man y computer vision applications. Mostl y moving foreground object is detected b y background subtraction techniques. In d y namic background, Gaussian Mixture Model (GMM) performs better for object detection. In this work, a GMM based Basic Background Subtraction (BBS) model is used for background modeling. The connected component and blob labeling has been used to improve the model with a threshold. Morphological operators are used to improve the foreground information with a suitable structure element. The experimental stud y shows that the proposed work performs better in comparison to considered state-of-the-art methods in term of error.

Moving Object Detection for Visual Surveillance Using Quasi-euclidian Distance

“…This model presents an effective scheme that improves convergence rate and accuracy. Haque et al [6] presents a improved versions of GMM-based method [5,9]. This model considers a fixed threshold that is independent of learning rate and overcomes the drawback of [5].…”

“…This model classifies the pixel as background using values of mixture components, i.e., larger components can be categorized as part of background and remaining as foreground. Lee [9] proposed a BG subtraction-based technique using adaptive GMM. This model presents an effective scheme that improves convergence rate and accuracy.…”

Histogram-based adaptive learning for background modelling: moving object detection in video surveillance

“…In literature, various methods and algorithms have been proposed using the background subtraction technique. Among these, Gaussian mixture model (GMM)-based probabilistic model proposed by Stauffer and Grimson (1999) for motion-based object detection and extended by Lee (2005), Yadav (2014) and Haque et al (2008) these models consists of background modelling and distribution of pixel values over time at each pixel with weighted mixture of Gaussian's. In visual background extractor (ViBe) ( Barnich and Van Droogenbroeck, 2011), parameters of GMM algorithm is determined automatically using particle swarm optimisation.…”

“…In various exiting methods, the initial learning rate (Stauffer and Grimson, 1999;Lee, 2005;Yadav, 2014;Haque et al, 2008) α = 0.01. Similarly, the proposed work also uses same initial value of learning parameter.…”

“…In this segment, the performance is evaluated quantitatively with the help of some parameters, let FP: false positive, FN: false negative. In literature, various metrics (Yadav, 2014;Yadav et al, 2014) are applied to evaluate the performance of various methods. In this work, the following metrics are used for analysis:…”