BMOG: Boosted Gaussian Mixture Model with Controlled Complexity

Abstract

Developing robust and universal methods for unsupervised segmentation of moving objects in video sequences has proved to be a hard and challenging task. The best solutions are, in general, computationally heavy preventing their use in real-time applications. This research addresses this problem by proposing a robust and computationally efficient method, BMOG, that significantly boosts the performance of the widely used MOG2 method. The complexity of BMOG is kept low, proving its suitability for real-time applications. The proposed solution explores a novel classification mechanism that combines color space discrimination capabilities with hysteresis and a dynamic learning rate for background model update.

Background subtraction in dynamic scenes using the dynamic principal component analysis

“…The proposed algorithm is compared with four state‐of‐the‐art methods: WeSamBE [36], boosted Gaussian mixture model (BMOG) [58], auto‐adaptive parallel architecture (AAPSA) [59] and co‐occurrence probability‐based pixels pairs (CP3) [60]. Tables 2–4 report the values of the recall, F‐measure and precision of these methods.…”

Statistical accuracy analysis of different detecting algorithms for surveillance system in smart city

“…In the recent years, several algorithms are developed for video surveillance recognition. For instance, Sehairi et al [8] proposed a simplified self-organized Background Subtraction algorithm, while Martins et al [9] used different methods to explain the difficulty and trouble of Background Subtraction algorithm that based on the Background model, known as Mixture of Gaussians (MoG). Henry et al [10] introduced a multi-object detection and tracking model based on stereo vision for use in surveillance systems.…”

Background subtraction for night videos

“…To improve the performance, researchers also propose to use descriptors. Some researchers transform the pixels from RGB space into other color spaces to separate color intensity from other color information ( Balcilar, Amasyali & Sonmez, 2014 ; Martins et al, 2017 ). Another effective descriptor for background subtraction is texture-based local binary pattern (LBP) ( Heikkila & Pietikainen, 2006 ).…”