Non-parametric Model for Background Subtraction

By Ahmed M. Elgammal, David Harwood, Larry S. Davis. European Conference on Computer Vision, 2000

2,516 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/40b55a07d841aae6246775d3e82a3109c3a75555

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Analysis of Rain and Snow in Frequency Space (2008, methodology): “In addition to layer extraction, techniques such as Gaussian mixture models (Stauffer and Grimson 1998), kernel density estimation (Elgammal et al. 2000), or robust PCA (de la Torre and Black 2001) could model the variation caused by rain and snow more effectively.”
Animal Detection From Highly Cluttered Natural Scenes Using Spatiotemporal Object Region Proposals and Patch Verification (2016, methodology): “[16] used a nonparametric background model to achieve better accuracy under the same constraints as the mixture of Gaussians.”
Detecting Moving Shadows: Algorithms and Evaluation (2003, methodology): “In the statistical methods (see [8], [9], [10], [11], [12]), the parameter selection is a critical issue.”
Deep Neural Network Concepts for Background Subtraction: A Systematic Review and Comparative Evaluation (2018, methodology): “n in the Hamming space to perform foreground detection. Experimental results [222] on the CDnet 2012 dataset [62] demonstrate that SDAE gives better performancethan traditionalmethods (MOG [180], KDE [49], LBP [77]), and recentstate-of-artmodel(PBAS[81]).Toaddressrobustnessagainststationarynoise, Ga”
SuBSENSE: A Universal Change Detection Method With Local Adaptive Sensitivity (2015, methodology): “Nonparametric models based on Kernel Density Estimation (KDE) [14] were also introduced early on and improved in more recent state-of-the-art methods [11], [15], [16].”
A survey on activity recognition and behavior understanding in video surveillance (2013, background): “[42] kernel density estimation i g – s little/background”
Compressive sensing: From theory to applications, a survey (2013, methodology): “Data captured by single pixel camera can also be used for background subtraction for automatic detection and tracking of objects [57], [58].”
Decomposition into Low-rank plus Additive Matrices for Background/Foreground Separation: A Review for a Comparative Evaluation with a Large-Scale Dataset (2015, methodology): “Moreover, we believe that we are living in a key transition in the field of background subtraction as we are progressively migrating from the conventional statistical models as MOG [357][33], KDE [95][30] and naive subspace learning models [29] to models based on robust decomposition into low-rank p”
Motion segmentation and pose recognition with motion history gradients (2000, methodology): “Although there is recent work on more sophisticated methods of background subtraction [[14][18][21]], we use a simplistic method here.”
ViBE: A powerful random technique to estimate the background in video sequences (2009, methodology): “Index Terms— Surveillance, Pattern recognition, Signal analysis, Video signal processing”
Machine Recognition of Human Activities: A Survey (2008, background): “scenarios with changing background [21].”
Improving shadow suppression in moving object detection with HSV color information (2001, methodology): “This is similar to what is called selective update in [5], although in that work, single moving points are used without the further validation to belong to moving objects.”
Traditional and recent approaches in background modeling for foreground detection: An overview (2014, background): “For example, Figure 2 shows the original frame 309 of the sequence from [84], the generated background, the ground-truth and the foreground mask.”
An Improved Adaptive Background Mixture Model for Real-time Tracking with Shadow Detection (2001, methodology): “Elgammal et al used a kernel estimator for each pixel [6].”
Statistical modeling of complex backgrounds for foreground object detection (2004, methodology): “Many existing methods utilize spectral features (distributions of intensities or colors at each pixel) to model the background [4], [5], [7]–[9].”
Maximum Margin Criterion (2018, methodology): “The KDE was implemented in Microsoft Visual C++ and the IRT and IMMC was implemented in Matlab.”
Segmenting foreground objects from a dynamic textured background via a robust Kalman filter (2003, methodology): “In [4, 17] the color distribution of each background pixel is learned using a Gaussian mixture model, and these mixture models are continuously updated to adapt to slow background changes.”