Improvements and Experiments of a Compact Statistical Background Model

Abstract

Change detection plays an important role in most videobased applications. The first stage is to build appropriate background model, which is now becoming increasingly complex as more sophisticated statistical approaches are introduced to cover challenging situations and provide reliable detection. This paper reports a simple and intuitive statistical model based on deeper learning spatial correlation among pixels: For each observed pixel, we select a group of supporting pixels with high correlation, and then use a single Gaussian to model the intensity deviations between the observed pixel and the supporting ones. In addition, a multi-channel model updating is integrated on-line and a temporal intensity constraint for each pixel is defined. Although this method is mainly designed for coping with sudden illumination changes, experimental results using all the video sequences provided on changedetection.net validate it is comparable with other recent methods under various situations.

Comparative study of motion detection methods for video surveillance systems

“…Bayesian decision 89 Subspace learning-principle component analysis(Eigen-Backgrounds) 42 Linear predictive filter 86 Wallflower method 86 Picardi 4 Running Gaussian average 14 Temporal median filter 90,91 Mixture of Gaussians 15 Kernel density estimation (KDE) 16 Sequential kernel density approximation 92 Subspace learning-principle component analysis(Eigen-Backgrounds) 85 Co-occurrence of image variations 93 Bayesian multi-layer 36 Histogram over time 13 Local-self similarity 110 Bianco et al 130 IUTIS-1 130 IUTIS-2 130 IUTIS-3 130 Flux tensor with split Gaussian models (FTSG) 124 Self-balanced local sensitivity (SuBSENSE) 123 Weightless neural networks (CwisarDH) 51,122 Spectral-360 121 fast self-tuning BS 120 K-nearest neighbor method (KNN) 119 Kernel density estimation(KDE) 16 Spatially coherent self-organized background subtraction (SC-SOBS) 22 Euclidean distance 97 Mahalanobis distance 97 Multiscale spatio-temp BG model 117 CP3-online 118 Mixture of Gaussians (MoG) 15 GMM Zivkovic 60 Fuzzy spatial coherence-based SOBS 65 Region-based mixture of Gaussians (RMoG) 131 CDnet 2014 dataset 125 Recall Specificity False Positive Rate False Negative Rate Percentage of wrong classifications Precision F-measure…”