Three Novell Analog-Domain Algorithms for Motion Detection in Video Surveillance

Expert video-surveillance system for real-time detection of suspicious behaviors in shopping malls

“…There are several background subtraction methods commonly employed in automated video-surveillance applications, as is exposed in Stefano et al (2011);Brutzer et al (2011), and their results have been greatly improved in the last years using varied techniques such as the formulated by Maddalena & Petrosino (2008); Reddy et al (2011);Verdant et al (2011). On the other hand, some classic segmentation algorithms continue being widely applied, as the ones derived from Gaussian mixture models (Grimson & Stauffer, 1999;Zivkovic, 2004;Baf et al, 2008a), fuzzy models (Zhang & Xu, 2006;Baf et al, 2008b) or a combination of both (Baf et al, 2008c).…”

A 120μW 240×110@25fps vision chip with ROI detection SIMD processing unit

“…Indeed their instruction set allows implementing various complex algorithms (e.g. enhanced versions of ΣΔ or Recursive Average [7]). The computations are performed on the local average of 10×10 pixels generated from 10×10 connected photodiodes spread among the considered macropixels.…”

“…The circuit has been optimized to implement various motion detection algorithms, from which derives the locations of ROIs. Those algorithms based on the difference between the actual scene and an estimate of its background, obtained from the temporal filtering of the macropixels' luminance, are among the most compact ones [7].…”

Background initialization and foreground segmentation for bootstrapping video sequences

“…On the other hand, within some existing approaches [30][31][32][33][34], temporal smoothing is not adopted in background updating. Chein et al [30] proposed a pixel-wise video segmentation approach with adaptive thresholding to determine each pixel as a moving or stationary one.…”

“…That is, this type of approaches might not work well in illumination-changing environments. Verdant et al [31] proposed three analog-domain motion detection algorithms in video surveillance, namely, the scene-based adaptive algorithm, the recursive average with estimator algorithm, and the adaptive wrapping thresholding algorithm, in which background estimation and variance of each pixel are computed with nonlinear operations to perform adaptive local thresholding. Lin et al [32] used a classifier to determine whether an image block belongs to the background for block-wise background updating.…”