Real Time Foreground-Background Segmentation Using a Modified Codebook Model

Abstract

Real time segmentation of scene into objects and background is really important and represents an initial step of object tracking. Starting from the codebook method [4] we propose some modifications which show significant improvements in most of the normal and also difficult conditions. We include parameter of frequency for accessing, deleting, matching and adding codewords in codebook or to move cache codewords into codebook. We also propose an evaluation method in order to objectively compare several segmentation techniques, based on receiver operating characteristic (ROC) analysis and on precision and recall method. We propose to summarize the quality factor of a method by a single value based on a weighted Euclidean distance or on a harmonic mean between two related characteristics.

A hybrid codebook background model for background subtraction

“…In [7], they proposed a modified CB method by adding the parameters of frequency for CB and for cache CB. Similarly, Yao et al [8] introduced a multi-layer ground subtraction method by using the local binary pattern (LBP) and color to model the background pixels. However, the major drawback for background models is the influence of the shadow and highlight.…”

Computer Vision Techniques for Background Modeling in Urban Traffic Monitoring

“…Other measures for fitness quantification, in the context of background subtraction techniques, have been proposed in the literature (Rosin & Ioannidis, 2003;White & Shah, 2007;Ilyas et al, 2009). The following are some examples:…”

“…The lower limit of this interval occurs when there are not matching pixels, while a perfect match would make the coefficient to hit the upper bound. Finally, Ilyas et al (2009) proposes a weighted Euclidean distance, considering the deviations of FPR and TPR from their respective ideal values, 0 and 1. It is defined as follows:…”

Cascade Method for Image Processing Based People Detection and Counting

“…In this approach, the background subtraction method is applied to segment the foreground blobs [22], [23] and ellipse detection is applied to identify the number of people in each blob. Other shapes such as Bernoulli shapes have been used by some researchers to count the number of people [24].…”