Real Time Illumination Invariant Background Subtraction Using Local Kernel Histograms

Abstract

Constant background hypothesis for background subtraction algorithms is often not applicable in real environments because of shadows, reflections, or small moving objects in the background: flickering screens in indoor scenes, or waving vegetation in outdoor ones. In both indoor and outdoor scenes, the use of color cues for background segmentation is limited by illumination variations when lights are switched or weather changes. This problem can be partially allievated using robust color coordinates or background update algorithms but an important part of the color information is lost by the former solution and the latter is often too specialized to cope with most of real environment constraints. This paper presents an approach using local kernel histograms and contour-based features. Local kernel histograms have the conventional histograms advantages avoiding their inherent drawbacks. Contour based features are more robust than color features regarding scene illumination variations. The proposed algorithm performances are emphasized in the experimental results using test scenes involving strong illumination variations and non static backgrounds.

Bayesian Formulation of Image Patch Matching Using Cross-correlation

“…However, this scene falls outside the scope of the considered algorithm which is targeted for outdoor traffic scenes. If the LightSwitch sequence is excluded this gives on average 6.78% misclassified pixels, which is slightly better than the results presented in [15], [12], 7.82% and 7.33% misclassified pixels respectively. The proposed algorithm is also significantly faster.…”

“…The proposed algorithm were also tested on the WallFlower dataset from [15] available online 1 , which is also used by [12]. This dataset consists of 7 sequences with resolution 160x120.…”

Global Illumination Compensation for Background Subtraction Using Gaussian-Based Background Difference Modeling

“…Furthermore, gradient in-formation used for pixel-based background subtraction is highly unreliable under camera motion (shaking). Even more advanced features like histograms of gradients [10], which are supposed to be robust against small camera movements, appeared to be very sensitive to vibrations of the camera in our experiments using i-LIDS parking vehicle sequences [6]. Seki et al [12] model the co-occurrence of variations between adjacent blocks, thereby achieving accurate segmentation under illumination variation and swaying vegetation, but this proposal seems compute intensive.…”

Background Subtraction With Adaptive Spatio-Temporal Neighborhood Analysis

“…In [13], pixel value probability densities, represented as normalized histograms, are accumulated over time, and BG label are assigned by the Maximum A Posteriori criterion. Region-based algorithms usually divide the frames into blocks and calculate block-specific features; change detection is then achieved via block matching, considering for example fusion of edge and intensity information [7]. In [8] a region model describing local texture characteristic is presented:the method is prone to errors when shadows and sudden global changes of illumination occur.…”