Wavelet Multi-Scale Transform Based Foreground Segmentation and Shadow Elimination

Abstract

An algorithm using wavelet multi-scale transform for segmenting foreground moving objects and suppressing shadow is proposed. The optimal selection of threshold is automatically determined which does not require any complex supervised training, manual calibration or hypothesis. The proposed algorithm is efficient enough to segment foreground moving objects with low contrast against the background. The reference image is used to extract foreground no matter the objects enter the field of view before captured or not. The developed method is highly computationally cost-effective since it does not concern with complex computation model, color model or background statistics at a time. By compari- sons, it has been shown that the proposed approach is more robust and efficient to detect foreground and suppress shadow during coping with different indoors or outdoors circumstances.