Shadow detection and elimination using geometric approach for static images

The approach presents the coarse approach to detect and eliminate unexpected shadow of single and multiple objects from the static background. The initial stage consists of moment based method to estimate the boundaries between shadows and static objects. The selected shadow is parameterized using features like orientation, mean and the center position of the shadow. Using these features model can detect and eliminate the shadow from the background scene. It improves the quality of detection for analysis. Experiments demonstrate the single object and multiple objects with their respective shadow can be successfully detected and eliminated from the scene background.