Detection of Change in Body Motion With Background Construction and Silhouette Orientation: Background Subtraction With GMM

Background subtraction techniques have been widely implemented and improvised to obtain a stable background model. The novelty of the proposed work is to generate a stable background model under dynamic changes in the environmental conditions where 1) an improved background subtraction algorithm is proposed based on GMM with EM algorithm for computing granulometry to run faster for the generation of a stable background model; 2) detecting the foreground by curvelet based denoising process with improvised semisoft thresholding techniques with morphological operations is done; 3) background maintenance is done by an adaptive algorithm in which the intensity values are mapped to remove the connected components with less than P pixels. The proposed scheme works for the spatio-temporal motion of the object in both spatial and temporal modes. The experimental outcome for the proposed model results in the accurate shape analysis of the object in motion thereby dipping the complexity.