Real-Time Image-Based Motion Detection Using Color and Structure

In this paper we propose a method for automating the process of detecting regions of motion in a video sequence in real time. The main idea of this work is to detect motion based on both structure and color. The detection using structure is carried out with the aid of information gathered from the Census Transform computed on gradient images based on Sobel operators. The Census Transform characterizes local intensity patterns in an image region. Color-based detection is done using color histograms, which allow efficient characterization without prior assumptions about color distribution in the scene. The probabilities obtained from the gradient-based Census Transform and from Color Histograms are combined in a robust way to detect the zones of active motion. Experimental results demonstrate the effectiveness of our approach.