Motion Segmentation Using Robust Statistics and Spatial Continuity

This paper presents a robust method for segmenting images based on the motion of different scene objects. The presented algorithm combines a robust estimator named Selective Statistical Estimator (SSE) with a median filtering technique to enforce the spatial continuity into a robust estimator. The algorithm is tested on a number of real image sequences contain multiple moving objects. The proposed method is also shown to be effective for human surveillance.