With the demand for the content based image processing, the segmentation for semantic objects is required. One of the first process to be done is the foreground extraction. Since this algorithm is the pre-processing part, its computing load should low. However, in most algorithms, either motion compensation or the extra parameters required. The former results in high computing load while the later requires the knowledge that may be unavailable. We propose a novel algorithm for fast foreground extraction. Motion compensation is replaced by algorithm consisting of only logical AND and OR operators. Test of the proposed algorithm has shown Promising results.