Efficient scheme of Hough transform computing for videosequence compression

With the rise of high-speed communication channels a need for new, more efficient video stream compression algorithms has appeared. These algorithms involve both image filtering and segmentation. These tasks can be efficiently solved using the Hough transform. Preliminary sample smoothing is needed for Hough transform stability. Generally, the smoothing level is not given in advance. So, a parameter fitting scheme should be used. However, that leads to multiple calculations of the Hough transform. We show that a smoothing transform in the Hough space exists, which is equivalent to preliminary sample smoothing. Such a scheme is much more efficient.