An apparatus used in a motion-compensated video signal encoder for determining a predicted current frame based on a current frame and a previous frame of a digital video signal using a predetermined number of search grids, each search grid being a set of search blocks formed in the current frame, comprising: a motion estimator for detecting a set of motion vectors for each of the search grids, each of the motion vectors being determined by calculating an error between each of the search blocks and candidate blocks and selecting one candidate block which yields a minimum error; an error accumulator for accumulating each minimum error for all of the search blocks contained in each of the search grids to provide an accumulated minimum error signal for each of the search grids; a selector for comparing each of the search grids by using the accumulated error signal to provide a search grid selection signal; a switch, in response to the search grid selection signal, for selecting one set of motion vectors corresponding to the selected search grid; and a motion compensator for assigning the value of each of the pixels in the previous frame, which corresponds to one of the pixels in the current frame through the selected one set of motion vectors, as the value of said one of the pixels in the current frame.
A method, for use in a segmentation-based coding system, for detecting motion vectors between a current frame and a previous frame of digital video signal comprises the steps of (a) generating segmentation data for each of the segmented regions, the segmentation data including contour information representing the shape and location for each segmented region and texture information representing a mean gray level of all pixels contained in each segmented region, (b) generating candidate segmentation data for each of the candidate regions, (c) producing a minimum-sized predetermined geometric pattern for encompassing a segmented region, (d) producing a minimum-sized predetermined geometric pattern for encompassing each of the candidate regions, (e) calculating an error value between the segmented region and each of the candidate regions, (f) selecting a candidate region yielding a smallest error value as a search region, (g) detecting one or more motion vectors between the segmented region and the search region; and (h) repeating said steps (c) to (g) for all the segmented regions.
A prediction error generating unit generates a predictive vector from the motion vectors of a plurality of adjacent blocks, and obtains a difference from a target vector. A plurality of variable-length coding units respectively encode the output of the prediction error generating unit with different encoding methods. A determining unit estimates the accuracy of the predictive vector generated by the prediction error generating unit based on the degrees of non-uniformity of the motion vectors of the plurality of adjacent blocks. A selecting unit selects one of the encoding results obtained by the plurality of variable-length coding units.
The invention concerns a method for synchronising a video image (IM) spatial position to find the position of an initial digital coding grid (GI), by said image (IM) coding blocks, characterised in that it consists in the following steps: a) moving along the direction of the image (IM) lines and/or along the direction perpendicular to the image (IM) lines, an analysis grid (GA) consisting of coding blocks and corresponding to at least part of the image (IM); b) determining for each position of the analysis grid (GA), the transformed coefficients Fu,v of each block of the analysis grid (GA), using a discrete cosine transform (TCD); and computing for each analysis grid (GA) position, an activity parameter (AM, ES) representing the image activity; c) determining the analysis grid (GA) position for which said activity parameter (AM, ES) has a minimal value.
The invention concerns a method for assessing the degradation of a video image coded by blocks of picture elements or pixels, comprising the following steps: a) selecting an input image and determining its spatial activity (SA) in an analysis window according to the following sub-steps: i) determining for each block (n, m) the transformed coefficients Fn m (i, y) by a block transform; ii) determining from the transformed coefficients Fn, m (i, y) each block spatial activity bsa; iii) determining the analysis window global spatial activity SA.sub.1 ; b) selecting the output image corresponding to the input image and determining said global spatial activity SA.sub.2 of the output image; c) comparing the spatial activities (SA.sub.2) and (SA.sub.1)
A soft windowing technique involves the use of a window for vector assisted frame differing, the window being larger than a macroblock of the structure defined over the video picture. The edges of each window taper so that overlapping windows provide the correct pixel weighting at the edges. For a scheme having 16*16 luminance macroblocks and 8*8 chrominance macroblocks, the luminance window is 24*24 and the chrominance window is 12*12.
