Compressed image data processing uses both gradient filtering and discontinuity detection for location of image blocks

摘要

Data processing method includes use of gradient filtering to identify natural contours within image data, and discontinuity detection to identify block artefacts. The results of these two methods are used in a pass band filter in order to extract the required image data. The data processing method for an input digital image includes a gradient filtering stage (GF) for the values of the pixels, allowing detection (THR) of a natural contour zone (NC) within the input digital image. It also includes a stage for detection of block artefacts (BAD) due to the technique of block coding, based on the calculation of a discontinuity value with respect to the values of a current pixel, and of the pixels adjacent to the current pixel. The method then includes a stage of pass band filtering (LPF) of the values (Y) of the pixels emerging from the block artefact detection stage, with the exception of the pixels contained in the natural contour zones determined by the gradient filtering stage.