Abstract: A method is provided for deinterlacing pixels in a video display system operable to display images as a plurality of pixels arranged in first and second fields of interlaced rows, at least one video component level quantified by a numerical video component value characterizing each pixel in an image with the video component levels for the pixels in the first and second fields being updated on alternate scans. A global mean video component value is computed from the video component values generated for a previous scan of at least one of the first and second fields. A global standard deviation is computed for the global mean. A local mean video component value is computed from the video component values generated for a plurality of pixels being updated as part of the current scan of the first field and defining a neighborhood of a pixel in the second field being deinterlaced. A local standard deviation is computed form the local mean video component value.