METHOD AND APPARATUS FOR WAVELET-BASED UNIVERSAL HALFTONE IMAGE UNSCREENING

A universal halftone image unscreening method includes: 1) decomposing the halftone image into a plurality of subband image levels, where each subband image level includes a plurality of subband images; 2) selectively filtering the plurality of subband images of each subband image level; and 3) recomposing the plurality of filtered subband images into a continuous tone image. The continuous tone image can then be further filtered to enhance its appearance. The subband images are preferably formed using the discrete wavelet transform, but can be formed using any known transform. The subband images are selectively filtered using a non-edge areas noise attenuating filter and a plurality of oriented filters. The halftone image is preferably decomposed using an N-stage halftone image decomposer, wherein each of the N decomposer stages comprises a set of low-pass and high-pass filters. Similarly, the continuous tone image is preferably recomposed using an N-stage continuous tone image recomposer, wherein each of the N recomposer stages comprises a set of low-pass and high-pass filters. The universal halftone unscreening method and apparatus are able to unscreen any halftone image to form the continuous tone image without needing any information about the halftone image halftoning process, and are able to unscreen both original, electronically stored, halftone image data and printed halftone images which have been scanned to generated the electronic image data.