Digital Signal Processing Seminar

Inverse halftoning is the process by which a halftone (binary) image is converted to grayscale. Some work on this topic has appeared in recent literature. However, current inverse halftoning algorithms tend to be computationally intensive. We demonstrate techniques of far lower computational complexity that do not sacrifice visual quality.

We present two simple and efficient algorithms for performing inverse halftoning. The first algorithm produces inverse halftones from both dithered and error diffused halftones at a fraction of the time taken by conventional algorithms. The second seeks to minimize the computation required to produce inverse halftones from error diffused images only, while achieving comparable visual quality to the best algorithms in the literature