Error Diffusion Device and Method for Color Images Using Halftoning

High-Speed Color Halftoning Using a 3×5 Diffusion Matrix with a Power-of-Two Denominator

This technology is an error diffusion device and method that creates an error diffusion filter by calculating the components of a 3×5 error diffusion matrix based on a filter determination parameter N, and performs halftoning by applying this filter to an input color image. The matrix denominator is set as a power of two to simplify calculations.

Error diffusion is a halftoning technique that distributes pixel quantization residuals to neighboring pixels, providing edge enhancement. However, it has historically suffered from long processing times due to the need for pixel-by-pixel sequential processing and buffering. Existing Floyd-Steinberg filters use denominators that are arbitrary integers, resulting in a high computational burden for division.

Based on the Floyd-Steinberg method, this technology sets N so that the diffusion coefficient denominator is a power of two, reducing division to simple bit-shift operations, and diffuses errors by separating red, green, and blue channels. It maintains structural simplicity while being faster than existing methods and producing visually natural images. Applicable to inkjet/laser printer controllers and e-paper display drivers, it enables fast binarized output of color documents and photos even on low-spec embedded chips.

‍

‍

Key Features:

• An input unit that receives the target image for error diffusion and a parameter determination unit that sets the filter determination parameter
• A filter formation unit that calculates the components of the error diffusion matrix based on the filter determination parameter to form an error diffusion filter
• A 3×5 error diffusion matrix that distributes diffusion errors of N/(10N-12) to the right of the current pixel and (N-1)/(10N-12) in the diagonal direction
• An error diffusion unit and output unit that separate the image into red, green, and blue channels, apply the error diffusion matrix, and output the error-diffused image

‍