Dither Error Diffusion for PWM Moire-Free Image Processing

The patent applies preliminary action by pre-calculating and storing correction values in a correction value table before actual image formation. The correction values, which compensate for PWM errors, are prepared in advance based on the non-linear relationship between pulse width and light amount. During image processing, these pre-prepared correction values are applied to the concentration values of pixels, eliminating the need for real-time complex calculations and directly addressing the PWM error problem.

The patent introduces a correction value table as an intermediary between the PWM signal generation and the light amount control. This table acts as a mediator that translates the non-linear PWM error characteristics into compensatable data. The correction values stored in the table serve as intermediate correction data that is added to or subtracted from pixel concentration values, thereby indirectly correcting the PWM errors without requiring direct modification of the PWM signal generation mechanism.

The patent applies the taking out principle by extracting and isolating the PWM error correction function from the general error diffusion process. Instead of applying comprehensive error diffusion that affects all pixels uniformly, the invention selectively applies correction only to pixels where PWM errors occur, based on their concentration values. This selective correction approach extracts the essential correction need from each pixel and applies targeted correction, avoiding the propagation of errors to pixels that do not require correction and thus preventing moire and pseudo contour formation.

The patent implements local quality by applying different correction strategies to different pixels based on their individual characteristics. The correction value application is not uniform across the entire image but is localized to specific pixels whose concentration values fall within ranges where PWM errors are significant. This localized correction approach ensures that correction is applied precisely where needed while leaving other pixels unaffected, thereby maintaining image quality and avoiding artifacts like moire and pseudo contour that result from uniform error diffusion.