Patent US 5,488,673

An image processing apparatus is arranged such that corrected data for an input objective pixel is obtained by adding the input data for that pixel and data distributed to the position of that pixel, the corrected data is then quantized (e.g., binarized), and, based on the error occurring in quantization, calculating values to be distributed to respective non-quantized neighboring pixels, based on predetermined weighting factors corresponding to the respective positions of those neighboring pixels. The values so calculated are then distributed to the appropriate neighboring pixel positions, and total sum of those values is subtracted from the value of the error, and this difference is in turn distributed to the position of one neighboring pixel located differently with respect to the objective pixel position, at least each time the position of the objective pixel is updated.