Image processing apparatus with improved dithering scheme

摘要

This invention utilizes the fact that "a maximum density difference in a window within a predetermined range including a pixel of interest is large for a character region, and is small for a photograph region." A maximum density difference of an image within a predetermined range is calculated as a feature amount. A quantization error of pixels around a pixel of interest is calculated, and then, a correction amount is calculated by proportionally distributing the quantization error in accordance with the feature amount calculated by a feature amount calculator. The correction amount is added to an image signal of the pixel of interest to form a compensation image signal, and the compensation image signal is quantized.