Unquantized resolution conversion of bitmap images using error diffusion

Abstract

In a method for converting from a first bitmap image resolution to a second bitmap resolution, be it interpolation (an increase in resolution) or decimation (a decrease in resolution) a bitmap image defined at resolution and density I.times.J.times.a is prefiltered as a first step to ensure the preservation of original image structure and compensate for device dependence of the input image, which is necessary in order to maintain the quality and sharpness of the original image. The filter may be selected from a variety of filter choices and may also perform pixel interpolation or decimation depending on the input source. As a second step, unquantized scaling of a prefiltered image (K.times.L.times.b) is performed using an f.times.g pixel window to identify a set of neighboring input pixels surrounding a single output pixel. The window contributes input pixel density information to generate a scaled gray output pixel. Each pixel in the window is assigned a weighting factor to establish a density contribution to the output pixel. As a third step, each pixel in the gray scale image (M.times.N.times.c) is quantized to a desired depth using error diffusion or other quantizing methods. The gray scale image is converted from an image defined at c levels to an output image defined at d levels where each of the d levels can be reproduced by a selected output device.