Method and apparatus for converting halftone images

摘要

The edge feature value of a M-valued original image is detected and converted to an N-valued image (where M>N) using an error diffusion method while suppressing the quantization error in edge areas of the image by edge area enhancement. The N-valued image data is scanned with a edge feature value detection filter to detect feature quantities and is scanned by plural spatial filters with different smoothing characteristics to normalize the output value of each spatial filter to a specific value. The normalized outputs of the spatial filters are weighted with a weighting coefficient based on the edge feature value to generate the interpolated image data, and the resulting image is normalized to L-valued image data (where N<L). The mixing ratio and image quality are also controlled by manipulating the edge feature value and weighting coefficients.