Image processing apparatus and method

An image processing method of obtaining binarization threshold values for converting a multi-valued image into a binary image, said method comprising the steps of:

• calculating luminance frequencies of said multi-valued image in units of blocks in said multi-valued image;
• calculating first binarization threshold values, in units of blocks, on the basis of the luminance frequencies by specifying an optimal luminance frequency region in the luminance frequencies, calculating an average luminance value in the optimal luminance frequency region, and calculating the first binarization threshold values in units of blocks on the basis of the average luminance value, wherein the specifying includes discriminating a feature of an image on the basis of a skew of the luminance frequencies, and specifying the optimal luminance frequency region on the basis of the feature of the image;
• setting a region for interpolating the first binarization threshold values in units of blocks;
• and
• calculating second binarization threshold values, in units of a pixels, by interpolating the first binarization threshold values in units of blocks, with respect to the set region,
• wherein, in said step of calculating luminance frequencies, the relation between luminance and frequencies of multi-valued levels of a plurality of pixels is obtained as the luminance frequencies, and in said step of calculating first binarization threshold values in units of pixels, the second binarization threshold values in units of pixels is calculated by interpolating the first binarization threshold values-obtained in a plurality of blocks.

An image processing apparatus, which obtains binarization threshold values for converting a multi-valued image into a binary image, comprising:

• first arithmetic operation means for calculating luminance frequencies in units of blocks in a multi-valued image;
• second arithmetic operation means for calculating binarization threshold values in units of blocks on the basis of the luminance frequencies, wherein the second arithmetic operation means comprises specifying means for specifying an optimal luminance frequency region in the luminance frequencies, calculating an average luminance value in the optimal luminance frequency region, and calculating the binarization threshold values in units of blocks on the basis of the average luminance value, and wherein the specifying means comprises discrimination means for discriminating a feature of an image on the basis of a skew of the luminance frequencies, and specifying the optimal luminance frequency region on the basis of the feature of the image;
• and
• interpolation means for calculating binarization threshold values in units of pixels by interpolating the binarization threshold values in units of blocks,
• wherein, said first arithmetic operation means obtains the relation between luminance and frequencies of multi-valued levels of a plurality of pixels as the luminance frequencies, and said interpolation means calculates the binarization threshold values in units of pixels by interpolating the binarization threshold values obtained in a plurality of blocks.

An image processing apparatus, which obtains binarization threshold values for converting a multi-valued image into a binary image, comprising:

• first arithmetic operation means for calculating luminance frequencies in units of blocks in a multi-valued image;
• second arithmetic operation means for calculating binarization threshold values in units of blocks on the basis of the luminance frequencies, wherein said second arithmetic means comprises specifying means for specifying an optimal luminance frequency region in the luminance frequencies, calculating an average luminance value in the optimal luminance frequency region, and calculating the binarization threshold values in units of blocks on the basis of the average luminance valued, and wherein said specifying means comprises discrimination means for discriminating a feature of an image on the basis of a skew of the luminance frequencies, and specifying the optimal luminance frequency region on the basis of the feature of the image;
• first setting means for setting a region for interpolating the binarization threshold values in units of blocks;
• and
• interpolation means for calculating binarization threshold values in units of pixels by interpolating the binarization threshold values in units of blocks with respect to the set region,
• wherein, said first arithmetic operation means obtains the relation between luminance and frequencies of multi-valued levels of a plurality of pixels as luminance frequencies, and said interpolation means calculates the binarization threshold values in units of pixels by interpolating the binarization threshold values obtained in a plurality of blocks.

A computer program product, which comprises a computer useable medium having computer readable program code of obtaining binarization threshold values for converting a multi-valued image into a binary image, comprising:

• computer readable first program code of calculating luminance frequencies in units of blocks in a multi-valued image;
• computer readable second program code of discriminating a feature of an image on the basis of a skew of the luminance frequencies calculated by said first program code;
• computer readable third program code of specifying an optimal luminance frequency region on the basis of the feature of the image discriminated by said second program code;
• computer readable fourth program code of calculating an average luminance value in the optimal luminance frequency region specified by said third program code, and calculating binarization threshold values in units of blocks on the basis of the average luminance value;
• and
• computer readable fifth program code of calculating binarization threshold value in units of pixels by interpolating the binarization threshold values in units of blocks calculated by said fourth program code,
• wherein, by said first program code of calculating luminance frequencies, the relation between luminance and frequencies of multi-valued levels of a plurality of pixels is obtained as luminance frequencies, and by said fifth program code of calculating binarization threshold values in units of pixels, the binarization threshold values in units of pixels is calculated by interpolating the binarization threshold values obtained in a plurality of blocks.