Method and apparatus for gradation reproduction of continuous tone image

1. A method for gradation reproduction of a continuous tone image based on an error diffusion method in which, when the gradation of input image data on a continuous tone image is converted into multivalue data based on at least three values using at least two sets of threshold values to reproduce a gradation, an error between a data item before multivaluing and the corresponding multivalued data item is spread, as a corrected error and under spreading conditions determined by weight coefficients, to a plurality of subsequent data items successively input to obtain corrected data, and the corrected data is multivalued, the gradation reproduction method comprising the step of determining the interval between the threshold values so that an amplitude of the corrected data according to rising and falling variations in the corrected error spikes outward beyond the threshold values at least at a halftone gradation between a darkest gradation and a lightest gradation, whereby generation of a transient region where the amplitude of the corrected data becomes excessively small at the halftone gradation of the corrected data is prevented.

2. A method for gradation reproduction of a continuous tone image according to claim 1 ; wherein an amplitude of the corrected error is produced in accordance with the spreading conditions determined by the weight coefficients.

3. A method for gradation reproduction of a continuous tone image according to claim 2 ; further comprising the step of determining the threshold values in accordance with the gradation of the input image data by referring to a threshold value look-up table in which the threshold values and the interval between the threshold values are set in advance.

4. A method for gradation reproduction of a continuous tone image according to claim 1 ; further comprising the step of adding pseudo-random numbers to the input image data.

5. A method for gradation reproduction of a continuous tone image according to claim 4 ; wherein an amplitude of the pseudo-random numbers is set so as not to exceed a certain limit value.

6. A method for gradation reproduction of a continuous tone image according to claim 5 ; wherein the limit value of the amplitude of the pseudo-random numbers varies in accordance with the value of input image data.

7. A method for gradation reproduction of a continuous tone image according to claim 6 ; wherein the limit value of the amplitude of the pseudo-random numbers is zero when the value of input image data is zero or at a maximum of a dynamic range.

8. A method for gradation reproduction of a continuous tone image according to claim 6 ; wherein the limit value of the amplitude of the pseudo-random numbers is a value varying smoothly with respect to the input image data.

9. A method for gradation reproduction of a continuous tone image according to claim 6 ; further comprising the step of determining the limit value of the amplitude of the pseudo-random numbers by referring to an amplitude limit value look-up table in which limit values of the amplitude of the pseudo-random numbers according to the input image data are set in advance.

10. A method for gradation reproduction of a continuous tone image according to claim 6 ; wherein the limit value of the amplitude of the pseudo-random numbers is locally changed with respect to a median of a dynamic range or a value in the vicinity of the median.

11. A method for gradation reproduction of a continuous tone image according to claim 10 ; wherein the limit value of the amplitude of the pseudo-random numbers is changed in an increasing direction with respect to the median of the dynamic range or a value in the vicinity of the median.

12. A method for gradation reproduction of a continuous tone image, comprising the steps of:

providing input image data corresponding to a target pixel among a plurality of pixels;

adding pseudo-random numbers to the input image data;

correcting the input image data to which the pseudo-random numbers have been added to form corrected data;

converting the corrected data into multivalue data by comparing the corrected data to at least two threshold values and determining an interval between the threshold values so that an amplitude of the corrected data spikes outward beyond the threshold values at a halftone gradation of the corrected data to thereby prevent generation of a transient region where the amplitude of the corrected data becomes excessively small at the halftone gradation of the corrected data;

computing an error between the corrected data and the multivalue data; and

spreading the computed error to pixels neighboring the target pixel in accordance with preselected weight coefficients.

13. A method for gradation reproduction of a continuous tone image, comprising the steps of:

providing input image data corresponding to a target pixel among a plurality of pixels;

adding pseudo-random numbers to the input image data;

setting an amplitude of the pseudo-random numbers so as not to exceed a certain limit value that varies in accordance with the value of the input image data;

correcting the input image data to which the pseudo-random numbers have been added to form corrected data;

converting the corrected data into multivalue data by comparing the corrected data to at least two threshold values;

computing an error between the corrected data and the multivalue data; and

spreading the computed error to pixels neighboring the target pixel in accordance with preselected weight coefficients.

14. A method for gradation reproduction of a continuous tone image, comprising the steps of:

providing input image data corresponding to a target pixel among a plurality of pixels;

adding pseudo-random numbers to the input image data;

setting an amplitude of the pseudo-random numbers so as not to exceed a certain limit value that varies smoothly with respect to the input image data;

correcting the input image data to which the pseudo-random numbers have been added to form corrected data;

converting the corrected data into multivalue data by comparing the corrected data to at least two threshold values;

computing an error between the corrected data and the multivalue data; and

spreading the computed error to pixels neighboring the target pixel in accordance with preselected weight coefficients.

15. A method for gradation reproduction of a continuous tone image, comprising the steps of:

providing input image data corresponding to a target pixel among a plurality of pixels;

adding pseudo-random numbers to the input image data;

setting an amplitude of the pseudo-random numbers so as not to exceed a certain limit value that is determined by referring to an amplitude limit value look-up table in which limit values of the amplitude of the pseudo-random numbers according to the input image data are set in advance;

correcting the input image data to which the pseudo-random numbers have been added to form corrected data;

converting the corrected data into multivalue data by comparing the corrected data to at least two threshold values;

computing an error between the corrected data and the multivalue data; and

spreading the computed error to pixels neighboring the target pixel in accordance with preselected weight coefficients.