JPH0427273A - Multilevel image estimating system - Google Patents

Abstract

PURPOSE:To prevent the diffusion of an image generated by repeating binarization and multilevel conversion by extending a shape of an opening at the time of multilevel conversion in the same direction as an error diffusion matrix, in the case of converting an image binarized by an error diffusion method to multilevel. CONSTITUTION:A binary image to be converted to multilevel is read in succes sively to a line buffer 101 at every raster. In this case, to each latch 102, a picture element corresponding to a multilevel conversion opening 603 is latched, and outputted to A2 - A14. The outputs A2 - A14 of each latch in this case are inputted to an address line of a ROM 103, and a multilevel estimated value set in advance to the ROM is outputted. Also, as for the contents of the ROM in this case, multilevel data corresponding to density of dots in a multilevel conversion matrix 603 is set in advance. As for the multilevel conversion opening 603, it will suffice that it is a shape for extending in the same direction as a diffusion matrix 403. By setting weighted data corresponding to a dot position, the multilevel conversion can more exactly be executed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention relates to a multi-value image estimation method for estimating a multi-value image from a binary image.

[Prior Art] When displaying a multivalued image on a CRT that can display only a limited number of colors, or printing it on a dot printer, it is common to binarize the image using the dither method or error diffusion method and display it in a pseudo manner. used.

[Problem to be solved by the invention] However, once a multivalued image is binarized, there is a problem that when processing such as enlarging or reducing a single image, the image becomes unrecognizable. .

Therefore, the above problem can be solved by estimating a multivalued original image from a binary image, performing processing such as enlarging, reducing, and rotating the multivalued image, and then binarizing the results. can.

As a method for estimating a multivalued original image from this binary image, there is, for example, a method for estimating a halftone image, disclosed in Japanese Patent Application Laid-Open No. 62-281673 by Hiratsuka and Matsunawa. What about this method? As shown in diagram s2, a certain aperture 202 is set for a binary image 201, and multi-value image data of a pixel of interest 203 is estimated based on the dot density within the aperture.

However, when the halftone image estimation method is applied to an image binarized by the error diffusion method, there is one difference between the two methods.

That is, in the binarization method using the error diffusion method, as shown in FIG. Value conversion is performed, and the binarization error generated at this time is diffused to the pixel of interest 401 shown in FIG. 4 using, for example, an error diffusion matrix 402.

Therefore, the binarization error caused by the error diffusion method is diffused downward in a fan shape with respect to the pixel of interest 501 as shown in FIG.

Therefore, in order to perform processing such as diffusion and reduction on a binary image, if the following processing is repeatedly performed, processing such as enlargement and reduction ↓ Multi-value image ↓ Conventional multi-value conversion This creates a problem in that the binarization error is diffused downward in a fan shape.

[Means and effects for solving the problem] According to the present invention, when an image binarized by the error diffusion method is multi-valued, the shape of the aperture during multi-value processing is the same as the error diffusion matrix. By expanding the image in the horizontal direction, it is possible to prevent the image from spreading out due to repeated binarization and multi-value conversion.

[Embodiment] FIGS. 1(a) and 1(b) show an embodiment of the present invention, in which 101a and 101b are line buffers that store one raster, 102a to 1020 are 1-bit latches, and 103 is a ROM. It is.

FIG. 6 is an explanatory diagram when performing multi-value conversion in the embodiment of the present invention, where 601 is a binary image to be multi-valued, 602 is a pixel of interest to be multi-valued, and 603 is a multi-value aperture.

Next, in the above configuration, the binary image to be multivalued is read into the line buffer sequentially for each raster. At this time, each latch 102 latches a pixel corresponding to the multilevel aperture 603 in FIG. 6, and outputs it to A2-A14.

At this time, the outputs A2 to A14 of each latch are stored in the ROM 10.
3, and a multivalued estimated value preset in the ROM is output.

In addition, the contents of the ROM at this time are set in advance with multivalue data corresponding to the density of dots in the multivalue matrix 603. For example, when the number of dots in the multivalue aperture is n, the output of the ROM By setting R to R=[-X255], it is possible to obtain a multi-value estimated value corresponding to the dot density within the multi-value aperture.

However, [ ] means truncated to the nearest whole number. Ru.

In the embodiment described above, the multi-level aperture 603 does not have the same shape as the diffusion matrix 402 of the error diffusion method, but may have a shape that spreads in the same direction as the diffusion matrix 402. .

Further, the above method is described for a single plane, eg, a black and white image, but can be easily extended to multiple planes, eg, RGB color binary images.

Further, in the above embodiment, multi-leveling is performed using a ROM, but
It is also possible to replace it with RAM and change the contents of RAM from the CPU.

In addition, in the above embodiment, multi-value data corresponding to the dot density of the multi-value aperture is set, but more accurate multi-value data can be achieved by setting weighted data corresponding to the dot position. It is.

[Effects of the Invention] As explained above, in this method, a multi-value original image is estimated from a binary image obtained by the error diffusion method using a multi-value aperture that spreads in the same direction as the spread of the error diffusion matrix. By suppressing image diffusion, it is possible to estimate a multivalued image that is faithful to the original from a binary image.

[Brief explanation of the drawing]

Fig. 1 is a configuration diagram of an embodiment of the present invention, Fig. 2 is a multi-level conversion matrix diagram of a conventional example, Fig. 3 is an explanatory diagram of image scanning during multi-level conversion, Fig. 4 is an error diffusion matrix diagram, and Fig. 5 is an explanatory diagram of binary error propagation, and Figure wS6 is a multilevel aperture diagram of the embodiment of the present invention. 101a, b are line buffers 1028 to 102o are latch 103 is ROM 201 is a binary image 202 is a conventional multi-value aperture 203 is a pixel of interest 301 is a multi-value image 302 is a pixel of interest 401 is a pixel of interest 402 is error diffusion The matrix 501 is the pixel of interest 60 ] and the binary image 602 is the pixel of interest 603 is the multilevel aperture? Mouthful

Claims (1)

[Claims] (1) In a method of estimating a multi-level original image for an image binarized by the error diffusion method, the pixel of interest is estimated in the same direction as the diffusion matrix used in the binarization process by the error diffusion method. A multilevel image estimation method that sets a widening aperture and estimates the multilevel original image of the pixel of interest based on the dot density within the aperture.