JP2821238B2 - Multi-valued image estimation method - Google Patents

Description

Description: TECHNICAL FIELD 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 multi-valued image on a CRT which can display only a limited number of colors or printing on a dot printer, a method of binarizing the image by a dither method or an error diffusion method and displaying the image in a pseudo manner is often used. Can be

[Problems to be Solved by the Invention] However, once a multi-valued image is binarized, there is a problem that the image is lost when a process such as enlargement or reduction of the image is performed.

Therefore, it is possible to solve the above problem by estimating a multi-valued original image from a binary image, performing processing such as enlargement, reduction, and rotation on the multi-valued image and binarizing the result. it can.

As a method of estimating a multi-valued original image from this binary image, there is, for example, Japanese Patent Application Laid-Open No. 62-281673 by Hiratsuka and Matsunawa, and a halftone image estimating method. In this method, as shown in FIG.
An opening 202 is set for 201, and the multivalued image data of the target pixel 203 is estimated based on the dot density in the opening.

However, when the halftone image estimation method is applied to an image binarized by an error diffusion method, the following problem occurs.

That is, in the binarization method by the error diffusion method, when the multi-valued image 301 is binarized while sequentially scanning the multi-valued image 301 in the raster direction as shown in FIG. The binarization error that occurs at this time is
The target pixel 401 shown in FIG. 4 is diffused using, for example, an error diffusion matrix 402.

Therefore, the binarization error due to the error diffusion method is diffused downward in a fan shape with respect to the target pixel 501 as shown in FIG.

Therefore, if the following processing is repeated to perform processing such as diffusion and reduction on a binary image, There is a problem that the binarization error is always diffused downward in a fan shape.

[Means and Actions for Solving the Problems] According to the present invention, when an image binarized by the error diffusion method is multi-valued, the shape of the aperture at the time of multi-value conversion is the same as the error diffusion matrix. Spreading in the direction prevents image diffusion caused by repeating binarization and multi-leveling.

[Embodiment] FIGS. 1 (a) and 1 (b) are views showing an embodiment of the present invention, wherein 101a and 101b are line buffers for storing one raster.
102a to 102o are 1-bit latches, and 103 is a ROM.

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

Next, in the above configuration, the binary image to be multivalued is sequentially read into the line buffer for each raster. At this time,
Pixels corresponding to the multivalued aperture 603 in FIG. 6 are latched on each latch 102 and output to A2 to A14.

The outputs A2 to A14 of the latches at this time are input to the address lines of the ROM 103, and a multi-valued estimated value set in the ROM is output.

The contents of the ROM at this time are based on the multi-valued matrix 6
Multi-value data corresponding to the dot density in 03 is set in advance. For example, when the number of dots in the multi-valued aperture is n, the output R of the ROM is By doing so, a multi-valued estimated value corresponding to the dot density in the multi-valued aperture can be obtained.

 However, [] means cutting to the nearest decimal.

In the above embodiment, the multi-valued aperture 603 has the same shape as the diffusion matrix 402 of the error diffusion method. However, the shape is not limited to this, and any shape may be used as long as it spreads in the same direction as the diffusion matrix 402.

Further, in the above-mentioned method, a single plane, for example, a black and white image is described, but it can be easily extended to a plurality of planes, for example, an RGB color binary image.

Further, in the above embodiment, the multi-valued data is stored in the ROM,
It is also possible to change the contents of RAM from CPU instead of AM.

Further, in the above embodiment, multi-value data corresponding to the dot density of the multi-valued aperture is set. However, by setting weighted data corresponding to the dot position, more accurate multi-valued data can be obtained. is there.

[Effects of the Invention] As described above, in the present method, a multi-valued original image is estimated from a binary image by the error diffusion method using a multi-valued aperture that spreads in the same direction as the spread of the error diffusion matrix. In addition, it is possible to suppress image diffusion and to estimate a multi-value image faithful to the original from the binary image.

[Brief description of the drawings]

 FIG. 1 is a block diagram of an embodiment of the present invention, FIG. 2 is a diagram of a conventional multi-valued matrix, FIG. FIG. 6 is an explanatory diagram of the binarization error propagation, and FIG. 6 is a multi-valued aperture diagram of the embodiment of the present invention. 101a, b are line buffers 102a to 102o are latches 103 are ROM 201 are binary images 202 are conventional multi-valued apertures 203 are target pixels 301 are multi-valued images 302 are target pixels 401 are target pixels 402 are error diffusion matrices 501 is the target pixel 601 is the binary image 602 is the target pixel 603 is the multi-valued aperture

Claims (1)

(57) [Claims] In a multi-valued image estimation method for estimating a multi-valued original image from an image binarized by an error diffusion method, a diffusion matrix used in a binarization process by the error diffusion method is applied to a target pixel. A multi-valued image estimating method that sets an opening that spreads in the same direction as the above and estimates the multi-valued original image of the pixel of interest from the dot density in the opening.