JP2738865B2 - Image processing device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for performing multi-value processing of image data into multi-value image data, and more particularly to input image data and multi-value image data after multi-value processing. The present invention relates to an image processing apparatus that performs multi-value processing while correcting an error with the image processing.

[Conventional technology]

Conventionally, an error diffusion method, an average error minimization method, and the like have been known as a density preserving type quantization method. FIGS. 7 (a) and 7 (b) are block diagrams showing multi-value processing of input image data by the error diffusion method. Multi-value (n-value)
The section has a configuration in which n-1 thresholds are used,
The threshold value is compared with the _{values of} t ₁ to t _n−1 and the image data to perform the image n-value conversion processing. The error (difference between the input image data and the output image data after the n-value processing) generated at the time of the n-value processing is calculated by the error calculation unit, and a plurality of the surrounding pixels that have not been binarized yet are calculated. It is added to a single pixel, and the density of the input image and the output image is preserved.
According to this type of quantization method, there is an advantage that it is possible to satisfactorily reproduce both a line image such as a character and a halftone image such as a photograph.

[Problems to be solved by the invention]

However, in the above-described conventional example, n-value
The image data is compared with _one threshold value t1 to tn _-1 . For up to this time t ₁ ~t _n-1 has a value close to each other, the error between the image data value of the n value conversion result and the original pixel is very small, around the error is very small pixel Will not be sufficiently diffused. For this reason, the error diffusion method is not multi-valued, but very similar to simple multi-valued conversion.

For this reason, when the original image is a halftone image having a gradation (density gradient), the boundary when performing multi-valued processing becomes clear, a pseudo contour occurs, and the image quality deteriorates.
There was a drawback that the image quality was sometimes worse than the binarization by the error diffusion method.

Further, the cause of the occurrence of the pseudo contour can be considered when outputting the result of the n-value conversion to a printer or the like. For example, depending on the density of an ink jet printer, a large density difference occurs when switching between light ink and dark ink to be output. For this reason, a pseudo contour occurs near the density corresponding to the ink switching. FIG. 8 (a) shows an example of the processing result when n-value conversion is performed by the error diffusion method. Since the amount of error diffusion is small, a clear boundary appears, and the output "1" is output. If the dark and light inks used are different from those of "2" and "2", there is a disadvantage that a clearer pseudo contour is generated.

The present invention eliminates the above-mentioned disadvantages of the prior art. By correcting error data generated at the time of multi-value processing, it is possible to reproduce an image excellent in both resolution and gradation and to perform multi-value processing. Of at least one of the plurality of sets of thresholds used in the step (b) is biased toward the higher density level, and at least one set of thresholds is biased toward the lower density level. It is an object of the present invention to provide an image processing apparatus capable of preventing the occurrence of false contours.

Means for Solving the Problems In order to achieve the above object, an image processing apparatus according to the present invention is an image processing apparatus for converting input image data into multi-valued image data. A plurality of thresholds for multi-value processing of one pixel as one set, a threshold generation unit that generates the plurality of sets of thresholds by periodically changing, and a set of a plurality of thresholds from the threshold generation unit; 1
A multi-level converting means for performing multi-level processing of the input image data into multi-level image data by comparing the input image data of the pixels; and a multi-level processing after the multi-level processing and the input image data generated by the multi-level processing. Correction means for correcting an error with the value image data, wherein the threshold value generated by the threshold value generation means is biased toward a higher density level of at least one of the plurality of threshold values, and further at least One set of threshold values is biased toward the lower density level.

〔Example〕

Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing one embodiment of the present invention.

An image input unit 1 reads an original image and outputs digital image data quantized to 8 bits. Further, the image input unit may be configured to input an image from a computer. The adder 2 adds the diffusion error from the line buffer 6 to the input image data input by the image input unit 1. This diffusion error is an error generated when multi-value processing is performed on a pixel before the currently input image data. Next, multi-level conversion is performed in the multi-level conversion unit 3. The result is output as a final output to an image output unit 7 including a laser beam printer, an ink jet printer, a monitor, and the like. This output is sent to the difference calculator 4. The difference calculator 4 calculates a difference between the input image data before multi-value conversion and the output image data after multi-value conversion. This calculation result becomes an error in multi-value conversion. This error is divided by the distribution calculation unit 5 as data to be distributed to peripheral pixels of the pixel currently being processed. The distribution calculation unit 5 performs weighting so that more errors are distributed to pixels close to the pixel currently being processed. Diffusion error data obtained at an appropriate division distribution ratio is added to an appropriate position corresponding to a peripheral pixel of the line buffer 6. Finally, the error assigned to the pixel of interest is added in the line buffer 6, and the addition of the error is performed by the adder 2 at the stage of multi-value processing. The error can of course be negative.

The error is positive when the output image data is smaller than the input image data, and is negative when the output image data is larger than the input image data.

Next, the multi-value conversion unit 3 will be described in detail. FIG. 2 is a diagram showing an embodiment of the detailed configuration of the multi-level conversion unit, where n-1 to 11- (n-1) are used for multi-level conversion (n-level conversion).
Are provided.

In the comparators 11-1 to 11 (n-1), threshold ROMs 10-1 to 10 (n-1) storing the input image data and the threshold, respectively.
Then, when the image data is larger than the threshold value, 1 is outputted, and when the image data is smaller than the threshold value, 0 is outputted to the encoder 12. In the encoder 12, the comparators 11-1 to 11- (n-1)
Is determined based on the number of comparators exceeding the threshold value, that is, the number of 1 values output from the comparators, based on the comparison result.

The threshold data stored in the threshold ROMs 10-1 to 10- (n-1) output a value determined according to the position of the pixel of interest (pixel data being processed). For this reason, the pixel position is determined according to the pixel clock input to the horizontal counter 8 and the number of horizontal synchronization signals (HSYNC) input to the vertical counter 9, and the outputs from the horizontal counter 8 and the vertical counter 9 are used as the threshold ROM 10. -1 to 10- (n-1) are supplied as address data.

FIG. 3A shows threshold data stored in the threshold ROMs 10-1 to 10-3 when n = 4, that is, when the input image data is quaternized.

Here, the input image data is assumed to be data quantized to 8 bits (0 to 255), and the data output after being subjected to the multi-value processing is assumed to be 4-value data of 0, 80, 160, 255. Fig. 3 (a)
X ₁₁ ~x ₂₂ shown in 13 indicate the positions of pixels in.

Each of the threshold matrices shown in 14, 15, and 16 is 2 × 2 in this embodiment. Therefore, the horizontal counter 8 and the vertical counter 9 use only the lower 2 bits as a threshold to specify the pixel position.
It would be good to supply it to ROM10. Eventually, the multi-level conversion unit 2 uses the threshold value of any one of the four types shown in FIG. 3B. For example, at a pixel position where a threshold value of [20, 40, 60] is selected, there is a high possibility that the density of the quaternized error diffusion processing result will be high.
(8 bits 255), the error diffusion amount is likely to be large. Conversely, if a threshold of [200, 220, 240] is selected, the processing result tends to output thinly,
Since the input image data less than 0 is output as 0 (8 bit 0), the possibility that the amount of error diffusion becomes large increases.

In this way, by making the threshold value higher or lower, the amount of error diffusion can be made larger than when the error diffusion method is performed using three simple threshold values, for example, (60, 120, 180).

In other words, according to the present embodiment, it is possible to prevent a simple n (4) -value processing without correcting the error.
As a result, it is possible to prevent the level after the n-value conversion process from clearly changing, and to prevent the occurrence of a false contour.

Moreover, in this embodiment, the threshold value is periodically changed according to the pixel position. For example, in FIG.
Have higher concentration levels of threshold _11, the order of _{x 22, x 12, x 21} . Therefore, when considered as a 2 × 2 matrix, the thresholds can be balanced without setting the thresholds to extremely low or high densities. The method of setting the threshold is not limited to the above embodiment. For example, when quaternary processing is performed using the threshold value of FIG.
The output of 0) and 3 (8 bit 255) increases, but if the combination of thresholds is assumed to be (20, 50, 80), 1 (8 bit 80) or 2
(8bit160) is output more frequently.

In other words, the smaller the data difference among a set of threshold values, the closer the processing result to binarization.

Therefore, by using this, when the original is a character or the like, to improve the resolution, use a set of threshold values with a small difference between the respective data, and when the original is a photograph or the like, to improve the gradation, If one set of threshold values having a large difference between the respective data is used, the image quality can be further improved.

[Other embodiments]

FIG. 4 is a block diagram showing a second embodiment of the present invention. In the figure, the same components as those in FIG. 1 are denoted by the same reference numerals, and detailed description thereof will be omitted.

The multi-level conversion unit 21 simply performs four-level conversion using, for example, the three threshold values shown in FIG. A dither signal is added to the data at the time of performing the quaternary conversion in the second adder 20 before performing the multi-value processing. The second adder 20 has a dither ROM 1
Nine outputs are given, and the position information obtained by counting the target pixel position by the horizontal counter 17 and the vertical counter 18 is given as an address to the dither ROM. Each counter is counted based on the image clock and the horizontal synchronization signal HSYNC. Example of data contents of dither ROM 19
It is shown in FIG. In this case, the threshold value of the multi-level conversion unit 31 is
Then, the same result as in the first embodiment can be obtained.

If a [20,40,60] in FIG. 6 is changed to [20,50,80], the output frequency of 1,2 becomes high in addition to the output of 0,3 in the case of multi-valued conversion. .

The effect on the output can be changed by the combination of the dither ROM 19 and the threshold value of the multi-value quantization unit 31. That is, the image output unit 7 is a display device, a laser beam printer, or
The combination described above can be arbitrarily set depending on whether the printer is an LBP or an inkjet printer, and image processing can be performed according to the output device. For example, in the combination of FIG. 5B and FIG. 6B, when quaternized, the appearance frequencies of “0”, “1”, “2”, and “3” can be prevented from fluctuating, In addition, in the combination of FIG. 5C and FIG. 6B, the appearance frequency of “0” and “3” can be reduced, and the appearance frequency of “1” and “2” can be increased. This is because the dither signal of FIG. 5B has a larger amplitude than the dither signal of FIG. 5C.

The larger the amplitude of the dither signal is, the larger the false contour prevention width shown in FIG. 8B is, and the occurrence of the false contour can be prevented. On the other hand, in this case, noise is generated in the reproduced image. To suppress the noise, a signal having a small dither signal amplitude may be used.

In other words, the output device having a large amplitude of the dither signal shown in FIG. 5B is used for an output device in which the pseudo contour is conspicuous according to the characteristics of the output device, and the output device is used for the output device in which the pseudo contour is not conspicuous. 5 By using a dither signal having a small amplitude as shown in FIG. 5 (c), it is possible to obtain an image without noise.

In the above-described embodiment, the dither signal is added to the portion shown in FIG. 4B, but it is also possible to add the dither signal to the position shown in FIG. 5A. In this case, from the viewpoint of preserving the density in the error diffusion method, it can be easily considered that the value of the dither ROM to be added takes both positive and negative values, and it is desirable that the sum thereof be 0.

Further, in the present invention, the size (size) of the dither ROM or the threshold ROM is not limited.

In the present embodiment, the case where the number of input image data is one has been described. However, the present invention can be applied to a color image by increasing this data to R, G, and B.

As described above, according to the present embodiment, by operating a threshold value, input image data, and the like at the time of multi-value conversion, the multi-value conversion result of error diffusion has a variation effect, which is useful for preventing false contours. It is possible to make it.

Further, depending on how to set the threshold value of the threshold ROM or the dither ROM, an effect of preventing the appearance of a texture structure peculiar to error diffusion also occurs.

〔The invention's effect〕

As described above, according to the present invention, it is possible to reproduce an image excellent in both resolution and gradation by correcting error data generated at the time of the multi-value processing, and a plurality of sets used at the time of the multi-value processing. By biasing at least one set of thresholds toward higher density levels and further biasing at least one set of thresholds toward lower density levels, the amount of errors that occur is increased, and pseudo contours are increased. Can also be prevented from occurring.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of the present invention, FIG. 2 is a diagram showing a detailed configuration of a multi-value conversion unit, FIG. 3 is a diagram showing a multi-value threshold matrix, and FIG. FIG. 5 is a block diagram showing a second embodiment of the present invention, FIG. 5 is a diagram showing a dither signal used in the second embodiment, and FIG. 6 is a multi-level coding used in the second embodiment. 7 (a) and 7 (b) are diagrams for explaining a general multi-valued error diffusion method, and FIG. 8 is a diagram showing a conventional multi-value processing result and the present embodiment. FIG. 14 is a diagram illustrating an example of a multi-value processing result. In the figure, 1 is an image input unit, 2 is an addition unit, 3 is a multi-value conversion unit, 4 is a difference calculation unit, 5 is a distribution operation unit, 6 is a line buffer, 7
Is an image output unit, 8 is a horizontal counter, 9 is a vertical counter,
10-1 to 10- (n-1) are threshold ROMs, and 11-1 to 11 (n-
1) is a comparator, 12 is an encoder, 17 is a horizontal counter, 18
Denotes a vertical counter, 19 denotes a dither ROM, 20 denotes an addition unit, and 21 denotes a multi-value conversion unit.

Claims (1)

(57) [Claims] 1. An image processing apparatus for converting input image data into multi-valued image data, comprising: a set of input means for inputting image data; and a plurality of thresholds for multi-value processing of one pixel. A threshold value generating means for periodically varying a plurality of sets of threshold values; and a plurality of sets of threshold values from the threshold value generating means and input image data of one pixel to compare input image data at a multi-level level. A multi-level processing unit for performing multi-level processing on the image data; and a correction unit configured to correct an error between the input image data generated by the multi-level processing and the multi-level image data after the multi-level processing. The threshold value generated by the generating means is such that at least one of the plurality of threshold values is biased toward a higher density level, and at least one of the threshold values is biased toward a lower density level. Image processing Apparatus.