JPH05136988A - Binary image scaling processor - Google Patents

Abstract

(57) [Abstract] [Purpose] To obtain multi-valued binary image data with excellent gradation and resolution by converting into multi-valued estimation data that maintains resolution. A multi-valued estimation block 2 for converting binary image data read from a hard disk device 1 storing a binary image to be scaled into multi-valued estimation data, and a multi-valued estimation block 2 from this multi-valued estimation block. A scaling block 3 that scales the value estimation data to a target scaling factor, and smoothes, subtracts, compares, multiplies, and adds the multivalued estimation data from the scaling block to enhance edges. An edge enhancement block 4 for converting into multi-valued estimation data and a re-binarization block 5 for binarizing the multi-valued estimation data from the edge enhancement block and outputting scaled binary image data. There is.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a binary image scaling apparatus suitable for scaling a pseudo-halftone image in a binary image processing apparatus such as a facsimile machine.

[0002]

2. Description of the Related Art A simple thinning method, an SPC method, a logical sum method, a 9-division method, a projection method and the like are used as a scaling process for a digital image including a binary image. These methods are
It is known that a certain degree of effect can be expected for a binary image such as a line drawing or a multivalued image.

However, in recent years, there has been an increasing number of opportunities to transmit images such as photographs even in facsimile machines and the like.
When the conventional scaling method is used for the pseudo halftone image obtained by processing the multi-tone data of these photographs by the systematic dither method or the error diffusion method, the gradation property is lost and the systematic dither image is changed. Depending on the magnification, moire may occur due to the correlation between the dither cycle and the scaling cycle, resulting in a problem that image quality is significantly deteriorated. As a method for solving such a problem, a scaling processing method suitable for a pseudo halftone image disclosed in JP-A-1-99178 has been conventionally known.

[0004]

However, in this conventional scaling processing method, when multi-valued data is estimated by the filter processing, the edge portion of the image is blurred due to the filter operation, and a resolution such as a character or a line drawing is required. There is a problem that the image quality is deteriorated with respect to the affected portion.

Therefore, the present invention is capable of obtaining multi-valued estimation data while maintaining resolution, and therefore is capable of obtaining scaled binary image data excellent in gradation and resolution. Is to provide.

[0006]

According to the present invention, binary image data is converted into multivalued estimation data by performing multivalued estimation processing with a weighting coefficient of a low-pass filter on a block including a pixel of interest and surrounding pixels. Multi-value estimating means, a scaling processing means for scaling the multi-value estimation data from the multi-value estimating means to a target scaling factor, and an edge component for the multi-value estimation data from the scaling processing means. Edge enhancement means for enhancing only the edge enhancement means, and re-binarization means for binarizing the edge-enhanced multivalued estimation data and outputting scaled binary image data. Is a smoothing means for filtering and smoothing multi-valued estimation data input from the scaling processing means with a predetermined weighting coefficient, and a multi-valued estimation data and smoothing procedure input from the scaling processing means. When the absolute value of the difference is equal to or more than a certain value, the subtraction means for calculating the difference from the data smoothed by is compared with the absolute value of the difference calculated by the subtracting means. Comparing means for outputting the difference data as it is and outputting 0 data when the absolute value of the difference is less than a certain value;
A multiplication means for multiplying the data from the comparison means by a constant and an addition means for adding the data from the multiplication means and the smoothed data from the smoothing means are provided.

[0007]

In the invention having such a structure, the multi-valued estimation data is obtained by performing the multi-valued estimation processing by the low-pass filter on the block of the binary image data including the peripheral pixel around the pixel of interest. Then, this multivalued estimated data is subjected to a scaling process to a target scaling factor and then edge-emphasized.
Edge enhancement smooths the scaled multivalued estimation data,
The difference between the smoothed data and the scaled multi-valued estimation data before smoothing is obtained, and when this difference is equal to or greater than a certain value, the difference data is multiplied by a constant and then added to the smoothed data. Quantize. If the difference is less than a certain value, only the smoothed data is binarized.

[0008]

Embodiments of the present invention will be described below with reference to the drawings.

In FIG. 1, reference numeral 1 is a hard disk device storing a binary image to be scaled, and 2 is a multivalued estimation means for converting the binary image data read from the hard disk device 1 into multivalued estimation data. Of the multivalued estimation block 3, 3 is a scaling block as scaling processing means for scaling the multivalued estimation data from the multilevel estimation block 2 to a target scaling factor, and 4 is a scaling block from the scaling block 3. An edge enhancement block 5 serving as an edge enhancement means for performing edge enhancement of multi-valued estimation data is binarized from the multi-valued estimation data from the edge enhancement block 4, and scaled and edge-enhanced binary image data. Is a re-binarization block as re-binarization means for outputting

The multi-level estimation block 2 receives the input 2
As shown in FIG. 2, the value image data is used for the 3 × 3 block including the target pixel Pm, n and the surrounding pixels as shown in FIG.
A low-pass filter having a weighting coefficient shown in (1) is used to perform multi-valued estimation processing to convert into multi-valued estimation data. Note that the low-pass filter shown in FIG.
From 16 to 15, multi-valued data of 16 gradations can be obtained. Such multi-value estimation processing is performed on all pixels to obtain multi-value estimation data for one page.

The multi-valued estimation data thus obtained is supplied to the scaling block 3. The scaling method in the scaling block 3 is performed by a linear interpolation method, an SPC method, or the like, similarly to the normal scaling method for multivalued data.

As shown in FIG. 5, the edge emphasizing block 4 is provided with a delay circuit 41 composed of a line buffer for synchronizing the pixels and a smoothing circuit 42 as a smoothing means, and the scaling block is provided. The scaled multivalued estimation data from 3 is taken into the delay circuit 41 and the smoothing circuit 42, respectively.

The smoothing circuit 42 filters the 3 × 3 block centered on the pixel of interest Pm, n shown in FIG. 2 with the weighting coefficient shown in FIG. 4, and scales the multivalued estimated data. Is smoothed.

The scaled multi-valued estimation data from the delay circuit 41 and the smoothed scaled multi-valued estimation data from the smoothing circuit 42 are supplied to a subtractor 43 as subtraction means, and the scaled multi-leveled estimation data is supplied. And the smoothed scaled multi-valued estimation data are calculated. Then, the obtained difference data is supplied to the comparator 44 as a comparison means.

The comparator 44 has a threshold value Th which is a constant value.
Is input, and it is determined whether or not the absolute value of the difference is greater than or equal to the threshold Th. If the absolute value of the difference is greater than or equal to the threshold Th, the difference data is supplied as it is to the multiplier 45 as a multiplication means, and if the absolute value of the difference is less than the threshold Th, the data of 0 is supplied to the multiplier 45. I am trying to do it. The multiplier 45 multiplies the input data by K and supplies it to an adder 46 as an adding means.

The smoothing circuit 4 is also included in the adder 46.
Smoothed scaling multi-valued estimation data from 2 is input. The adder 46 adds the data from the multiplier 45 and the data from the smoothing circuit 42 and outputs the result. The data from the adder 46 becomes the output of the edge enhancement block 4 and is supplied to the re-binarization block 5.

As the binarization method in the re-binarization block 5, the systematic dither method or the error diffusion method may be used in consideration of the reproducibility of the pseudo halftone image. Considering the scaling of an image having periodicity such as systematic dither, the binary reproduction by the error diffusion method gives a high-definition output with less moire.

In the embodiment having such a configuration, one page of binary image data is read from the hard disk device 1 and input to the multi-value estimation block 2, and the multi-value estimation block 2 uses a low-pass filter as a filter. Processing is performed and multi-valued estimation data for one page is output.

Subsequently, the multi-valued estimation data for one page is input to the scaling block 3 and scaled to a target scaling ratio. Then, the scaled multivalued estimation data is subsequently input to the edge enhancement block 4.

In the edge emphasizing block 4, the scaled multi-valued estimation data is input to the delay circuit 41 and the smoothing circuit 42, and in the smoothing circuit 42, a 3 × 3 scanning window centered on the target pixel Pm, n is formed. Smoothing is performed by performing the filtering process shown in FIG. 4, and smoothed data P′m, n is output. On the other hand, the delay circuit 41 outputs the original data Pm, n at the same position in synchronism with the output P'm, n from the smoothing circuit 42.

The two data Pm, n and P'm, n
Is input to the subtractor 43 and Pm, n-P'm, n is subtracted. As a result of this subtraction, the edge component is detected. This edge component also includes a noise component in the flat portion of the image. Therefore, in order to remove this noise component, the absolute value of the difference that is the subtraction result is subsequently compared by the comparator 44 with the threshold Th. If the absolute value of the difference is greater than or equal to the threshold value Th, it is determined to be an edge component, and the difference data is output as it is to the multiplier 45 of the next stage. If the absolute value of the difference is less than the threshold Th, it is determined to be a noise component, and the data of 0 is output to the multiplier 45 of the next stage. Thus, the comparator 44 outputs only the edge component from which the noise component has been removed. In the multiplier 45, the difference data from the comparator 44 is further multiplied by K to further emphasize the edge component. The data of 0 is 0 even when multiplied by K.

Thus, the more emphasized edge component is supplied from the multiplier 45 to the adder 46. And adder 4
Reference numeral 6 adds the more emphasized edge component data and the data P'm, n output from the smoothing circuit 42. As a result, the noise component is removed, and the multilevel estimation block 2
The edges of the scaled image that has become blurred can be restored to a sharp image.

The resulting data added by the adder 46 is finally binarized by the re-binarization block 5. Therefore, it is possible to obtain, from the re-binarization block 5, variable-magnification binary image data in which blurring of the edge portion is eliminated and which is excellent in gradation and resolution. Needless to say, the coefficients of the low-pass filter of FIG. 3 and the smoothing filter of FIG. 4 are not limited to the coefficients of the above embodiment.

[0024]

As described above in detail, according to the present invention, it is possible to obtain multivalued estimation data having sharp edges and maintaining resolution, and thus a variable scale binary image excellent in gradation and resolution. It is possible to provide a variable-magnification processing device for a binary image that can obtain data.

[Brief description of drawings]

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

FIG. 2 is a diagram showing a pixel block at the time of multi-value estimation processing in the embodiment.

FIG. 3 is a diagram showing a configuration of a low-pass filter in the example.

FIG. 4 is a diagram showing a configuration of a smoothing filter in the example.

FIG. 5 is a block diagram showing a configuration of an edge emphasis block in the embodiment.

[Explanation of symbols]

2 ... Multi-value estimation block, 3 ... Variable scale block, 4 ... Edge enhancement block, 5 ... Re-binarization block, 42 ... Smoothing circuit, 43 ... Subtractor, 44 ... Comparator, 45 ... Multiplier, 46
… Adder.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Office reference number FI technical display location G09G 5/36 8121-5G

Claims (1)

[Claims] 1. A multi-valued estimating means for performing multi-valued estimation processing on a block including surrounding pixels centering on a pixel of interest with binary image data using a weighting coefficient of a low-pass filter, and converting the multi-valued estimation data Scaling processing means for scaling the multi-valued estimation data from the multi-valued estimation means to a target scaling factor; and edge enhancement means for enhancing only the edge component in the multi-valued estimation data from this scaling processing means. , The edge-enhanced multivalued estimation data from this edge enhancing means is 2
Re-binarizing means for performing binarization processing and outputting scaled binary image data, wherein the edge emphasizing means filters the multi-valued estimation data input from the scaling processing means with a predetermined weighting coefficient. Smoothing means for performing the smoothing, and subtracting means for calculating the difference between the multivalued estimated data input from the scaling processing means and the data smoothed by the smoothing means, and the subtracting means. If the absolute value of the difference is greater than or equal to a certain value, the difference data is output as it is, and if the absolute value of the difference is less than the certain value, 0 data is output. And a multiplying means for multiplying the data from the comparing means by a constant, and an adding means for adding the data from the multiplying means and the smoothed data from the smoothing means. A binary image scaling device.