JPH029268A - Picture processor - Google Patents

Abstract

PURPOSE:To attain optimum interpolation processing in response to kinds of a picture by discriminating kinds of the picture and selecting one of plural interpolation processings in response to the result of discrimination. CONSTITUTION:A discriminator 22 based on a reception signal 101 discriminates any picture of a character, dither, ED(Error Diffusion Method) picture. In the case of discriminating the dither picture, a discrimination signal 107 selects a signal 105. When a character picture is discriminated, the signal 104 selects it. In the case of discriminating the ED picture, the signal 106 is selected. If a picture is not matched with any of the said 3 pictures, the signal 103 is selected. A character interpolation device 14 executes the optimum interpolation method to a character picture, a dither interpolation device 15 applies the dither intermediate tone picture, the ED interpolation device 16 applies the picture subject to binarizing processing such as the error diffusion method. Moreover, the single interpolation device 13 is the 0-order interpolation.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image processing apparatus such as a digital facsimile machine, and more particularly to an image processing apparatus that performs interpolation processing on an image and performs pixel density conversion.

[Conventional technology]

Conventionally, in this type of device, when converting the pixel density to double (for example, from 8 dots/mm to 16 dat/m
When converting the density to m), simply double the binarized image (
After enlarging the area (4 times), the image was output using high-density output equipment. For this reason, image quality could not be improved even though high-density output equipment was used.

As a countermeasure against this problem, various interpolation methods have been proposed in the past. For line drawings such as characters, methods are being considered to interpolate the jaggedness of diagonal lines and output them smoothly using high-density output equipment.

Further, methods for improving the resolution and gradation of halftone images such as dithered images are also being considered.

[Problem that the invention is trying to solve] However,
Conventionally, in these methods, either dithering or character interpolation was performed on all originals sent from the sending side, for example. For this reason, when an image contains a mixture of text images and dithered images, interpolation processing for text may be performed on the dithered image, or interpolation processing for dithering may be performed on the text image, and even if interpolation is performed, image quality may deteriorate. There was a problem with causing it.

[Means and operations for solving the problems] The present invention eliminates the above-mentioned drawbacks, and includes a discriminating means for discriminating the type of image, a processing means for performing a plurality of interpolation processes of different types, and By providing a selection means for selecting one of the plurality of interpolation processes in the processing means according to the determination result of the determination means, the optimum interpolation process according to the type of image is performed.

〔Example〕

An embodiment of the present invention will be described in detail below with reference to the drawings.

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

This embodiment will be explained using a facsimile machine as an example. Image data from the transmitting section 10 of the other party's facsimile machine is received by the receiving section 11 via the communication line 100. When image data is encoded and sent/received, each
Encoder and decoder functions will be added.

The data received by the receiving section 11 is transferred to the line buffer 1.
2, the image is sent to the image judger 22. The data in the line buffer 12 is processed by a simple interpolator 13, a character interpolator 14, a dither interpolator 15, an error diffusion interpolator (hereinafter referred to as an ED interpolator) 16
to go into.

Each interpolator has the function of converting pixel density from low density pixels to high density pixels. The character interpolator 15 is a circuit that performs an optimal interpolation method on character image data. The dither interpolator also performs an optimal interpolation method on the dither halftone image. The ED interpolator performs an optimal interpolation method, such as an error diffusion method, on a binarized image.

The simple interpolator performs O-order interpolation, and converts low-density pixel data vertically and
It is a circuit that can be expanded twice horizontally (four times in terms of area) and can be configured with simple hard logic. These outputs 103.
104. 105. 106 is switched by switch 21, and the selected signal 108 is sent to high-density printer 20, where image formation is performed.

Based on the received signal 101, the determiner 22 determines whether characters, dither,
It is determined whether the image is an ED (Error Diffusion method) image. The determiner 22 is composed of a character determining section 17, and the dither image determining section [8] is constituted by an ED determining section 19.

FIG. 2 is a flowchart showing the order of image determination. Note that this determination is performed by the determiner 22. The judgment flow is as shown in FIG.
, character image determination 23 , and ED both image determination are preferably performed in this order, but the determination method is not limited thereto. When it is determined that the image is a dithered image, the signal 105 is selected by the determination signal 107. When it is determined that the image is a character image, signal 104 is selected. When both ED images are determined, signal 106 is selected. When none of the above three images apply, signal 103 is selected.

FIG. 3 is a block diagram showing details of the dithered image determination section 18.

The signal 101 received by the receiving section 11 enters the line buffer 30, and data for the lines necessary for determination is buffered. - For boat fishing, use a 4x4 type dither such as a Beyer type dither or a halftone type dither as shown in Figures 4 (a) and (b).
Pixel-sized dithering is often used. In this case, determination is made in units of 4×4 pixel blocks. The dot counting section 31 counts the number of black dots in the block and outputs a count signal 111 to the pattern table 32. On the receiving side, since it is unknown which dither was used to perform the binarization, several types of patterns are switched using the switching signal 110 and outputted to the pattern matching circuit 33.

For example, when the number of black dots is 5, patterns such as those shown in FIGS. 4(c) and 4(d) are output.

The pattern matching circuit performs pattern matching between the image signal 113 from the line buffer 30 and the pattern signal 112 from the pattern table 32, and when they match, outputs a determination signal 120 indicating that the image is a dithered image.

FIG. 5 is a block diagram of the character image determination section 17.

The signal 101 enters the line buffer 50, where data for the lines necessary for determination is buffered. Similar to the dither image determination section, the number of black dots in each 4×4 block is counted by a dot counting section 51, and the result is stored in a line buffer 52.

Output signal 114 from line buffer 52. At 115° and 116, signals delayed by one line of dot count data are output. This signal is applied to the latch 54a~
Adjust the timing with 54g and latch 54 to the judge 53.
The outputs of b, 54c, 54d, 54e, and 54g are sent.

The determiner 53 calculates a difference signal as shown in the following equation. In addition, the output from each latch is (latch 54d)
Write it like this.

D= (latch 54d) X4- (latch 54b)
-C latch 54c) -(latch 54e) -(latch 54g) or D= max] (latch 54b) -(latch 54
d) l, l (latch 54C) (latch 54d)
l, l (latch 54g) - (latch 54d) (
Latch 54e) - (Latch 54d,)l) When the difference signal calculated by the above equation exceeds the set value, a determination signal 121 indicating that the image of the block corresponding to the latch 54d is a character image is output. .

Further, when the determination signal 121 is not a character, an inverted signal 117 thereof is generated by the inverter 55.

Further, the output of the latch 54d enters the comparator 55 and is compared with a specific threshold value, and when the output is equal to or greater than the threshold value, the signal 117 becomes 1. In the AND circuit 57, the signal 117 and the signal A of 18
The ND is taken and the result becomes the decision signal 122.

A signal 117 is a signal indicating a non-edge portion, and a signal 118 is a signal indicating that there is a dot exceeding a threshold value within the block.

By combining this with the determination signal 120, when the image is not a dithered image (determination signal 120=o) and when the number of dots is large in the non-edge area (determination signal 122=1), a halftone image with many dots, i.e. A judgment is made regarding both ED images.

The above can be summarized as follows.

Signal 120 Signal 121 Signal 122
Judgment 1 × ×
Dithered image 0 1 0
Character image 0 0 1
ED image 0 0 0 undetermined FIG. 6 is a diagram for explaining the character interpolation method.

Figure 6(a) shows a low-density pixel block, and Figure 6(b) shows a low density pixel block.
) indicates a high-density pixel block, and when processed using the simple interpolation method, the dots shown by hatching in the figure are constructed.

In the character interpolation of the present invention, as shown in FIG. 3(c), the four pixels 60.degree. 61, 62.63 to be interpolated from pixel M are determined from three adjacent pixels.

For example, the pixel 60 is (AnBn
C) ■M=1 or (λn B n c) ■M=1 −
・-------When (1), the pixel becomes 60-■. However, the opening indicates the logical AND, and the symbol ■ indicates the exclusive OR18 indicates the complement of A.

That is, when M is a white pixel (=O) and A, B, and D are black pixels (-1), the pixel 60 becomes black (=1).

Also, M is a black pixel (=1) and A, B, and D are white pixels (=
0), the pixel 60 becomes white (=0). Similarly, the same figure (
e).

(f), (g)l: As shown, pixel 61 is B,
Pixel 62 is from C and E, and pixel 63 is from F and G.

Determine from H and G.

FIG. 7 is a block diagram of the character interpolation circuit. signal 101
The line buffer 70 holds data for the number of lines required for interpolation. This data is processed by interpolators 71a and 71b.
, 71c, and 71d. 71a is shown in Figure 6 (d
) The pixel data of AB, D, and M shown in ) are input, and the data of pixel 60 is output.

Similarly, 71b is shown in FIG. 6(e), 71c is shown in FIG. 6(f),
7] d receives and outputs data corresponding to FIG. 6(g). The interpolation method is shown in equation (1) above, and can be configured with simple hardware logic.

Interpolators 71a, 7], b, 71c, 71d
The output of the dot alignment controller 72 inputs the address of each corresponding pixel position and stores it in the line memory 7.
It is stored in 3. The output from here becomes signal 104.

FIG. 8 is an explanatory diagram of a dither image interpolation method.

FIG. 5(a) is a diagram expressing five levels of halftones using Heyer type dithering. When conventional simple interpolation processing is applied, the same figure (b
). Another method is to enlarge the image twice in units of 4×4 blocks, which is shown in FIG. 4(c). In this method, the pixel density becomes finer, but 8X8 (4X
2) There is a drawback that the block texture of each unit is noticeable.

This method shifts the phase between blocks of low-density pixels and high-density pixels to perform re-dithering, as shown in FIG. 9(b). The solid lines in the figure represent 4x4 blocks of low-density pixels, and the broken lines represent 4x4 blocks of high-density pixels.

In addition, FIG. 5A shows the results of calculating the number of dots in a 4×4 block of low-density pixels (3, 4, 5, 6°7). This value is taken as the average density level of this block. Next, as shown in the same figure (b), the average value for each high-density dust 4x4 pixel block is recalculated. This is block 90 =
(3+4+4+5)=4, block 91=% (4+
5) = 4.5, block 92 = z (4 + 5) = 4.5
, block 93:5 is calculated.

The result of re-dithering based on this average value is shown in FIG. 8(d). The density is higher than that in the same figure (b), and the same figure (c)
) 8x8 block textures also do not exist.

Taking this method further, the decimal point that occurs during calculation is 1〕L
If the lower fraction is carried forward when calculating the average value of the adjacent block, accurate density conservation will be achieved and the image quality will further improve.

FIG. 10 is a block diagram of the dither image interpolator 15 of FIG. The signal 101 enters the line buffer 80, and then the dot count section 81 counts the number of dots in the 4×4 block. This data is sent to the in-memory 82 and latched to the latch 83 as a density level signal for the low density block.
a, 83b, 83c, and 83d. This online data is shown in Figure 9(a)l: 3, , 1. ．． 1
．． This corresponds to data 1 of 5.

The stiffness data enters an allocation determiner 84, where an allocation ratio is calculated according to the block position signal, and block average value data is sent to a comparator 86. Threshold data is sent from the dither matrix and binarized by a comparator 86.
The result is sent to line memory 87. This output becomes high-density pixel data +05.

FIG. 11 is a block diagram of the ED interpolator 16.

The signal 101 enters the buffer 401, and then smoothed data is calculated in the smoothing circuit 41.1 For example, using a 5×5 smoothing filter as shown in FIG. The data is for the center pixel (hatched area in the figure).

Next, the smoothed data is stored in the line buffer 42. Using this line buffer, pixels are enlarged twice by the same process as simple interpolation, and the enlarged data is binarized by a 2&i conversion circuit 43. As this binarization method, error diffusion method, dither method, etc. are suitable.

The binarized data is input to the line memory 44, the output of which becomes the high-density pixel signal 106.

Although the embodiments of the present invention have been described with respect to two-fold enlargement of an image, it goes without saying that the present invention can also be applied to enlargement/reduction at other magnifications.

As described above, by identifying the type of binary image and switching the optimal interpolation process accordingly, it is possible to obtain an image with higher quality on the receiving side output device than the low density image on the transmitting side.

Note that this embodiment has been explained using a facsimile machine as an example.
The present invention can also be used in digital copiers.

〔Effect of the invention〕

As described above, according to the present invention, it is possible to perform optimal interpolation processing depending on the type of image, and it is possible to obtain a high-quality image.

[Brief explanation of the drawing]

FIG. 1 shows one embodiment of the present invention, and the block diagram shown in FIG.
Figure 3 is a block diagram of a dither image determination circuit, Figure 4 is a diagram showing Bayer type/halftone dithering and pattern images at a specific level, Figure 5 is a character image determination process flow diagram. A block diagram of the circuit, Fig. 6 is an explanatory diagram of the character image interpolation method, Fig. 7 is a block diagram of the character image interpolation circuit, Fig. 8 is an explanatory diagram of the dither image interpolation method, and Fig. 9 is an explanatory diagram of the dither image interpolation method. Another explanatory diagram, FIG. 10 is a dither image interpolation block diagram,
FIG. 11 is an interpolation block diagram of the El'1 image, and FIG. 12 is an example diagram of a smoothing filter.

Claims (1)

[Claims] A discriminating means for discriminating the type of image, a processing means for performing a plurality of interpolation processes of different types, and a selection means for selecting one of the plurality of interpolation processes in the processing means according to a determination result of the discriminating means. An image processing device comprising: