JPH045303B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an image conversion method when transmitting an image signal obtained by scanning an original by reducing the number of pixels.

For example, in a facsimile device, when transmitting an image from a transmitter with a main scanning line density of 8 pixels/mm to a receiver with a main scanning line density of 6 pixels/mm, the transmitter side reduces 4 pixels to 3 pixels. It is necessary to reduce the number of pixels.

You can also reduce the size of the original (for example, reduce the size of a B4 size original).
If the transmitter has a main scanning line density of 8 pixels/mm, the number of pixels for one scanning line must be reduced from 2048 to 1728 before transmission. .

Now, taking as an example the case of converting the above-mentioned 8 pixels/mm to 6 pixels/mm, there is a method of simply reducing the fourth pixel.

That is, FIG. 1 is an explanatory diagram illustrating an example of a conventional image conversion method, and an example of converting the image signal of a into the image signal of a-1 is the image conversion method using the fourth pixel reduction method.

In the figure, a ₁ to a ₁₁ are added to each pixel constituting the image signal before conversion from the left for explanation, and a ₁ -1 to a ₁₁ -1 are added to each pixel before conversion. These are the corresponding pixels after conversion, and the shaded pixels are shown to distinguish black pixels.

As is clear from this figure, in the case of a method that simply reduces the fourth pixel, the run length (run length) of pixels a ₄ and a ₈ (black pixels in this case)
One thin line (hereinafter abbreviated as RL) is missing, causing a decline in image quality.

Note that the image signal of a in _{FIG .}
A thin line of RL3 consisting of a ₃ , pixels a ₄ , a ₅ , a ₆ , a ₇ , a ₈ , a ₁₁ which are all thin lines of RL1, and pixels a ₉ ,
A thin line of RL2 consisting of a ₁₀ is shown.

Other methods include a method in which pixels are reduced using pixels other than RL1. In this case, as shown in b of FIG. 1, a pixel signal in which RL1 is a continuous thin line is converted into b-1.

In FIG. 1, the numerical value below the pixels from a ₁ -1 to a ₁₁ -1 or from b ₁ -1 to b ₁₁ -1 represents the amount of image distortion. For example, if b ₁ is converted to b ₁ −1, −
This indicates that distortion of 1/4 pixel occurs in the received image. When the thin line of RL1 is continuous as shown in Figure 1b, there are no pixels other than the thin line of RL1 in the continuous range of RL1, so pixels cannot be reduced, so image distortion accumulates and increases. This results in a large received image, which does not fully satisfy the image quality (in the case of a-1 in image signal a, pixels a ₄ and a ₈ are reduced and there is no -4/4 image distortion ).

The present invention has been made in view of the above points, and an object of the present invention is to reduce the number of pixels and transmit the RL1.
In turn, while reducing the loss of thin lines in the entire RL,
An object of the present invention is to provide an image conversion method that can reduce image distortion and maintain good image quality.

In order to achieve the above object, the present invention proposes the following image conversion method.

In other words, in image conversion that reduces the image signal obtained by scanning a document to the number of pixels n/m (n<m), it is possible to perform image conversion by exclusively reducing pixels in thin lines other than RL1. If the image distortion is within the set tolerance limit, perform image conversion by reducing pixels in the thin lines other than RL1 without reducing the thin line in RL1,
Due to a large number of continuous thin lines of RL1, the RL1
If image distortion exceeds the above-mentioned allowable limit only by image conversion by reducing pixels in thin lines other than
The thin line of RL1 is also targeted for reduction, and this pixel reduction of RL1 reduces the pair of white and black thin lines.

According to such an image conversion method, an image signal in which many thin lines of RL1 are not continuous, in other words, even if the thin lines of RL1 are continuous, if the image distortion is within the permissible limit, only thin lines other than RL1 are used. Perform pixel reduction. Then, due to the image signal with many consecutive RL1s,
Only when it is unavoidable that the image distortion of the thin line in RL1 exceeds the allowable limit, is that RL1 targeted for pixel reduction. Therefore, it is possible to perform image conversion with less image distortion and fewer missing thin lines in RL1.

In addition, when reducing a large number of consecutive RL1 lines, the thin lines of RL1 that are paired with white and black are removed, so the thin lines of RL1 that remain after the reduction can maintain a relationship in which white and black lines are alternately continuous. , one by one as a result
By reducing RL1, the loss of thin lines can be reduced, and in turn, the loss of fine lines in the entire RL can be reduced (see b-2 and b in Figure 2 below).
Please refer to the detailed explanation in Comparison of Image Conversion Methods in Section 3).

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

FIG. 2 is an explanatory diagram including an explanation of the principle of the image conversion method according to the present invention in the case where four pixels are reduced to three pixels for conversion as an example.

Here, an example in which the allowable limit of image distortion is set to -1 pixel will be described.

Converting 4 pixels to 3 pixels as described above means multiplying the RL included in the image signal before conversion by 3/4 and transmitting the result as the RL after conversion. This results in a decimal remainder.

Therefore, the remainder created by converting the previous RL is added to the RL before conversion multiplied by 3/4, the integer part is transmitted as the converted RL, and the remainder of the decimal part is used as image distortion for the next image. The method is to add it to the RL of .

The image distortion of each RL in the image signal after conversion is as follows: If the number of pixels of each RL before image conversion is M, the number of pixels after conversion is N, and the reduction ratio is n/m (n<m). , M×(n/m)−N, which will be added. For example, if the RL before the change is 3 and the number of pixels after the change is 2, the image distortion itself will be +1/4 (this means that the integer 2 is transmitted as the pixel after the change, and the remainder is +1 /4 means image distortion). In addition, even if RL1 before image conversion remains with a pixel count of 1 after image conversion (this means that RL1 is preferentially left and transmitted, and thin lines other than RL1 are targeted for reduction), RL1
Image distortion also occurs due to the influence of the overall image conversion,
According to the above calculation formula, the image distortion of RL1 itself is −1/
4, negative distortion results.

Therefore, when a large number of thin lines of RL1 are continuous, negative distortion increases, and here, the allowable limit thereof is set to within -1 pixel. That is, when the present invention described above is applied to this embodiment, first, even if the pixels in the thin lines other than RL1 are exclusively reduced and the image is converted by 3/4, the image distortion of the thin line in RL1 will not accumulate. If it is within -1 pixel, RL1
Pixel reduction in thin lines other than RL1 will be performed without reducing the thin lines in . This image conversion is illustrated in a and a-2 of FIG.

A in FIG. 2 is an image signal before conversion (same as a in FIG. 1), which becomes a-2 after conversion. Here, the numerical value below each pixel after conversion in FIG. 2 represents the integrated value (amount) of image distortion. In the case of FIG. 2, the thin line of RL1 does not continue until the image distortion exceeds -1 pixel.

According to such image conversion, an image signal a-2 with fewer missing thin lines of RL1 can be obtained than the conventional image signal a-1 in FIG. Specifically, as shown in a in Figure 2, a thin line of RL3 consisting of a ₁ , a ₂ , and a ₃ ,
A ₄ RL1 thin line, A ₅ RL1 thin line, A ₆ RL1 thin line, A ₇ RL1 thin line, A ₈ RL1 thin line, A ₉ and A ₁₀
In the case of a pixel signal where many RL1s are not consecutive, such as a thin line of RL2 with a ₁₁ and a thin line of RL1 with a 11, as in a-2, the combination of a ₁ -2 and a ₂ -2
Thin line of RL2, thin line of RL1 of a ₄ -2, RL1 of a ₅ -2
Thin line of RL1 of a ₆ -2, thin line of RL1 of a ₇ -2, thin line of RL1 of a ₈ -2, thin line of RL1 of a ₉ -2,
The number of thin lines is the same, such as the thin line of RL1 of a ₁₁ -2, that is, transmission is possible without missing thin lines, and images with good reproducibility can be transmitted.

Furthermore, in this embodiment, if a large number of consecutive thin lines in RL1 cause image distortion (integrated value) to exceed -1 pixel by image conversion by reducing pixels in thin lines other than RL1, Thin lines with RL1 exceeding the allowable limit are also subject to reduction. This RL1
Pixel reduction reduces the thin lines in RL1 of the white and black pairs.

Figure 2b shows the case where many RL1s are consecutive (Figure 1b
similar), RL1 where b-2 exceeds -1 pixel
one by one (here, each RL1 of b5 and b9)
In b-3, the white and black pair (in this case, the RL1 pair of b4 and b5) is reduced.

Case b-2 in Figure 2 has the advantage that even if many RL1s are consecutive, the image distortion after conversion will not be large compared to b-1 in Figure 1, but the white RL1 pixel is missing. Put it away.

On the other hand, in the case of b-3 in FIG. 2, even if the thin line of RL1 is to be reduced, the white and black RL1 are alternately continued while suppressing the image distortion within the permissible limit.

In other words, as you can see by comparing b-2 and b-3 in Figure 2, the amount of distortion is the same, and b-2 is better.
b ₁ −2 RL1 thin line, b ₂ −2 RL1 thin line, b ₃ −
Thin line of RL1 of 2, RL2 consisting of b ₄ -2 and b ₆ -2
thin line of RL1 of b ₇ -2, b ₈ -2 and b ₁₀ -2
The number of thin lines is 7, such as the thin line of RL2 with b ₁₁ -2 and the thin line of RL1 with b 11 -2.
For -3, thin line of RL1 of b ₁ -3, RL1 of b ₂ -3
Thin line of RL1 of b ₃ -3, thin line of RL1 of b ₆ -3, thin line of RL1 of b ₇ -3, thin line of RL1 of b ₈ -3,
b ₉ -3 RL1 thin line, b ₁₀ -3 RL1 thin line, b ₁₁
The number of thin lines is nine, such as the thin line of RL1 of -3, and the reproducibility of the thin lines is better in b-3 than in b-2.

Therefore, according to this embodiment, it is possible to transmit an image with good image reproducibility by eliminating image distortion and reducing the loss of thin lines in RL1, and even in the entire RL, and to maintain good image quality.

FIG. 3 is a block diagram of an embodiment of a facsimile transmitter for implementing the invention described above.

In the figure, reference numeral 10 denotes a reading device that scans a document and converts information on the document into binary white and black image signals.

20 is a line memory that temporarily stores the image signal, and is a random access memory (Raudom).
It consists of (Access Memory, RAM), counters, gates, etc.

30 scans the above line memory 20,
A run length calculation device that calculates the run length (RL), which is the number of consecutive white and black pixels, and is composed of a flip-flop, a counter, a gate, etc.

Reference numeral 40 denotes an image processing device that performs image conversion processing, encoding processing, etc. according to the present invention, and is composed of a microcomputer and the like.

50 is a modulation/demodulation device, and 60 is a network control device.

Next, FIG. 4 shows the image processing device 4 of FIG.
0 is a program diagram showing an example of a program for performing the image conversion process b-3 shown in FIG. 2 above.

This program will be explained.

In the process 100, the run length calculation device 30
Enter the RL and register it as RL(A).

Judgment 110: RL(A) is 1 and the remainder is -3/4
In this case, the process moves to process 130, and in other cases, the process moves to process 120.

Here, the above remainder -3/4 means RL1
This represents three consecutive thin lines.
In other words, moving to process 130 means that the thin line of RL1 is 4.
It represents a series of consecutive items.

In process 120, RL(A) is registered as RL(U) as the image signal to be subjected to image processing next, and the process moves to process 190.

On the other hand, in process 130, the next RL is input and registered as RL(B), and the process moves to determination 140.

In this determination 140, it is determined whether RL(B) is 1. If RL(B) is 1, the process moves to process 170; otherwise, the process moves to process 150.

In process 150, RL(A), which is a thin line of RL1, is encoded, and the remainder corresponding to image distortion is set from -3/4 to -1.

This is because RL(B), which is the next RL after RL(A), is not the thin line of RL1, so even if the thin line of RL(A), which is RL1, is transmitted with priority, the image distortion equivalent to -1 pixel will occur. R.L.
This is because it is corrected during processing (B).

Here, the encoding process mentioned above is, for example, a modified Huffman code, which is an international standard code for high-speed facsimile.
This refers to the process of encoding RL.

Returning to the beginning, in process 160, RL(B) is registered in RL(U) as the RL to be processed next, and the process moves to determination 190.

In process 170, the next RL after RL(B) is input, this is registered as RL(C), and the process moves to process 180.

In process 180, RL(A) and RL(B) are added to RL(C),
This is registered in RL(U) as the RL to be processed next, and the process moves to determination 190.

This corresponds to forcibly deleting the thin lines of RL(A) and RL(B) in order to suppress image distortion.

In determination 190, the RL obtained by performing the process of converting 4 pixels into 3 pixels is obtained for RL(U), and if this is 1 or more, the process moves to process 200, otherwise the process moves to process 210.

That is, in determination 190, it is determined whether the result of multiplying RL(U) by 3/4 and adding the remainder is 1 or more.

In process 200, the result of the above process is set as RL' and a remainder, and the process moves to process 220.

However, in process 200, the remainder is set to a positive value smaller than 1.

On the other hand, in process 210, from the standpoint of prioritizing prevention of thin line reduction, the result of converting 4 pixels into 3 pixels in RL(U) is forcibly set to 1 and registered in RL', and the remainder is set as negative.

In process 220, RL' is encoded and the process ends.

By performing processing from the beginning to the end of this series of programs, it is possible to perform image conversion with fewer missing thin lines and less image distortion.

In the above embodiment, b-3 in FIG.
Although I have only shown an example of a program for the image conversion method of
It is possible to realize the following image conversion method.

Further, image conversion methods according to the present inventions in a conversion mode in which pixels other than the conversion from 4 pixels to 3 pixels are reduced can also be created as a similar program.

Furthermore, as can be seen from the program example shown in Figure 4, the encoding process and image conversion process mentioned above can be performed in parallel, so there is no need to add a special device for image conversion process. This has the effect that there is no such thing.

Note that various allowable limits for image distortion can be selected depending on the type of image, user's request, etc.

According to the present invention, including the details described above, it is possible to provide an image conversion method capable of performing image conversion with less image distortion and fewer missing thin lines, and which is excellent and practical. It can be said that this invention is effective.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram illustrating a conventional image conversion method;
FIG. 2 is an explanatory diagram of each embodiment including an explanation of the principle of the image conversion method according to the present invention, and FIG. 3 is a block diagram of an embodiment of a facsimile transmitter used for implementing the present invention. FIG. 4 is a program diagram showing an example of a program for the image processing apparatus shown in FIG. 3 to perform image conversion processing. 10...Reading device, 20...Line memory, 30...Run length calculation device, 40...Image processing device, 50...Modulation/demodulation device, 60...Network control device, a, b, a-2, b- 2, b-3... Image signal, a ₁ to a ₁₁ , b ₁ to b ₁₁ , a ₁ -2 to _{a 11} -2, b ₁
-2 to b ₁₁ -2, b ₁ -3 to _{b 11} -3...pixel.

Claims (1)

[Claims] 1. In image conversion that reduces the number of pixels of an image signal obtained by scanning a document by a factor of n/m (n<m), pixels in thin lines other than run length 1 are exclusively used. If the image distortion is within the set allowable limit even if the image is converted by reducing the run length, image conversion is performed by reducing pixels in the thin lines other than the run length 1 without reducing the thin line of run length 1. If a large number of continuous thin lines with a run length of 1 result in image distortion exceeding the permissible limit by image conversion by simply reducing pixels in the thin lines other than the run length 1, the thin line with a run length of 1 is also targeted for reduction; This pixel reduction with a run length of 1 is an image conversion method characterized by reducing a pair of white and black thin lines. 2. The image conversion method according to claim 1, wherein the permissible limit of the image distortion is -1 pixel.