JPH0453142B2 - - Google Patents

Description

[Detailed Description of the Invention] (Industrial Application Field) The present invention relates to an image processing method for high-resolution display of an image encoded by a two-dimensional sequential encoding method such as a modified read (MR) method. It is.

(Prior art) Regarding facsimile, the Consultative Committee on International Telegraph and Telephone (CCITT) has recommended resolution, encoding method, etc. Takahiko Fukinuki, "Image Signal Processing for FAX and OA," first edition (October 20, 1982), Nikkan Kogyo Shimbun, p.69-71, p.79-83, p.165. According to the above recommendation, international standard G3
For facsimile (Group3FAX), main scanning resolution is 8 pixels/mm, sub-scanning resolution is 7.7 lines/mm or
3.85 lines/mm, the Modified Human (MH) method is recommended as a one-dimensional encoding method, and the Modified Read (MR) method is recommended as a two-dimensional encoding method.

However, recently, the resolution has been further increased,
There is an increasing demand for a less redundant and more efficient encoding method that can improve the quality of reproduced images and uses highly reliable communication lines. To meet this requirement, we have doubled the resolution in both main and sub-scanning to 400 pixels/inch (approx.
15.7 pixels/mm), extend the two-dimensional encoding method as an encoding method, and set the K parameter to infinity (extended MR method), as described in the Journal of the Institute of Image Electronics Engineers, "CCITT SG Conference Report" 11 [1] ( 1982), p.37
and the same magazine “CCITT SG Joint Working Group Report” 11
[3] (1982), p. 196. However, as stated in the 1981 IEICE General Conference, ``Compression Effects of MR Methods on Resolution,'' p. 168, increasing the resolution improves the image quality, but increases the number of transmission bits. Therefore, both transmission and reception have a reading and recording resolution of 400 pixels/inch (approximately 15.7 pixels/mm), and 200 pixels/inch when encoding documents or parts of documents that do not require detailed reproduction. A method of encoding with a resolution of (approximately 7.9 pixels/mm) is considered. However, with this method, when reproducing an image encoded with a resolution of 200 pixels/inch (approximately 7.9 pixels/mm) at 400 pixels/inch (approximately 15.7 pixels/mm), data for one pixel is simply If both the main scanning and sub-scanning are expanded to two pixels, the stepped binarization error indicated by diagonal lines will be emphasized.

As a method to improve this point, the method described in Takahiko Fukinuki, "Image signal processing for FAX, OA", first edition (October 20, 1981), Nikkan Kogyo Shimbun, p. 35-37. There is. This method converts the original image into two
Divide into partial patterns of about ×2 or 4×4,
This is matched with a pre-prepared pattern, and if the two match, the pre-prepared n×n (to double the resolution, n
are replaced with patterns 4 to 8), respectively.

(Problems to be Solved by the Invention) However, when performing image processing with improved image quality using the method of the prior art as described above, it is necessary to It is necessary to perform processing by pattern matching. Therefore, the processing is performed in two stages, resulting in an increase in the amount of hardware or software processing time.

The present invention has been made in order to eliminate such drawbacks of the prior art, and it is an object of the present invention to provide an image processing method that can easily obtain good high-resolution image quality.

(Means for Solving the Problems) The image processing method of the present invention is such that when signals obtained by scanning an original image are sequentially sampled and converted into binary signals, the values change from one value to the other. Decoding is performed based on the codes of the pass mode, horizontal mode, and vertical mode created by encoding the changed pixel position that changed to the value, and
During decoding, detecting the mode-specific code indicating the displacement of the changed pixel position of the decoded row to be decoded with respect to the changed pixel position of the reference row immediately before the decoded row,
In the image processing method, the detected code is inserted between the reference line and the decoded line according to the magnification rate, one or more interpolation lines depending on the detected mode-specific code. mode, the horizontal mode, or the vertical mode, when the amount of change is 0, the contents of one or more interpolation lines inserted between the reference line and the decoded line according to the enlargement ratio are The content in the pixel range of the decoded row related to the mode is adopted as is, and the detected code is in the vertical mode and the displacement amount is 0.
If the number is 3 or less, including 3, then the contents of one or more interpolation lines inserted between the reference line and the decoded line according to the expansion rate, depending on the displacement amount and expansion rate of each. It is characterized by employing a predetermined pattern.

(Function) According to the present invention, since the image processing method is configured as described above, it functions as follows.

That is, by inserting one or more interpolation rows between the reference row and the decoded row, the data in the interpolation row acts to correct the discontinuity between the reference row and the decoded row, and the encoding Good high-resolution image quality is provided when the resolution during decoding is greater than the resolution during decoding and when enlarged display is performed with the same pixel density. Therefore, the problems of the prior art described above can be solved.

(Example) Figure 1 shows a first example to which the image processing method of the present invention is applied.
FIG. 2 is a block diagram of a circuit according to an embodiment of the present invention. here
400 pixels/inch (approx. 15.7 pixels/mm), 3456 pixels/inch
line, 400 lines/inch (approx. 15.7 lines/mm) image vertically,
The following explanation assumes that the data is thinned out in half horizontally and encoded, then enlarged twice and reproduced. In addition, in the same scanning line, a pixel whose color (white or black) has changed from the previous pixel is defined as a changed pixel, and the five types of changed pixels a ₀ , a ₁ , a ₂ , b ₁ , b ₂ are Define it as follows.

a ₀ : Reference change pixel or start change pixel in the scan line to be encoded (hereinafter referred to as encoded scan line). Its position is determined by the previous encoding mode. At the start (left end) of the encoded scan line, Let a ₀ be a pixel virtually placed immediately before (to the left) of the first pixel.

a ₁ : Next change pixel to the right of a ₀ in the encoded scan line. This is the opposite color to a ₀ and is the next change pixel to be encoded.

a ₂ : The next change pixel further to the right of a ₁ in the encoded scan line.

b ₁ : The next changing pixel to the right of a ₀ in the scanning line immediately before the encoding scanning line (hereinafter referred to as the reference scanning line). Same color as a ₁ .

b ₂ : Next change pixel to the right of b ₁ in the reference scan line.

If there are no a ₁ , a ₂ , b ₁ , b ₂ other than a ₀ , then
These pixels are assumed to be pixels immediately after (to the right) the last (rightmost) pixel of each scanning line.

Moreover, three modes in this method are defined as follows.

(i) Pass mode (hereinafter referred to as P mode): When b ₂ is located to the left of a ₁ .

(ii) Vertical mode (hereinafter referred to as V mode): | a ₁ b ₁
| In case of 3.

(iii) Horizontal mode (hereinafter referred to as H mode): | a ₁ b ₁
| In case of 4.

It is expressed as H+M(a ₀ +a ₁ )+M(a ₁ a ₂ ). however,
M is a MH (Modified Huffman) code word.

In FIG. 1, 101 is the received extended MR
102 is a control circuit that controls decoding and interpolation, 103 is a reference row memory that has a capacity of 1728 pixels and stores reference row data, 104 is a detection circuit that detects b ₁ and b ₂ , and 105 is reproduced image data. a pixel reproducing unit 106 that writes the data into the decoding row memory 106;
107 has a capacity of 1728 pixels and stores decoded row data, and 107 has a capacity of 3456 pixels and stores the contents displayed in the middle row between the reference row and the decoded row (hereinafter referred to as the interpolation row). This is an interpolated row memory.

The operation of this embodiment will be described below with reference to the flowchart of FIG.

In the initial state of each row, the position of a ₀ is virtually to the left of the leftmost pixel, and the color is white (step
201). Next, the detection circuit 104 detects the reference row memory 1.
The data stored in pixel 03 is read out and the pixel _b1 is searched for (step 202). Then, it decodes which of the three modes (P mode, V mode, H mode) the received MR code is (step
203, step 204).

In the case of P mode, the detection circuit 104 detects _b2 , which is the next change point, from the data stored in the reference row memory 103 (step 205). Then, the pixel from a ₀ in the decoded row to the pixel at the same position as b ₂ is reproduced as the same color as a ₀ (step 206). Then, using the pixel reproducing unit 105, the reproduced image data is written into the decoding row memory 106. Next, _a0 is changed to the pixel position of _b2 for the next decoding (step 207).

In the case of H mode, the position of a ₁ is calculated by decoding the run length following the H mode identification code and adding it to the current a ₀ (step 208). Then, the area from a ₀ to the pixel to the left of a ₁ is reproduced as the same color as a ₀ (step 209), and the pixel reproduction unit 105 writes the reproduced image data into the decoding row memory 106. Next, the second run length of H mode is decoded and added to a ₁ to calculate the position of a ₂ (step 210). Then, the pixels between a ₁ and the pixel on the left of a ₂ are reproduced as the opposite color of a ₀ , that is, when the pixel at the position of a ₀ is white, it is reproduced as black, and when it is black, it is reproduced as white (step 211). The reproduction means 105 writes the reproduced image data into the decoding row memory 106. Next, for the next decoding, a ₀ is changed to the pixel position of a ₂ and made the same color as the previous a ₀ (step 212).

In the case of the above-mentioned P mode and H mode, when enlarging the image twice vertically and horizontally, the same pixel is simply displayed twice, and no special processing for interpolation is added.

Here, the positional relationship between the interpolated row, the reference row, and the decoded row is shown in FIG. In the initial state of page 1, reference row memory 1
03, it is assumed that both the interpolation row memory 107 are all in a white state. Every time one row of decoding is completed, the 1728 pixels in the decoding row memory 106 are expanded twice in the main scanning direction, transferred to the interpolation row memory 107, and written to 3456 pixels in the memory 107 (step 213). . Therefore, if the result of decoding the next row is P mode or H mode, the contents of interpolation row memory 107 are the same as decoding row memory 1 at the time of decoding the previous row.
The contents of 06 are enlarged.

In the case of V mode, the distance of a ₁ b ₁ is decoded from the received code, and the position of a ₁ is calculated (step 214).
Next, as shown in Figures 4 and 5, the color of a ₀ and a ₁ b ₁
The contents of the interpolation row memory 107 are corrected according to the distance between the
215), and the area from a ₀ to the pixel on the left of a ₁ .
It is reproduced as the same color as a ₀ (step 216) and written to the decoding line memory 106. Next, for the next decoding, _a0 is changed to the pixel position of _a1 , and the color is set to the opposite color (step 217). The above operations are repeated until the decoding of 1728 pixels per row is completed (step 218). 1
When the row decoding is completed, first the interpolation row memory 107
The 3456-pixel image is displayed (step 219), and then the 1728-pixel image in the decoding row memory 106 is enlarged by 2 pixels in the main scanning direction, and the 3456-pixel image is displayed (step 220). . Next, in preparation for decoding the next row, the contents of the decoded row memory 106 are transferred to the reference row memory 103 (step 221), and then the contents of the decoded row memory 106 are expanded twice as described above, and the interpolated row memory is Transfer to 107 (step
213).

The above operations are repeated until the decoding for one page is completed, that is, until the page end signal is decoded (step 222).

As described above, in this embodiment, in V mode, except for the case where a ₁ - b ₁ = 0, the change point in the interpolation row is set to be midway between the change points in the reference row and the decoded row. (Figures 4 and 5). As a result, the image displayed at high resolution can reduce the jagged appearance of the shaded areas. Figure 6 is a sampled image of CCITT facsimile test chart No. 1 at approximately 200 pixels/mm, and a portion of it has been enlarged approximately 8 times. On the other hand, FIG. 7 is an image obtained by correcting the same image only for the V mode using the method of the above embodiment.

Next, a second embodiment of the present invention will be described.

In the first embodiment, one interpolation row is inserted between the reference row and the decoded row in order to double the resolution, but in the second embodiment, a plurality of interpolation rows are inserted.

FIG. 8 is a block diagram of the circuit of the second embodiment, in which 801 is the received extended MR.
802 is a control circuit, 803 is a reference row memory, 804 is a detection circuit, 805 is a pixel reproduction unit, 8
06 is a decoding row memory. These elements are the first
Similar operations are performed corresponding to elements 101 to 106 in the figure. 1728×8= for each of 807 to 813
Interpolation row memory (#0
~#6).

In this embodiment, when the V mode is detected, for example, when a ₀ is white and a ₁ - b ₁ = 3, the contents of interpolation row memories #0 to #6 are changed as shown in FIG.
Interpolation is performed with eight times the resolution both vertically and horizontally. FIG. 9 is a diagram illustrating a method of interpolating between corresponding change points of a reference row and a decoded row using a linear function. Note that the figure shows the case where a ₀ is white. This method will be generalized and described. If we consider the case where l interpolated rows are inserted between the reference row and the decoded row to increase the resolution by m times, the amount of displacement at the change point between the reference row and the decoded row is n pixels at the encoding resolution. At some point, the i number (1
The change point of the interpolated row of il) is displaced from the reference row by the integer closest to m×n×i÷(l+1) in the number of pixels after high resolution. In addition,
n, m, i, l are positive integers.

FIG. 10 shows an image in which the resolution of the image in FIG. 6 has been made eight times higher using the method of the second embodiment.

(Effects of the Invention) As explained in detail above, according to the present invention,
There is no need to perform logical judgment such as pattern matching for interpolation, which is done in conventional enlargement methods, and good high-resolution image quality can be obtained with a smaller amount of hardware or simpler software. Therefore, this method is effective when displaying an image encoded at a low pixel density at a high pixel density, or when enlarging and displaying an image at the same pixel density.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a circuit according to a first embodiment of the present invention, FIG. 2 is a flowchart showing an operation sequence of the circuit according to the first embodiment, and FIG. 3 is a diagram explaining the position of interpolation rows. Figures 4 and 5 are diagrams explaining pixel patterns of interpolated rows, Figure 6 is a diagram showing an example of an enlarged image without interpolation, and Figure 7 is a diagram showing the interpolation of one row according to the first embodiment. FIG. 8 is a diagram showing an example of an image displayed at twice the resolution in the second embodiment of the present invention.
9 is a diagram explaining the pixel pattern of the interpolated rows, and FIG. 10 is an image displayed at 8 times the resolution by interpolating 7 rows according to the second embodiment of the present invention. It is a figure which shows an example. 101,801...MR code, 102,802
...Control circuit, 103,803...Reference row memory, 104,804...Detection circuit, 105,80
5...Pixel reproduction unit, 106,806...Decoding row memory, 107,807-813...Interpolation row memory.

Claims (1)

[Claims] 1. When a signal obtained by scanning an original image is sequentially sampled and converted into a binary signal, the changed pixel position where the value changes from one value to the other value is encoded. Decoding is performed based on the created pass mode, horizontal mode, and vertical mode codes, and during decoding, the changed pixel position of the decoded row to be decoded is compared to the changed pixel position of the reference row immediately before the decoded row. Image processing that detects the mode-specific code indicating displacement, and inserts one or more interpolation lines between the reference line and the decoded line according to the magnification rate according to the detected mode-specific code. In the method, when the detected code has a displacement of 0 in the pass mode, the horizontal mode, or the vertical mode, it is inserted between the reference row and the decoded row according to an enlargement ratio. As the contents of one or more interpolation lines, the contents in the pixel range of the decoded line related to each mode are adopted as they are, and the detected code is in the vertical mode and the displacement amount is 0.
If the number is 3 or less, including 3, then the contents of one or more interpolation lines inserted between the reference line and the decoded line according to the expansion rate, depending on the displacement amount and expansion rate of each. An image processing method characterized by employing a predetermined pattern. 2 The amount of displacement is not 0 in the vertical mode,
As an enlarged pixel inserted into multiple interpolation rows when the number is 3 or less including 3, a pixel close to a changed pixel in the reference row or the same changed pixel is adopted in an interpolation row closer to the reference row. Featured image processing method.