JPH09200541A - Coder, decoder, coding method, decoding method and image processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the coding efficiency by dividing an image into blocks consisting of M-row×N-column (1<=M, 2<=N) pixels and applying pixel rearrangement to shift picture pixels with a prescribed value in each block toward one side so as to improve a prediction hit rate. SOLUTION: Multi-value image data 1 are converted into binary error spread data 3 by an error spread section 2. A pixel rearrangement section 4 applies block processing to the error spread data 3 in the unit of M-row×N-column (1<=M, 2<=N) matrices and 1, 0 are rearranged. In the rearrangement processing, 1s or 0s are packed to the left in the unit matrix in the unit of L(1<=L) rows. For example, in the case of L=1, 1, 0 are packed to the left alternately in the unit of one row. The rearranged data 5 subjected to rearrangement processing are prediction-coded by a prediction coder 6 and converted into coding data 7.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a coding device, a decoding device, a coding method, a decoding method, and an image processing device suitable for image data, particularly random dither image data.

[0002]

2. Description of the Related Art As one of image coding techniques, there is known a predictive coding technique for predicting a target pixel level from peripheral pixel levels and coding prediction error data. In this technique, it is necessary to improve the predictive predictive value in order to increase the coding efficiency.

JP-A-5-316370 and JP-B-6
In the encoding technique described in Japanese Patent Laid-Open No. 3-9426, a predictor having a high hit rate is adaptively switched from among a plurality of predictors, and finally one predictor is selected to determine the value of the value of the current image. Used for prediction to generate prediction error data.

Further, in the encoding method and the decoding method described in Japanese Patent Publication No. 6-95645, prediction is performed by using the predicted values output by the table lookup method with the values of a plurality of peripheral pixels as addresses. Although the error data is generated, the positions of the predicted pixels are not all fixed, but the positions of only one pixel can be made variable, and the predicted position is changed according to the predictive hit ratio.

As described above, in the conventional technique, the predictor is adaptively switched based on the prediction result to improve the prediction efficiency.

However, when the predictive coding technique is applied to the random dither image represented by the error diffusion image, the occurrence of "0" and "1" has no periodicity or regularity, and therefore the prediction efficiency is inevitable. In many cases, the prediction error data does not become data that can improve the compression efficiency.

In order to increase the predictive predictive value of such a random dither image, in the image data compression apparatus described in Japanese Patent Laid-Open No. 4-1160460, the density of the peripheral pixel of the pixel of interest is calculated and the peripheral pixel is calculated. The predicted value is determined so that the image densities are closer to each other compared with the densities in a larger range including.

However, the predictive value determination method disclosed in Japanese Patent Laid-Open No. 4-11460 cannot cope with the case where there is a change in density, etc., and the predictive value is determined by a quantitative judgment, so that the predictive value cannot be obtained in many cases.

[0009]

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned circumstances, and in an apparatus for encoding and compressing a random dither image represented by an error diffusion image,
An encoding device, a decoding device, an encoding method, a decoding method, and an image processing device, which can use a predictive encoding method and improve the predictive predictive value, are provided to improve encoding efficiency and provide halftone. The purpose is to perform decoding with good image reproducibility.

[0010]

According to a first aspect of the present invention, in an encoding apparatus, an image has M rows × N columns (1 ≦ M,
2 ≦ N) pixels are divided into blocks, and a pixel rearranging unit that moves pixels of a predetermined value in each block to one of the blocks, and a pixel signal of N pixels before at least the pixel of interest It is characterized by having a predictive coding means for predicting a value and coding based on the prediction result.

According to a second aspect of the present invention, in the encoding device, an image is divided into blocks of M rows × N columns (2 ≦ M, 2 ≦ N) pixels, and pixels having a predetermined value in each block are L pixels. (1 ≦
L) Pixel rearranging means for alternately arranging each of the blocks on one side and the other of the block, and a pixel signal of 2L rows before the target pixel and a pixel signal of N pixels before the target pixel It is characterized by having a predictive coding means for predicting the value of and predicting the value based on the prediction result.

According to a third aspect of the present invention, in a decoding device, a decoding means for decoding the code encoded by the encoding device according to the first or second aspect, and a prediction error signal decoded by the decoding means. It is characterized by having an inverse prediction means for performing an inverse prediction process based on the above, and a correction means for performing a correction process on an image output from the inverse prediction means.

According to a fourth aspect of the present invention, in the decoding device according to the third aspect, the correction means detects a change point of the pixel value and appropriately changes the pixel value of an area including the change point. It is characterized by.

According to a fifth aspect of the present invention, in the decoding apparatus according to the third aspect, the correction means detects a change point at which the pixel value changes from the predetermined value to another value, and the change point is detected. The feature is that the values of the sandwiched pixels are exchanged. Pixels that sandwich the change point are not limited to pixels that sandwich the change point adjacent to both sides of the change point, but also include pixels that sandwich the change point at a position distant from the change point.

According to a sixth aspect of the present invention, in the encoding method, the image is divided into blocks of M rows × N columns (1 ≦ M, 2 ≦ N) pixels, and pixels having a predetermined value in each block are divided into blocks. Pixels are rearranged so as to be closer to one of the blocks, the value of the pixel of interest is predicted using a pixel signal of N pixels before the pixel of interest, and encoding is performed based on the prediction result. Is.

According to a seventh aspect of the present invention, in the encoding method, the image is divided into blocks of M rows × N columns (2 ≦ M, 2 ≦ N) pixels, and pixels having a predetermined value in each block are L. (1 ≦
L) The pixels are rearranged so as to alternately approach one side and the other side of the block for each row unit, and the pixel signal of 2L rows before the target pixel and the pixel signal of N pixels before the target pixel are used for attention. It is characterized in that the value of a pixel is predicted and encoded based on the prediction result.

According to an eighth aspect of the invention, in the decoding method, the code encoded by the encoding method according to the sixth or seventh aspect is decoded, and inverse prediction processing is performed based on the decoded prediction error signal. The correction processing is performed on the image after the inverse prediction processing.

According to a ninth aspect of the present invention, in the decoding method according to the eighth aspect, the correction processing detects a pixel value change point and appropriately changes a pixel value of an area including the change point. It is characterized by.

According to a tenth aspect of the present invention, in the decoding method according to the eighth aspect, the correction processing detects a change point at which the pixel value changes from the predetermined value to another value, and detects this change point. The feature is that the values of the sandwiched pixels are exchanged. Pixels that sandwich the change point are not limited to pixels that sandwich the change point adjacent to both sides of the change point, but also include pixels that sandwich the change point at a position distant from the change point.

An eleventh aspect of the present invention is an image processing device, comprising the encoding device according to the first or second aspect and the decoding device according to any one of the third to fifth aspects. It is what

[0021]

FIG. 1 is a block diagram showing an embodiment of an encoding apparatus according to the present invention. In the figure, 1 is image data, 2 is error diffusion unit, 3 is error diffusion data, 4 is pixel rearrangement unit, 5 is rearrangement data, 6 is predictive coding unit,
Reference numeral 7 is code data.

The multivalued image data 1 is converted into binary error diffusion data 3 in the error diffusion unit 2. In the pixel rearrangement unit 4, after the error diffusion data 3 is divided into pixels by a matrix unit of M rows × N columns (2 ≦ M, 2 ≦ N), rearrangement processing of “1” and “0” is performed. Do.
In the rearrangement process, “1” or “0” is packed to the left within the matrix unit for each L (1 ≦ L) row unit.

FIG. 2 is an explanatory diagram of a specific example of the rearrangement process in the embodiment of the encoding apparatus of the present invention. In this specific example, M, N = 4, and L = 1. FIG. 2 (A)
Shows one matrix unit in the error diffusion data, and FIG. 2B shows the rearranged image data after the rearrangement processing.

In this embodiment of the rearrangement process, since L = 1, "1" and "0" are alternately packed to the left for each row unit. First, in the first line, "1" is packed to the left,
The data of “1101” is rearranged into “1110”. In the second line, 0s are left-aligned this time, and the input of "1010" is rearranged to "0011". Below, "1" is left-justified on the 3rd line, and "0" is left-justified on the 4th line.
The rearrangement process of "1100" and "0111" is performed.

Returning to FIG. 1, description will be made. The rearrangement data 5 rearranged by the pixel rearrangement unit 4 as described above.
Is predictively coded by the predictive coding unit 6 and converted into coded data 7.

FIG. 3 is an explanatory diagram showing an example of prediction processing in the embodiment of the encoding apparatus of the present invention. This embodiment is a configuration example of the predictor corresponding to the rearrangement data 5 rearranged by the matrix of M, N = 4, L = 1 described in FIG.

First, in the horizontal direction, the pixel data at the position X ₃ four pixels before the pixel of interest P becomes the first reference pixel data. The reason why it is set to 4 pixels before is that M = 4, so that the pixel data at the positions of the target pixel P and X ₃ are at the same position in different matrix units, and it is considered that the correlation is strong.

In the vertical direction, the pixel data at the position Y ₁ on the second row with respect to the target pixel P becomes the second reference pixel data. Since this is not L = 1, it is 2 × L 2
It is considered that the pixel data at the position of Y ₁ on the row and the pixel of interest P are filled with “1” or “0” in the same direction in different matrix units, respectively, so that the correlation is considered to be strong. .

For the final determination of the predicted value, the one having the higher hit ratio is selected from the first reference pixel and the second reference pixel.

FIG. 4 is a block diagram showing an example of prediction value selection processing in the embodiment of the encoding apparatus of the present invention.
In the figure, 5 is rearrangement data, 8 and 9 are coincidence number counting units, 10 and 11 are count values, 12 is a comparator, 13 is a comparison result signal, 14 is a selecting unit, and 15 is a predicted value. The rearrangement data 5 is input to the matching number counting units 8 and 9. The coincidence number counting unit 8 determines the target pixels P and X until now.
_The number of coincidences with the pixel data at the position ₃ is sequentially counted, and the coincidence count unit 9 similarly counts the number of coincidences between the pixel data of the target pixel P and the pixel data at the position Y ₁ up to the present. To do. The count values 10 and 11 of the coincidence number counting units 8 and 9 are sent to the comparator 12. By the comparison result signal 13 of the comparator 12, in the selection unit 14,
From the pixel data at the position of X ₃ and the pixel data at the position of Y ₁ that are currently input, the value of the position with the greater number of matches is output as the prediction value 15 and is used for prediction of the current pixel of interest. .

In addition to this configuration, as shown in FIG. 5, a prediction value is determined by a prediction processing method using a prediction memory centered on the reference pixel X ₃ in the row direction and the reference pixel Y _{1 in the} column direction. Good. In FIG. 5A, 5 is rearrangement data,
Reference numeral 15 is a prediction value, and 16 is a prediction memory.

In the prediction process using the prediction memory of FIG. 5, first, from the rearranged data 5, the pixel data at the positions of X ₃ , Y ₁ , X ₀ and Y ₀ in FIG. Input to memory. In the prediction memory 16, the threshold values shown in FIG. 5 corresponding to the values at the positions of X ₃ , Y ₁ , X _U , and Y ₀ are written in advance.

In this example, the prediction value written in the prediction memory is determined by the following rules. When both X ₃ and Y ₁ are “0” or “1”, the predicted value is X ₃ and Y regardless of X ₀ and Y _{0.
It has} a value of ₁ . When X ₃ and Y ₁ are different, X ₀ and Y ₀ Is referred to. In this case, if X ₃ = 0 and Y ₁ = 1 then X
_{If 0} U is “0”, according to the rule of the rearrangement method in the matrix described above, when X ₃ and X ₀ are the same, all the values of X _{3 to} X ₀ are “0”, and the values in the horizontal direction are Since there is a high possibility that “0” s will be arranged and P will be “0”, the predicted value is set to “0”. Similarly, X ₃ =
If 1, Y ₁ = 0 and X ₀ = 1 then the predicted value is 1.
Further, when X ₃ = 0 and Y ₁ = 1 and Y ₀ = 1 there is a high possibility that “1” s are arranged in the vertical direction according to the rule of the sorting method in the matrix described above. Therefore, the predicted value is set to "1". Similarly, X ₃ = 1 and Y ₁
= 0 and Y ₀ = 0, the predicted value is set to “0”. X ₃ , Y ₁ , X ₀ , and Y ₀ are “1001” or “0
In the case of “110”, the prediction value may be arbitrarily determined, or may be a value referring to another pixel, for example, the same as the previous prediction value.

FIG. 5 is a table showing the above-mentioned rules.
(B). In this way, X ₃ , Y ₁ , X ₀ , Y
_The predicted value written in advance in the prediction memory 16 is read with reference to the value of ₀ , and the predicted value 15 is determined.

Furthermore, the determination of the predicted value is performed by the method disclosed in Japanese Patent Laid-Open No. 31
The method described in Japanese Patent No. 6370 may be used to adaptively select the predictive hit count in one line, or the predictive memory 16 in FIG. May be.

Further, in the concrete example of FIGS. 2 to 5, M, N =
However, the present invention is not limited to this, and the matrix size may be changed according to the image quality and the compression rate. For example, if L = 2, the first two lines are left-aligned with "1" (or "0"), and the second two lines are left-aligned with "0" (or "1").
In this way, "1" or "0" is alternately packed to the left every two lines. In the case of this example, for the reference pixel in the vertical direction described with reference to FIG. 3, the pixel data at a position four rows above the pixel of interest P becomes the second reference pixel data. Since L = 2, the pixel data at the position on the 4th row which is 2 × L and the target pixel P are filled with “1” or “0” in the same direction in different matrix units. It is considered that the correlation is strong due to the existence of

Further, it is also possible to pack "1" or "0" to the left for all the rows without alternately packing "1" and "0" to the left.

Next, the decoding of encoded coded data according to the present invention will be described. FIG. 6 is a block diagram showing an embodiment of the decoding device of the present invention. In the figure, 3 is error diffusion data, 4 is a pixel rearrangement unit, 5 is rearrangement data, 7 is code data, 17 is a decoding unit, 18 is prediction error data, and 19 is an inverse prediction unit.

The coded data 7 is decoded into the prediction error data 18 by the decoding unit 17, and then the inverse prediction unit 19 performs the reverse processing of the prediction method used at the time of coding to be converted into the rearranged data 5. It At the time of decoding, necessary information about the prediction method is given and the process in the inverse prediction unit is performed. The obtained rearrangement data 5 may be output as the image data 3, but by the rearrangement processing at the time of encoding,
Since the arrangement of "1" and "0" is biased to some extent, it is preferable to perform correction processing.

FIG. 7 is an explanatory diagram of a specific example of the rearrangement process in the embodiment of the decoding apparatus of the present invention. If both "1" and "0" exist in the same row in the matrix, the process of exchanging "1" and "0" is performed. For example, since the first row in FIG. 7A is an array of “1110”, the “1” of the third pixel and the “0” of the fourth pixel are exchanged, and the first row of FIG. 7B is replaced. Of “1101”. Also for the second row, "0" of the second pixel in the row of "0011"
Then, "1" of the third pixel is replaced with "0101". In the correction processing by this rearrangement, the pixel value is "1".
From "0" to "0" or from "0" to "1" is detected, and pixels on both sides adjacent to this change point are replaced.

In this way, when the values of the pixels sandwiching the change point are exchanged, not only the method of exchanging the pixels sandwiching the change point adjacently but also the appropriate pixels apart from the change point on both sides of the change point are exchanged. You may do it. For example,
In the first line of FIG. 7A, “1” of the first pixel is replaced with “0” of the fourth pixel to correct “0110” or 2
In the row, “0” of the first pixel and “1” of the fourth pixel may be replaced with each other to perform a correction process such as “1010”.

Further, when the matrix size at the time of encoding is large, conversion may be performed by using a pattern conversion table or the like so that the arrangement of 1,0 on the same row is arranged more randomly.

As a result of this correction processing, the rearranged image is
Since 1 and 0 are randomly arranged as compared with the image before rearrangement, there is an advantage that the reproducibility of the halftone image is improved.

An image processing apparatus can be constructed using the encoding apparatus and the decoding apparatus according to the present invention. The encoding apparatus can increase the data compression rate, which is advantageous when storing or transferring image data. The data read from the memory or transferred is
The decoding device can decode the image data with good reproducibility.

[0045]

As is apparent from the above description, according to the present invention, deterioration of image quality is minimized with respect to a random dither image represented by an error diffusion image having no regularity or periodicity of pixels. Thus, by providing a certain degree of regularity, it is possible to significantly improve the predictive coding efficiency.

Further, in encoding, by performing a rearrangement process of bringing pixels in each block closer together, the predictive encoding efficiency can be remarkably improved by a simple method. On the other hand, in decoding, there is an effect that the reproducibility of a halftone image can be improved by a simple correction process of replacing the values of pixels sandwiching a change point.

[Brief description of drawings]

FIG. 1 is a block diagram illustrating an embodiment of an encoding device according to the present invention.

FIG. 2 is an explanatory diagram of a specific example of a rearrangement process in the embodiment of the encoding device of the present invention.

FIG. 3 is an explanatory diagram showing an example of a prediction process in the embodiment of the encoding device of the present invention.

FIG. 4 is a configuration diagram showing an example of a predicted value selection process in the embodiment of the encoding device of the present invention.

FIG. 5 is an explanatory diagram of a prediction process using a prediction memory in the embodiment of the encoding device of the present invention.

FIG. 6 is a block diagram showing an embodiment of a decoding device of the present invention.

FIG. 7 is an explanatory diagram of a specific example of rearrangement processing in the embodiment of the decoding apparatus of the present invention.

[Explanation of symbols]

1 ... Image data, 2 ... Error diffusion part, 3 ... Error diffusion data, 4 ... Pixel rearrangement part, 5 ... Rearrangement data, 6 ... Predictive coding part, 7 ... Code data, 8, 9 ... Matching count part 10, 11 ... Count value, 12 ... Comparator, 13 ... Comparison result signal, 14 ... Selection unit, 15 ... Prediction value, 16 ... Prediction memory, 17 ... Decoding unit, 18 ... Prediction error data, 19 ...
Inverse predictor.

Claims (11)

[Claims] 1. A pixel rearranging unit that divides an image into blocks each having M rows × N columns (1 ≦ M, 2 ≦ N) pixels and shifts pixels having a predetermined value in each block to one of the blocks. An encoding apparatus comprising: a predictive encoding unit that predicts a value of a pixel of interest using a pixel signal N pixels before the pixel of interest and encodes the value based on the prediction result. 2. The image is divided into blocks of M rows × N columns (2 ≦ M, 2 ≦ N) pixels, and pixels of a predetermined value in each block are divided into L (1 ≦ L) row units of the block. The value of the pixel of interest is predicted by using the pixel rearranging means that alternately shifts to one side and the other, and the pixel signal of 2L rows before the pixel of interest and the pixel signal N pixels before of the pixel of interest to predict the prediction result. An encoding apparatus having a predictive encoding means for encoding based on the above. 3. Decoding means for decoding the code coded by the coding device according to claim 1, and inverse prediction means for carrying out inverse prediction processing based on the prediction error signal decoded by the decoding means. A decoding device comprising: a correction unit that performs a correction process on an image output from the inverse prediction unit. 4. The decoding device according to claim 3, wherein the correction unit detects a change point of the pixel value and appropriately changes the pixel value of an area including the change point. 5. The correction unit according to claim 3, wherein the correction unit detects a change point at which the pixel value changes from the predetermined value to another value, and replaces the values of pixels sandwiching the change point. Decoding device. 6. The image is divided into blocks of M rows × N columns (1 ≦ M, 2 ≦ N) pixels, and pixels are rearranged so that pixels of a predetermined value in each block are closer to one of the blocks, An encoding method characterized by predicting a value of a pixel of interest at least using a pixel signal N pixels before the pixel of interest and encoding based on the prediction result. 7. The image is divided into blocks of M rows × N columns (2 ≦ M, 2 ≦ N) pixels, and pixels of a predetermined value in each block are divided into L (1 ≦ L) row units of the block. Pixels are rearranged so as to alternately approach one side and the other side, and the value of the target pixel is predicted using the pixel signal of 2L rows before the target pixel and the pixel signal of N pixels before the target pixel, and the prediction is performed. An encoding method characterized by encoding based on a result. 8. A code encoded by the encoding method according to claim 6 or 7 is decoded, inverse prediction processing is performed based on the decoded prediction error signal, and an image after the inverse prediction processing is performed. A decoding method characterized by performing a correction process. 9. The decoding method according to claim 8, wherein the correction processing detects a change point of a pixel value and appropriately changes a pixel value of an area including the change point. 10. The correction process according to claim 8, wherein a change point at which the pixel value changes from the predetermined value to another value is detected, and the values of pixels sandwiching the change point are replaced. Decryption method. 11. An image processing apparatus comprising the encoding device according to claim 1 or 2, and the decoding device according to any one of claims 3 to 5.