JPH08263645A - Digital copying machine - Google Patents

Abstract

PURPOSE: To compress full color and multi-level digital picture data in accordance with the property of a picture such as a character and a photograph. CONSTITUTION: A quantization coefficient matrix for photograph mode or a quantization coefficient matrix for character mode are previously set in a code quantity control part 7. The function key of an operation part 8 is used as the mode designation key of a 'photograph mode' and a 'character mode'. When copying is started, picture data which is read out by a scanner 1 is transmitted to an encoding part 3 through a picture processing part 2. When the 'photograph mode' is set, the quantization coefficient matrix for photograph mode is set in a quantization table 41 from a code quantity control part 7 and it is compressed by using the quantization coefficient matrix for photograph mode. When the 'character mode' is set, data is compressed by using the quantization coefficient matrix for character mode.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital copying apparatus for compressing full-color or multi-tone digital image data.

[0002]

2. Description of the Related Art Generally, image memory capacity and image transmission cost can be reduced by compressing digital image data. Conventionally, as a digital copying apparatus for compressing digital image data, for example, an actual Kaihei 3-123 is used.
It is disclosed in Japanese Patent Laid-Open No. 355,355 and Japanese Patent Laid-Open No. 3-97372.

[0003]

By the way, in the copying machine, the image processing method can be changed based on the characteristics of the image such as "photo mode" and "character mode". The user can specify the mode accordingly or automatically select the mode. In this case, the smoothing filter is used for the photographic image and the emphasis filter is used for the character image. Further, the copying machine is not provided with means for designating the image quality based on the image data amount (communication data amount) such as "normal" or "fine character" which is provided in a facsimile or the like.

Here, for example, the JPEG method is known as a standard method for compressing color / multi-tone digital image data. Among the JPEG methods, the DCT (Discrete Cosine Transform) method is a so-called lossy compression method in which the original image is not completely restored even if the digital image data is compressed and then restored.
The compression ratio can be easily changed by controlling the compression parameter. That is, the quality of the restored image can be controlled by the compression rate from the level where the compression rate is small and the image deterioration is small and cannot be visually recognized to the level where the compression rate is increased and the image deterioration is visually recognized.

In general, as a method of quantifying the deterioration of a compressed / decompressed image in lossy compression, S / S, which is a statistic of the least square error between the original image data and the image data after the compression / decompression. N ratio is used. It is said that the S / N ratio is not necessarily in a fixed proportional relationship with the degree of deterioration of an image by subjective evaluation, but as a general tendency, the degree of deterioration is well represented.

Therefore, if the original image is compressed, the capacity of the image memory can be reduced when the image data is stored in the memory, and the image data can be efficiently transferred when it is transferred via the transmission path. be able to.

Also, the Institute of Image Electronics Engineers, Proceedings 92-03-0
3, P13-P20, D in the JPEG system
It has been reported that the degree of deterioration of a compressed / decompressed image in a CT-based baseline system according to a subjective evaluation substantially follows the “color difference” between the original image and the decompressed image. "Color difference" is also a statistical amount of error due to compression / decompression,
It can be seen that it can be used as a means for quantifying the degree of image deterioration.

Therefore, it is an object of the present invention to provide a novel digital copying apparatus capable of compressing full-color or multi-tone digital image data according to the nature of an image such as a character or a photograph.

Another object of the present invention is to provide a novel digital copying apparatus capable of efficiently compressing full-color or multi-tone digital image data. .

[0010]

In order to achieve the above object, a first means responds to an image compression means for compressing image data by a lossy method and a property of image data compressed by the image compression means. And a control means for controlling the degree of deterioration of the image when the image data compressed by the image compression means is restored.

A second means is an orthogonal transform means for orthogonally transforming image data into orthogonal transform coefficient data by the image compressing means in the first means, and an orthogonal transform coefficient transformed by the orthogonal transform means. It is characterized in that it comprises a quantizing means for quantizing the data based on the quantization table, and the control means changes the contents of the quantization table according to the property of the image data.

A third means is a value for dividing the orthogonal transform coefficient data obtained by transforming the contents of the quantization table by the orthogonal transform means in the first or second means,
The control means changes the value of the quantization table according to the property of the image data.

A fourth means is an image compression means for compressing image data in an irreversible system, and a deterioration degree of an image when the image data compressed by the image compression means is restored to a degree that cannot be visually recognized. And a control unit for controlling the image compression unit to compress the image.

A fifth means is that, in the fourth means, the control means calculates a statistical amount of a difference between the image data amount before being compressed by the image compressing device and the image data amount after compression and restoration, It is characterized in that the statistic amount is controlled to be a predetermined amount.

A sixth means is characterized in that in the fifth means, the statistic is a color difference.

[0016]

According to the first means, when the image data is compressed by the lossy method, the degree of deterioration of the image when the compressed image data is restored is controlled according to the property of the image data. It is compressed depending on whether it is a character or a photograph, and therefore full-color or multi-tone digital image data can be compressed according to the nature of an image such as a character or a photograph.

In the second means, when the orthogonal transform coefficient data is quantized based on the quantization table, the content of the quantization table is changed according to the property of the image data, so that the original is a character or a photograph. The contents of the quantization table are changed depending on whether or not it is possible to compress full-color or multi-tone digital image data according to the nature of the image such as characters or photographs.

In the third means, when the content of the quantization table is a value that divides the orthogonal transform coefficient data, the value of the quantization table is changed according to the property of the image data. , The contents of the quantization table are changed depending on whether it is a photograph or the like, so that full-color or multi-tone digital image data can be compressed according to the nature of the image such as characters or photographs.

According to the fourth means, when the image data is compressed by the irreversible method, since the deterioration degree of the image is compressed to such a degree that it cannot be visually recognized, the full-color or multi-tone digital image data is efficiently processed. Can be compressed into.

In the fifth means, the statistic of the difference between the image data amount before compression and the image data amount after compression / restoration is controlled to be a predetermined amount. Image data can be efficiently compressed.

In the sixth means, since the statistic is the color difference, full-color digital image data can be efficiently compressed.

[0022]

Embodiments of the present invention will be described below with reference to the drawings. 1 is a block diagram showing an embodiment of a digital copying apparatus according to the present invention, FIG. 2 is a block diagram showing in detail the encoding unit and the decoding unit of FIG. 1, and FIG. 3 is an explanation showing the operation unit of FIG. 4 and 5 are flow charts for explaining the compression process of the digital copying apparatus of FIG. 1, and FIG. 5 is a flow chart for explaining the compression process of the modified example of FIG.

FIG. 1A shows a configuration having an image memory 4, and FIG. 1B shows a transmission line 10 instead of the image memory 4.
Shows the configuration provided with. The original is read by the scanner 1 and output as an analog image signal, and the analog image signal is corrected by the image processing unit 2 for shading correction
Processing such as A / D conversion is performed.

This digital image data is compressed by the encoding unit 3 shown in detail in FIG. 2, stored in the image memory 4 in the example shown in FIG. 1 (a), and in the transmission line in the example shown in FIG. 1 (b). Sent to 10. The compressed data read from the image memory 4 or the compressed data received via the transmission path 10 is decoded by the decoding unit 5 shown in detail in FIG. 2 and recorded on the recording paper by the printing unit 6.

FIG. 2 shows an example of D in the JPEG system.
3 shows an encoding unit 3 and a decoding unit 5 according to a CT system baseline system. Input image data in the encoding unit 3 is divided into blocks of 8 × 8 pixels, and this data is subjected to 8 × 8 orthogonal transform (DC) by the DCT unit 31.
T) Converted into matrix data of coefficients. Next, the quantizing unit 32 quantizes the 8 × 8 orthogonal transform coefficient by dividing it by the quantized coefficient of the quantization table 41, and each element of the quantized orthogonal transform coefficient is a one-dimensional coefficient string. Is converted to. Then, the entropy coder (code assigning unit) 33 converts the one-dimensional coefficient sequence into an entropy code (Huffman code or arithmetic code).

When the compressed data is restored in the decoding unit 5, the entropy decoder (code decoding unit) 5 is used.
The data matrix of the quantized orthogonal transformation coefficient is obtained by regenerating the one-dimensional coefficient sequence by 3 and reconstructing it into the matrix of the block size. Next, the inverse quantization unit 52 multiplies this by the quantization coefficient of the quantization table 41 equal to that of the encoding unit 3 to obtain the orthogonal transform coefficient before quantization, and then the inverse DCT.
The unit 51 restores the pixel data to 8 × 8 pixels.

Here, in the quantizer 32, the DCT
The coefficient orthogonally transformed by the unit 31 is reduced to a small value such as “0” or “1”, and as a result, when converted into a one-dimensional coefficient sequence, the coefficient “0” is continuous or “0”,
The appearance probability of the specific pattern including “1” increases, and the bias of the appearance probability can be used to perform compression by the entropy code. Therefore, the compression rate can be increased by causing many small values such as “0” and “1” to occur in the coefficient sequence.

However, the coefficient reduced to "0" remains "0" even if it is inversely quantized by the inverse quantization unit 52, except when the coefficient value before quantization is "0". The original coefficient value is not restored, and the coefficient reduced to a small value such as “1” has a large error from the original coefficient value even when inversely quantized. That is, the factor of this coefficient becomes the main factor of image deterioration in the JPEG system.

Therefore, by setting each element of the quantization table 41 in advance so as to quantize the high-frequency component coefficient element in the image data, which is less likely to be deteriorated due to visual characteristics, into a smaller coefficient, the number of elements is reduced. The compression rate can be earned depending on the degree of image deterioration. As a result, the compression rate can be controlled by changing the quantization table 41, and the degree of image deterioration can be controlled. In this case, it is possible to store a plurality of quantization coefficient matrices (having 8 × 8 elements) as the quantization table 41 and use one of them. Then, whichever quantized coefficient matrix is used for the quantized table 41, one scale factor f _SN is multiplied by 64 elements of the quantized coefficient matrix (that is, 64 quantized table 41 The compression ratio can be made variable (by multiplying by the element).

According to the first embodiment, the compression table and the degree of deterioration of the image can be controlled by making the quantization table 41 variable, but the image deterioration depends on the nature of the original image. The reproducibility is different. That is, since an image that emphasizes gradation expression has relatively few high-frequency components,
The compression rate of the low frequency component can be set small and the compression rate of the high frequency component can be set large. However, in an image that emphasizes edges such as a character image, on the contrary, a relatively large amount of the high frequency component is included. By increasing the compression rate and decreasing the compression rate of the high frequency component, efficient compression can be achieved regardless of the nature of the original image.

In FIG. 4, the quantized coefficient matrix for the photo mode and the quantized coefficient matrix for the character mode are preset in the code amount control section 7 shown in FIG. 1, and the function key 8a of the operating section 8 shown in FIG. It shows a process of setting the quantization table 41 from the code amount control unit 7 according to the designated mode by using it as a mode designation key such as "photograph mode" and "character mode".

That is, in FIG. 1, when copying is started, the image data read by the scanner 1 is sent to the encoding unit 3 via the image processing unit 2. When the "photo mode" is set, the photo mode quantization coefficient matrix is set in the quantization table 41 from the code amount control unit 7, and the photo mode quantization coefficient matrix is compressed using the photo mode quantization coefficient matrix. (Steps S11, S1
2). When the "character mode" is set, the character mode quantization coefficient matrix is used for compression (steps S13 and S14).

FIG. 5 shows a modification of the first embodiment shown in FIG. 4, in which an "automatic mode" is added to the first embodiment, and the image processing unit 2 is either "photograph" or "photograph" in the automatic mode. It is configured to automatically determine "character". In this case, when the "automatic mode" is set, the quantization coefficient matrix for the photo mode or the character mode is used according to the determination result of the image processing unit 2 (step S2).
1 to S26). When the "auto mode" is not set, when the "photo mode" is set, compression is performed using the quantized coefficient matrix for the photo mode (steps S23 and S24), and the "character mode" is set. If is set, compression is performed using the character mode quantized coefficient matrix (steps S25 and S26).

Further, the compression rate can be changed by multiplying all 64 elements of the quantization table 41 by the scale factor f _SN from the code amount control unit 7 shown in FIG. That is, if the scale factor f _SN is increased, the compression rate increases, the compression code amount decreases, and the image quality deteriorates more. On the contrary, if the scale factor f _SN is reduced, the compression rate is reduced and the compression code amount is increased.
Image quality deterioration is reduced. In the JPEG method, the image quality varies depending on the nature of the original image. However, in the case of a full-color image such as a photograph or a multi-gradation image, if the compression rate is made smaller than a certain level (for example, 1/5 or less), the image quality It is said that the deterioration of is almost invisible to human eyes.

When controlling the scale factor f _SN , as shown in FIG.
"Photo mode" and "Character mode" set via a
Alternatively, as shown in FIG. 5, the scale factor f _SN is set to the control unit 9 with respect to the quantization table 41 set in the quantization coefficient matrix determined using the property of the image automatically determined in the “auto mode”. Is set in the code amount control unit 7.

When the start key 8b of the operation unit 8 is pressed in the state where the image characteristics are thus determined, the original image read by the scanner 1 is sent to the encoding unit 3 via the image processing unit 2. Then, the encoding unit 3 quantizes based on the scale factor f _SN and the quantization table 41 set by the code amount control unit 7. Further, at the time of decoding, the compressed data is the scale factor f _SN and the quantization table 41.
Is inversely quantized based on.

Next, a second embodiment will be described with reference to FIG. In this embodiment, the ΔE calculation unit 11 is added, and ΔE
The calculation unit 11 calculates the difference ΔE in image quality between the original image data input to the encoding unit 3 and the image data decoded by the decoding unit 5. When calculating this difference ΔE,
A predetermined data amount unit (for example, a unit of 8 lines) of the image data read by the scanner 1 is sent to the encoding unit 3 via the image processing unit 2 and at the same time to the buffer memory in the ΔE calculation unit 11.

In the encoding unit 3, quantization is performed based on the initial value f _SN1 of the scale factor f _SN , and at the time of decoding, compressed data is dequantized based on this initial value f _SN1 and decompressed to restore ΔE calculation. It is also sent to the section 11 to calculate the difference ΔE. When the “color difference” is used as the difference ΔE, it can be calculated based on the description in P16 of the IEICE Technical Paper 92-03-03.

Here, in an irreversible image compression method such as the JPEG method, the deterioration of the image can hardly be visually recognized below a certain compression rate. Since the degree of visual deterioration can be quantitatively approximated by the difference ΔE, the difference ΔE ′ when the degree of deterioration cannot be visually recognized is set in the code amount control unit 7. In the second embodiment,
Difference ΔE 'depending on the nature of the image such as "text" or "photograph"
May be set in the code amount control unit 7.

Then, in the code amount control unit 7, the ΔE calculation unit 1
The scale factor f _SN2 is calculated so that the difference ΔE calculated by 1 becomes equal to the designated difference ΔE ′, and this scale factor f _SN2 becomes the scale factor f _SN that determines the compression rate during actual copying. The image to be calculated by the ΔE calculation unit 11 may be the whole or a part of the document to be read. In the case of a monochrome multi-tone image, the S / N ratio can be used instead of the "color difference".

[0041]

As described above, according to the first aspect of the invention, when the image data is compressed by the lossy method, the deterioration degree of the image when the compressed image data is restored depends on the property of the image data. It is controlled according to the
It is compressed according to whether it is a photograph or the like. Therefore, full-color or multi-tone digital image data can be compressed according to the nature of an image such as a character or photograph.

According to the second aspect of the present invention, when the orthogonal transform coefficient data is quantized based on the quantization table, the contents of the quantization table are changed according to the property of the image data. , The contents of the quantization table are changed depending on whether it is a photograph or the like, so that full-color or multi-tone digital image data can be compressed according to the nature of the image such as characters or photographs.

According to a third aspect of the invention, when the content of the quantization table is a value by which the orthogonal transform coefficient data is multiplied,
Since the value of the quantization table is changed according to the property of the image data, the content of the quantization table is changed according to whether the original is a character or a photo. It can be compressed depending on the nature of the image, such as text or photos.

According to the fourth aspect of the invention, when the image data is compressed by the lossy method, the image deterioration is compressed to such a degree that it cannot be visually recognized. When compressing, the capacity of the image memory can be reduced, and the data can be efficiently transferred via the transmission path.

According to the fifth aspect of the present invention, since the statistical amount of difference between the image data amount before compression and the image data amount after compression and restoration is controlled to be a predetermined amount, full color or multi-gradation is performed. The digital image data can be efficiently compressed.

According to the sixth aspect of the invention, since the statistic is color difference, full-color digital image data can be efficiently compressed.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a digital copying apparatus according to the present invention.

2 is a block diagram showing details of an encoding unit and a decoding unit in FIG. 1. FIG.

FIG. 3 is an explanatory diagram showing an operation unit of FIG. 1.

FIG. 4 is a flowchart for explaining a compression process of the digital copying apparatus of FIG.

FIG. 5 is a flowchart for explaining a compression process of the modified example of FIG.

FIG. 6 is a block diagram showing a digital copying machine according to a second embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Scanner 2 Image processing unit 3 Encoding unit 4 Image memory 5 Decoding unit 7 Code amount control unit 8 Operation unit 8a Image property designation key 8b Start key 9 Control unit 10 Transmission path 11 ΔE calculation unit

Claims (6)

[Claims] 1. An image compression means for compressing image data by a lossy method, and a method for restoring the image data compressed by the image compression means according to the property of the image data compressed by the image compression means. A digital copying apparatus comprising: a control unit that controls the degree of image deterioration. 2. The image compression means orthogonally transforms image data into orthogonal transform coefficient data, and the orthogonal transform coefficient data transformed by the orthogonal transform means is quantized based on a quantization table. 2. The digital copying apparatus according to claim 1, further comprising: a quantizing unit for converting the contents of the quantization table according to a property of image data. 3. The content of the quantization table is a value for dividing the orthogonal transform coefficient data transformed by the orthogonal transform means, and the control means is a value of the quantization table according to a property of image data. The digital copying apparatus according to claim 1 or 2, characterized in that: 4. An image compression means for compressing image data in an irreversible manner, and the image compression means to such an extent that the degree of deterioration of the image when the image data compressed by the image compression means is restored cannot be visually recognized. And a control means for controlling the compression so that the digital copying apparatus. 5. The control means calculates a statistic of a difference between the amount of image data before being compressed by the image compressing means and the amount of image data after being compressed and restored so that the statistic becomes a predetermined amount. 5. The digital copying machine according to claim 4, wherein 6. The digital copying apparatus according to claim 5, wherein the statistic is color difference.