JPH07212598A - Device for encoding picture compression - Google Patents

Abstract

PURPOSE:To enable controlling code quantity in accordance with the difference of quality in respective signal components constituting a picture signal by respectively generating scale factors which are prepared at every signal component based on calculated distribution code quantity so as to be multiplied by a quantization table. CONSTITUTION:A DCT coefficient quantized by a quantization circuit 34 is supplied to a Huffman encoding circuit 92, executed a variable encoding processing at a block unit and, then, supplied to Y code quantity calculating circuits 43-46 which are set in accordance with respective kinds of group picture data G1-G4 at the time of processing a luminance signal so as to be integrated at every group. They are supplied to C code quantity calculating circuits 47-50 which are arranged in accordance with the respective kinds of group picture data G1-G4 at the time of processing color difference signal so as to be integrated at every group. Integrated values respectively outputted from the circuits 43-46 and 47-50 are supplied to a Y/C distribution code quantity caluculating circuit 51. The circuit 51 calculates Y distribution code quantity NYt and C distribution code quantity NCt based on code quantity set in a code quantity setting circuit 52.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an improvement of an image compression coding apparatus for performing compression coding processing on digitized image data.

[0002]

2. Description of the Related Art As is well known, an image compression coding apparatus such as the one described above is conventionally constructed as shown in FIG. First, CCD (charge coupled device)
The luminance signal and the color difference signal obtained by photoelectric conversion by 11 are supplied to the image pickup processing circuit 12 and converted into digital data, respectively, and then recorded in the image memory 13. Each digital data of the luminance and color difference signal components recorded in the image memory 13 is scanned twice. In the first scan, the code amount and scale factor αt are calculated, and in the second scan,
Actual encoding is performed using the calculated scale factor αt.

That is, in the first scan, the digital data read from the image memory 13 is DCT.
(Discrete cosine transform) The DCT coefficient is generated by being supplied to the circuit 14. In this case, from the image memory 13, FIG.
As shown in, the horizontal direction 8 pixels × the vertical direction 8 pixels 64
Block reading is performed for each block of pixels, in which digital data is read in the order shown by the numerical values in FIG. Then, in the DCT circuit 14, the generated DCT coefficients are grouped into four groups G as shown in FIG.
1, G2, G3, G4, and each group G1, G
The image data for one group is sequentially output to the quantization circuit 15 for each of G2, G3, and G4.

The quantizing circuit 15 converts the input group image data from the Y quantization table 17 and the C quantization table 18 selected by the selector 16 and the scale factor output from the MPU (microprocessor unit) 19. And are divided by the value multiplied by the multiplication circuit 20. In this case, the scale factor is the group image data G1,
Values α1, α2, α3, α4 corresponding to G2, G3, G4
Are controlled by the MPU 19, and are generated. That is, the group image data Gi (i = 1, 2, 3, 4)
Are scale factors αi (i =
1, 2, 3, 4) will be quantized. Further, the selector 16 is switching-controlled to use the Y quantization table 17 when handling the group image data of the luminance signal and to use the C quantization table 18 when handling the group image data of the color difference signal.

Then, D quantized by the quantization circuit 15
The CT coefficient is supplied to the Huffman coding circuit 21 and subjected to variable length coding processing in block units, and then the code amount calculation circuit 22 installed corresponding to each group image data G1, G2, G3, G4. , 23, 24, 25, and are added up for each group. Here, since one image data is divided into four groups G1, G2, G3, and G4, each code amount calculation circuit 22, 23, 2
The integrated value output from 4 and 25 is 1/4 of the code integrated amount of one image data. Therefore, in each code amount calculation circuit 22, 23, 24, 25, the integrated value is multiplied by 4 to obtain the code amount N1, N2, N3, N4.

FIG. 7 shows the scale factors α1 and α.
2 shows the relationship between α2, α3 and α4 and the code amounts N1, N2, N3 and N4. Now, assuming that the code amount Nt set in the code amount setting circuit 26 is at the position shown in FIG. 7, A (α
The scale factor αt corresponding to the code amount Nt is calculated from the linear equation linearly approximated by the two points of the (2, N2) point and the B (α3, N3) point. Therefore, each code amount calculation circuit 22,
Code amount N1, N2, N output from 23, 24, 25
3 and N4 are supplied to the code amount comparison circuit 27, and the set code amount N
The value closest to t and the value closest to t are selected. Then, the selected two code amounts are supplied to the αt calculation circuit 28, and the set code amount Nt is calculated from the linear equation obtained by the two-point linear approximation.
The scale factor αt corresponding to is calculated.

Next, in the second scan, the digital data read from the image memory 13 is transferred to the DCT circuit 1.
4 to produce DCT coefficients. Also in this case,
Reading of digital data from the image memory 13 is performed by block reading. Then, the DCT circuit 14
The DCT coefficient generated in step 1 is quantized by the quantization circuit 1 in block units.
5 is output. This quantization circuit 15 receives the input D
The CT coefficient is divided by a value obtained by multiplying the Y quantization table 17 and the C quantization table 18 selected by the selector 16 and the scale factor αt output from the MPU 19 by the multiplication circuit 20. In this case, the scale factor αt is generated by controlling the value obtained from the αt calculation circuit 28 in the first scan with the MPU 19. And
The DCT coefficient quantized by the quantization circuit 15 is supplied to the Huffman coding circuit 21, subjected to variable length coding processing in block units, and taken out from the output terminal 29.

By the way, in the above-mentioned conventional image compression encoding apparatus, both the luminance signal component and the color difference signal component are processed using the same scale factor αi (i = 1, 2, 3, 4). However, there is a large difference in image resolution and the like between the luminance signal and the color difference signal, and in general, the color difference signal has a lower resolution than the luminance signal. Nevertheless, performing the same quantization processing on both the luminance signal and the color difference signal results in redundant processing on the color difference signal. In other words, the code amount to be distributed to the high-resolution luminance signal is distributed to the low-resolution chrominance signal, which causes a problem of low processing efficiency from the viewpoint of image quality or resolution. ing.

[0009]

As described above, in the conventional image compression coding apparatus, the luminance signal and the color difference signal are subjected to the same quantization processing regardless of the difference in resolution. There is a problem that the efficiency of signal processing is poor because the amount of codes corresponding to the property is not distributed.

Therefore, the present invention has been made in consideration of the above circumstances, and provides an extremely good image compression coding apparatus capable of controlling the code amount corresponding to the difference in the property of each signal component forming the image signal. The purpose is to do.

[0011]

An image compression coding apparatus according to the present invention has a plurality of signal components constituting an image signal,
An orthogonal transform unit that performs orthogonal transform in block units each including a plurality of pixels, and transform coefficients for each of a plurality of signal components output from the orthogonal transform unit, respectively. A quantization unit that performs a quantization process based on a value obtained by multiplying a quantization table by a scale factor, and a plurality of conversion coefficients for each signal component output from the quantization unit are encoded for each signal component. Coding means for performing processing, and computing means for calculating a distributed code amount for each signal component based on the code amount for each of a plurality of signal components respectively output from the coding means,
Based on the distribution code amount output from this arithmetic means, there is provided a generating means for respectively generating a scale factor for multiplying the quantization table prepared for each signal component.

[0012]

According to the above configuration, the distribution code amount for each signal component is calculated based on the code amount for each of the plurality of signal components that have been subjected to the encoding process, and the calculated distribution code amount is set to the calculated distribution code amount. Since the scale factor for multiplying the quantization table prepared for each signal component is generated based on each, the code amount control corresponding to the difference in the property of each signal component forming the image signal can be performed. Like

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described in detail below with reference to the drawings. In FIG. 1, the luminance signal and the color difference signal obtained by photoelectric conversion by the CCD 30 are supplied to the image pickup processing circuit 31 and converted into digital data, respectively, and then recorded in the image memory 32. The digital data of the luminance and color difference signal components recorded in the image memory 32 are each scanned twice. In the first scan, the code amount is calculated and the scale factor αt is calculated, and in the second scan, the actual encoding is performed using the calculated scale factor αt.

That is, in the first scan, the digital data read from the image memory 32 is supplied to the DCT circuit 33 to generate DCT coefficients. In this case, as shown in FIG. 5, digital data is read from the image memory 32 for each block consisting of 64 pixels of 8 pixels in the horizontal direction × 8 pixels in the vertical direction in the order shown by the numerical values in FIG. Reading is performed. And DCT
In the circuit 33, the generated DCT coefficient is divided into four groups G1, G2, G3 in block units as shown in FIG.
The images are classified into G4, and one group of images for each group G1, G2, G3, G4 are sequentially output to the quantization circuit 34.

The quantizing circuit 34 converts the input group image data into the Y quantizing table 36 and the C quantizing table 37 selected by the selector 35, and the selector 38.
The scale factor output from the MPU 39 or 40 selected in step 1 is divided by the value multiplied by the multiplication circuit 41. In this case, the scale factor is the group image data G
It corresponds to 1, G2, G3, G4, and at the time of luminance signal processing, the MPU 39 scale factors αY1, αY2, αY3,
The value that controls αY4 is selected by the selector 38, and the scale factors αC1, αC2,
A value that controls αC3 and αC4 is selected by the selector 38.

That is, the group image data Gi (i =
1, 2, 3, 4) correspond to the scale factors αYi (i = 1, 2, 3, 4) and αCi (i = 1, 1, respectively).
2, 3, 4) will be used for quantization. Also,
The selector 35 is switched and controlled to use the Y quantization table 36 when handling the group image data of the luminance signal and to use the C quantization table 37 when handling the group image data of the color difference signal.

Then, D quantized by the quantization circuit 34
The CT coefficient is supplied to the Huffman coding circuit 42 and subjected to variable length coding processing in block units, and then each group image data G1, G2, G3, G4 at the time of luminance signal processing.
Corresponding to the Y code amount calculation circuits 43, 44, 4
5 and 46 and are integrated for each group, and when the color difference signal processing is performed, each group image data G1, G
2, G3 and G4 are respectively supplied to C code amount calculation circuits 47, 48, 49 and 50, which are integrated for each group. Here, one image data is G1, G
Since it is divided into four groups of G2, G3, and G4, Y and C code amount calculation circuits 43 to 46 and 47 to 50
The integrated value output from is 1/4 of the code integrated amount of one image data. Therefore, in the Y code amount calculation circuits 43 to 46, the integrated values are multiplied by 4 to obtain the code amounts NY1 and NY.
2, NY3 and NY4 are obtained, and the C code amount calculation circuit 47 is obtained.
˜50, the integrated value is multiplied by 4 and the code amount NC1, NC2,
We have obtained NC3 and NC4.

Y and C code amount calculation circuits 43 to 46 and 4
The integrated value output from each of 7 to 50 is supplied to the Y / C distribution code amount calculation circuit 51. The Y / C distribution code amount calculation circuit 51 calculates the Y distribution code amount NYt and the C distribution code amount N based on the code amount set in the code amount setting circuit 52.
Calculate Ct. This calculation formula is expressed as NYt = NY / C × Nt / (NY / C + 1) NCt = Nt / (NY / C + 1). However, in the above equation, NY / C = [(NY1 / NC1) + (NY2 / NC2) +
(NY3 / NC3) + (NY4 / NC4)] / 4.

FIG. 2 shows the scale factors αY1 and αY.
2, αY3, αY4 and code amount NY1, NY2, NY3, NY
4 shows the relationship with 4. Now, assuming that the Y distribution code amount NYt calculated by the Y / C distribution code amount calculation circuit 51 is in the position shown in FIG. 2, the AY (αY2, NY2) point and the BY
A scale factor αYt corresponding to the code amount NYt is calculated from a linear equation linearly approximated by two points (αY3, NY3). Therefore, each Y code amount calculation circuit 43, 44, 4
5, 46, output code amounts NY1, NY2, NY
3 and NY4 are supplied to the Y code amount comparison circuit 53, and the value closest to the Y distribution code amount NYt and the value closest to the second are selected. Then, the selected two code amounts are supplied to the αYt calculation circuit 54, and the scale factor αYt corresponding to the Y distribution code amount NYt is calculated from the linear equation obtained by the two-point linear approximation.

Further, FIG. 3 shows the above scale factor αC.
1, αC2, αC3, αC4 and code amount NC1, NC2, NC
3, the relationship with NC4 is shown. Now, assuming that the C distribution code amount NCt calculated by the Y / C distribution code amount calculation circuit 51 is at the position shown in FIG. 3, two points of AC (αC2, NC2) and BC (αC3, NC3) The scale factor αCt corresponding to the code amount NCt is calculated from the linear equation linearly approximated by the points. Therefore, each C code amount calculation circuit 47, 4
Code amounts NC1, NC2 output from 8, 49, 50
NC3 and NC4 are supplied to the C code amount comparison circuit 55, and the value closest to the C allocation code amount NCt and the value closest to the second are selected. Then, the selected two code amounts are supplied to the αCt calculation circuit 56, and the scale factor αCt corresponding to the C distribution code amount NCt is calculated from the linear equation obtained by the two-point linear approximation.

Next, in the second scan, the digital data read from the image memory 32 is transferred to the DCT circuit 3.
3 to generate DCT coefficients. Also in this case,
Reading of digital data from the image memory 32 is performed by block reading. Then, the DCT circuit 33
The DCT coefficient generated in step 3 is quantized by the quantization circuit 3 in block units.
4 is output. This quantization circuit 34 receives the input D
The CT coefficient is divided by a value obtained by multiplying the Y quantization table 36 and the C quantization table 37 selected by the selector 35, and the scale factors output from the MPUs 39 and 40 selected by the selector 38 by the multiplication circuit 41. .

In this case, the scale factor is generated by controlling the values obtained from the αYt and αCt calculation circuits 54 and 56 in the first scan by the MPUs 39 and 40, respectively. Then, when processing the luminance signal, the MPU 39
The value that controls the scale factors αY1, αY2, αY3, and αY4 is selected by the selector 38, and MP is selected at the time of color difference signal processing.
A value which controls the scale factors αC1, αC2, αC3, and αC4 in U40 is selected by the selector 38. Further, the selector 35 is set to Y when handling the group image data of the luminance signal.
The quantization table 36 is used, and when the group image data of the color difference signals is handled, the switching is controlled to use the C quantization table 37. Then, the DCT coefficient quantized by the quantization circuit 34 is supplied to the Huffman coding circuit 42, subjected to variable length coding processing in block units, and taken out from the output terminal 57.

Therefore, according to the configuration of the above embodiment, the distribution code amount of each luminance and color difference signal component is calculated based on the code amount of the luminance and color difference signal components after Huffman coding, and this is calculated. Since the scale factors αYt and αCt for multiplying the Y and C quantization tables 36 and 37 prepared for the respective luminance and color difference signal components are generated based on the distributed code amount, It becomes possible to control the code amount corresponding to the difference in the resolution of the luminance and color difference signal components. The present invention is not limited to the above-described embodiment, and can be variously modified and implemented without departing from the scope of the invention.

[0024]

As described above in detail, according to the present invention,
It is possible to provide an extremely good image compression encoding device capable of controlling the code amount corresponding to the difference in the property of each signal component forming the image signal.

[Brief description of drawings]

FIG. 1 is a block configuration diagram showing an embodiment of an image compression encoding device according to the present invention.

FIG. 2 is a diagram for explaining a method for obtaining a scale factor for a luminance signal in the same embodiment.

FIG. 3 is a diagram for explaining a method of obtaining a scale factor for a color difference signal in the embodiment.

FIG. 4 is a block configuration diagram showing a conventional image compression encoding device.

FIG. 5 is a diagram shown for explaining the order of block reading.

FIG. 6 is a diagram for explaining group classification of blocks of one image data.

FIG. 7 is a diagram shown for explaining a method of obtaining a scale factor from a set code amount.

[Explanation of symbols]

11 ... CCD, 12 ... Imaging processing circuit, 13 ... Image memory, 14 ... DCT circuit, 15 ... Quantization circuit, 16 ... Selector, 17 ... Y quantization table, 18 ... C quantization table, 19 ... MPU, 20 ... Multiplier circuit, 21 ... Huffman coding circuit, 22-25 ... Code amount calculation circuit, 26 ... Code amount setting circuit, 27 ... Code amount comparison circuit, 28 ... .alpha.t calculation circuit, 29 ... Output terminal, 30 ... CCD, 31 ... Image pickup processing circuit, 32 ... Image memory, 33 ... DCT circuit, 34 ... Quantization circuit, 35 ... Selector, 36 ... Y quantization table, 3
7 ... C quantization table, 38 ... selector, 39, 40 ...
MPU, 41 ... Multiplication circuit, 42 ... Huffman coding circuit,
43 to 46 ... Y code amount calculation circuit, 47 to 50 ... C code amount calculation circuit, 51 ... Y / C distribution code amount calculation circuit, 52 ... Code amount setting circuit, 53 ... Y code amount comparison circuit, 54 ... αYt calculation Circuit, 55 ... C code amount comparison circuit, 56 ... αCt calculation circuit, 57 ... output terminal.

Claims (1)

[Claims] 1. A plurality of signal components constituting an image signal,
An orthogonal transformation unit that performs orthogonal transformation in block units each including a plurality of pixels, and transformation coefficients for each of the plurality of signal components output from the orthogonal transformation unit are prepared corresponding to each signal component. Based on a value obtained by multiplying a quantization table by a scale factor, a quantizing means for performing a quantizing process, and a transform coefficient for each of the plurality of signal components output from the quantizing means, for each signal component, respectively. Coding means for performing coding processing, computing means for computing the distributed code quantity for each signal component based on the code quantity for each of the plurality of signal components respectively output from the coding means, and the computing means Generating means for respectively generating the scale factor for multiplying the quantization table prepared for each signal component based on the distribution code amount output from Image compression encoding apparatus characterized by comprising Te.