JP3200189B2 - Image data encoding / restoring method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for encoding and restoring image data, and more particularly to a method and apparatus for encoding and restoring a moving image. It stores image data whose information amount is orders of magnitude larger than numerical data, especially data of halftone images and color images.
For high-quality transmission, it is necessary to efficiently encode grayscale values for each pixel. As an efficient compression method for image data, for example, an adaptive discrete cosine transform coding method (Ad
aptive Discrete CosineTransform
CT ”). The ADCT method uses 8 ×
The image signal of each block is divided into blocks of 8 pixels, and the image signal of each block is converted into a DCT coefficient of a spatial frequency distribution by a two-dimensional discrete cosine transform (hereinafter, referred to as “DCT transform”), quantized with a threshold suitable for vision, The obtained quantized DCT coefficients are encoded by a Huffman table statistically obtained.

[0002]

2. Description of the Related Art A first conventional method for encoding and restoring a moving image is to treat a moving image as a set of still images and independently encode each image by a still image encoding device.
FIG. 8 is a configuration diagram of a conventional still image encoding device using ADCT. The image is divided into blocks of 8 × 8 pixels shown in FIG. 9 and input from a terminal 31 to a two-dimensional DCT transform unit 32. In the two-dimensional DCT transform unit 32, the input image signal is orthogonally transformed by DCT, converted into DCT coefficients of a spatial frequency distribution shown in FIG. 10 and output to the linear quantization unit 33. Specifically, the image signal input from the terminal 31 is one-dimensionally DCT-transformed by the one-dimensional DCT conversion part 32a, and is output to the transposition part 32b. The transposition unit 32b replaces the rows and columns of the coefficients in one block and converts the one-dimensional DCT
output to c. The one-dimensional DCT transform unit 32c performs one-dimensional DCT transform on the input transposed coefficients again, and
Output to d. The transposition unit 32d outputs a signal obtained by replacing the rows and columns of the coefficients in one block again to the linear quantization unit 33 in the same manner as the transposition unit 32b. The linear quantization unit 33 multiplies each of the input DCT coefficients by a quantization control parameter with each of the quantization matrices 34 composed of the threshold values shown in FIG. Then, the DCT coefficient is divided by the quantization threshold to perform a linear quantization to obtain a quantized DCT coefficient. As a result of this quantization, as shown in FIG.
The T coefficient becomes zero, and a quantized DCT coefficient having only the DC component and a small AC component has a value.

[0003] The quantized DCT coefficients arranged two-dimensionally are converted into one dimension by zigzag scanning shown in FIG. The variable length coding unit 35 calculates the DC component at the head of each block and the D component of the previous block.
The difference from the C component is variable-length coded. For the AC component, a variable-length coding is performed on the value of a non-zero effective coefficient (hereinafter, referred to as an “index”) and the length of a run of an invalid coefficient that becomes zero up to the index (hereinafter, referred to as a “run”). Further, each of the DC and AC components is a Huffman code table 36 composed of a Huffman table created based on the statistics of each image.
And the obtained coded data is sequentially output from the terminal 37.

On the other hand, code data is restored to an image by the following method. FIG. 14 is a configuration diagram of a conventional still image restoring device using ADCT. In FIG. 14, the terminal 41
Are input to the variable length decoding unit 42. In the variable length decoding unit 42, the Huffman code table 3
6 (FIG. 8), the input code data is decoded into fixed-length data of an index and a run by a Huffman decoding table 43 composed of a table reverse to the Huffman table, and output to an inverse quantization unit 44. I do. The inverse quantization unit 44 inversely quantizes the input quantized DCT coefficient by multiplying each of the quantization matrices 45 by a value (quantization threshold value) multiplied by the quantization control parameter, thereby obtaining a DCT coefficient. The data is restored and output to the two-dimensional inverse DCT transform unit 46. The two-dimensional inverse DCT transform unit 46 orthogonally transforms the input DCT coefficients by inverse DCT, and converts the DCT coefficients of the spatial frequency distribution into image signals. Specifically, the input DCT coefficients are subjected to one-dimensional inverse DCT transform by a one-dimensional inverse DCT transform unit 46a and output to a transpose unit 46b. The transposition unit 46b replaces the rows and columns of the coefficients in one block and performs a one-dimensional inverse DCT transformation unit 46c.
Output to The one-dimensional inverse DCT transform unit 46c performs one-dimensional inverse DCT transform on the input transposed coefficients again, and
Output to 6d. The transposed unit 46d restores the image by replacing the rows and columns of the coefficients in one block again and outputting the same as in the transposed unit 46b. The still image encoding / restoring device is capable of high-speed encoding and restoring by LSI, and is capable of encoding and restoring at 30 frames / sec.
It is possible to encode and restore a moving image with a simple configuration.

A second conventional method for encoding and restoring a moving image employs a moving image encoding technique for a continuous scene such as a video conference. FIG. 15 is a configuration diagram of a conventional moving picture encoding apparatus. In the moving image encoding technique, the orthogonal transform, quantization, and variable length encoding are performed on the difference between the restored data up to the previous image obtained by restoring the encoded data of the previous image and the original image. The compression ratio is increased by assistive technology such as motion compensation (not shown). And, for example, the code amount per frame (code rate)
Is controlled to change the quantization threshold value in the quantization unit so that becomes a predetermined optimum code rate. That is, the current image is divided into blocks each including 8 × 8 pixels, input to the difference image generation unit 52 from the terminal 51, calculates the difference from the previous frame image, and stores the difference image in the block buffer 53. Thereafter, the difference image is stored in the block buffer 53 as 2
It is input to the dimensional DCT transform unit 54. The two-dimensional DCT transformer 54 orthogonally transforms the input difference image signal by DCT, converts the DCT coefficient into a spatial frequency distribution DCT coefficient, and outputs the DCT coefficient to the quantizer 55. The quantization unit 55 obtains a quantized DCT coefficient by dividing the input DCT coefficient by a quantization threshold. The two-dimensionally arranged quantized DCT coefficients are converted into one dimension by zigzag scanning, and
Is input to the inverse quantization unit 57.

The inverse quantization unit 57 inversely quantizes the input quantized DCT coefficient by multiplying by the quantization threshold to restore the DCT coefficient, and outputs the DCT coefficient to the two-dimensional inverse DCT transform unit 58. The two-dimensional inverse DCT transformer 58 converts the input DC
The T coefficient is orthogonally transformed by the inverse DCT transform, and the DCT coefficient of the spatial frequency distribution is transformed into a difference image signal. Adder 59
Generates a current image signal by adding the previous image signal and the difference image signal stored in the image memory 60 to generate the current image signal.
To be stored. The image signal newly stored in the image memory 60 is used when the next image signal is input from the terminal 51.
It is input to the difference image generator 52 as a previous image signal.

In parallel with the above, the variable-length coding unit 56 performs variable-length coding on the differential image signal, codes using the Huffman code table 61, and stores the obtained code data in the code buffer 62. Thereafter, by performing the above processing for each block, the code data of the current image for one frame is stored in the code buffer 62, and is sequentially output from the terminal 63. The record control unit 64 obtains the code amount (code rate) of the code data of the current image stored in the code buffer 62, compares the code rate with a predetermined optimum code rate, and calculates a code having the optimum code rate. The quantization threshold in the quantization unit 55 is changed and controlled so as to achieve the rate. That is, when the code rate of the current image is high, the quantization threshold is increased to reduce the code rate of the next image, and when the code rate of the current image is low, the quantization threshold is reduced. , The code rate of the next image is increased, and control is performed so that the code rate becomes the optimum code rate on average.

[0008] The optimum code rate is determined by setting the code data to I
When transmission is performed via a communication network such as SDN, it is predetermined in consideration of the maximum transmission speed of the communication network and the like. When code data is stored in a storage medium (for example, a floppy disk, a hard disk, or the like), Is determined in consideration of the storage capacity of the storage medium. For example, in order to transmit a moving image displaying 30 frames per second, both the encoding side and the restoration side need to capture and display an image every 1/30 second. Therefore, when this image is transmitted by ISDN (64 Kbps), the code amount of one image is 267 Byte (64000/30 = 2133bit → 267)
If it is less than (byte), transmission of 30 moving images per second is possible. Therefore, the optimal code rate (code amount of one frame)
Is set to a value of 267 bytes or less, for example, 260 bytes.

On the other hand, the restoration of the code data is performed as follows. FIG. 16 is a block diagram of a conventional moving image restoration apparatus. The difference image restored in the reverse order of the encoding is sequentially added to the contents of the image memory up to the previous screen, and written again in the image memory. Restore. That is, the code data input from the terminal 71 is input to the variable length decoding unit 72. In the variable length decoding unit 72, the Huffman code table 61
The input code data is decoded into fixed-length data of an index and a run by a Huffman decoding table 73 composed of a table opposite to the Huffman table in FIG. . The inverse quantization unit 74 performs inverse quantization by multiplying the input quantized DCT coefficient by a quantization threshold to restore the DCT coefficient, and performs a two-dimensional inverse DCT transformation unit 75.
Output to The two-dimensional inverse DCT transform unit 75 orthogonally transforms the input DCT coefficients by inverse DCT transform, and transforms the DCT coefficients of the spatial frequency distribution into an image signal (difference image signal). The adder 76 adds the difference image signal and the previous image signal stored in the image memory 77 to generate a current image signal,
The data is rewritten into the image memory 77. As described above, the image of one block is restored, and if the same processing is performed for all the blocks of one frame image thereafter, the current image is stored in the image memory 7.
7 is stored. In FIG. 16, reference numeral 78 denotes a pixel write control unit, and 79 denotes a block address generation unit. The pixel write control section 78 generates a pixel write address in each block, and the block address generation section 79 generates an address indicating a block position composed of 8 × 8 pixels, and stores the address in the image memory 77 designated by both addresses. Write the image signal. In the moving image encoder described above, the code rate can be controlled to be equal to the optimum code rate defined by the transmission speed, the capacity of the storage medium, and the like.

[0010]

According to the first method in which a moving image is regarded as a set of still images and each image is independently encoded by a still image encoding device, a simple method and a small circuit are used. There is an advantage that moving images can be encoded and restored on a large scale. However, the first method does not control the code rate for each frame image. For this reason, in the first method, the worst (hard to compress) image data is assumed, and the quantization threshold is determined in advance so that the code rate of the image data does not exceed the maximum code rate defined from the transmission speed and the like. It is fixed. Therefore, there is a problem in that even image data that is easily compressed is coarsely encoded, and the image quality obtained by restoration is poor. On the other hand, according to the second method that employs a moving image encoding technique for a continuous scene, the code rate of each frame image is set equal to the optimal code rate defined by the transmission speed, the capacity of the storage medium, and the like. It has the advantage that it can be controlled to be. However, the moving picture encoder has a problem that the encoding method is complicated and the circuit scale is large. As described above, an object of the present invention is to provide a moving image encoding / restoring method and apparatus capable of maintaining a constant code rate by a simple method and achieving high image quality. Another object of the present invention is to provide a moving image encoding / decompression method and apparatus capable of maintaining a constant code rate with a small circuit scale.

[0011]

FIG. 1 is a diagram illustrating the principle of the present invention. Reference numeral 10 denotes an image data encoding device, 12 denotes an encoding unit that encodes one frame of image data, 15 denotes a reference code amount holding unit that holds a reference code amount (code rate) F per frame, and 16 denotes a reference code amount holding unit. A code amount determination unit 17 that determines the magnitude of the code amount f when the one-frame image is encoded and the reference code amount F. When the code amount f is smaller than the reference code amount F, the code amount determination unit 17 determines the code data of the one-frame image. A code data output unit for outputting a copy code when the code amount f is larger than the reference code amount F, and updating the reference code amount F based on the result of the magnitude discrimination between the code amount f and the reference code amount F A reference code amount updating unit. Reference numeral 20 denotes an image data decoding device, 21 denotes a code analysis unit that analyzes whether the code data for each frame is data obtained by coding an image or data (copy code) instructing duplication, 22 Represents a restoration unit for restoring the original image data from the code data;
Is an image memory that holds the restored image data, and 24 is a frame display control unit that holds the restored image of the previous frame in the image memory as a restored image of the current frame and displays it if the code data is data instructing duplication. is there.

[0012]

The optimum code rate (reference value) fr as the reference code amount F for one frame image is previously stored in the reference code amount holding unit 15.
Set to. In such a state, the encoding unit 12 in the encoding device 10 encodes the input image data.
The code amount determination unit 16 determines the magnitude of the code amount f of one frame image and the reference code amount F, and inputs the determination result to the code data output unit 17. The code data output unit 17 outputs the code data of one frame image input from the encoding unit 12 when the code amount f is smaller than the reference code amount F, and outputs the code data when the code amount f is larger than the reference code amount F. Outputs coded data (copy code) instructing the duplication of the previous frame image. The reference code amount updating unit 18 updates the reference code amount F based on the result of the magnitude discrimination between the code amount f and the reference code amount F. On the other hand, the code analysis unit 21 of the decompression device 20 analyzes the code data of each frame, and if the code data is data obtained by coding an image, the decompression unit 22 restores the code data to the original image. And stores it in the image memory 23. However, if the code data is data instructing the duplication, the frame display control unit 24 causes the restored image of the previous frame to be held in the image memory 23 and displays the restored image of the current frame.

That is, an optimum code amount (reference value fr) desired to be obtained in advance is set as a reference code amount F, and the code data exceeding this is stopped and the transfer of only the copy code is stopped. Transfer data. Then, the same transfer control is performed by updating the reference code amount F.
In the restoration, only valid code data (code data of an image) is restored, and in the case of a copy code, the previous frame image data is used as it is. As described above, the code amount is compared with the reference code amount, and if the code amount is larger than the reference code amount, the code data of the image is not transmitted, and only the copy code instructing the copy of the previous image is transmitted instead. Therefore, the code rate can be controlled to be constant with a simple and small-scale configuration. Also, unlike the first method of the related art, it is not necessary to determine the quantization threshold value in consideration of the worst image, and high image quality can be obtained.

[0014] Also, not to update the reference code amount F based on the level decision result of the code amount f and the reference code amount F
You. For example, when the code amount f is larger than the reference code amount F, the reference code amount F is updated so as to be larger by the following equation: F + fr → F (1), and when the code amount f is smaller than the reference code amount F, , The reference code amount F is updated by the following equation: F−f + fr → F (2) That is, when f> F, equation (1) is used.
Therefore, copy code is frequently used because F is increased.
Can be prevented from being sent,
Slow motion can be prevented. Also, f <F
At this time, F is updated according to the equation (2).
If the signal amount f is smaller than the reference code amount fr (f <fr),
F becomes large and image code data is sent as it is
The movement of the moving image becomes slow motion
(B) When the code amount f is equal to the reference code amount fr
If larger (f> fr), F becomes smaller and copies
The code is sent and the average transmission rate is
The transmission rate according to the signal volume fr can be prevented from exceeding.

[0015]

EXAMPLES encoder Figure 2 is an example block diagram of a coding apparatus of a moving image according to the present invention. Reference numeral 11 denotes a frame image holding unit that holds one frame of image data to be encoded, 12 denotes an encoding unit that encodes image data, and encodes image data by, for example, the ADCT method. A code buffer 14 holds code data obtained by encoding the frame image. A code buffer 14 stores a code (copy code) for instructing the duplication of the image data of the previous frame, code data indicating the start and end of the moving image, and the start and end of the frame. A fixed data holding unit for holding and appropriately generating these code data, a reference code amount holding unit 15 for holding a reference code amount F per frame, and an optimum code amount (code rate) to be obtained in advance is a reference value. The code amount determination unit 16 determines the magnitude of the code amount f when the image of one frame is encoded and the reference code amount F. Reference numeral 17 denotes a code amount. If the code amount f is smaller than the reference code amount F, code data of one frame image is output. If the code amount f is larger than the reference code amount F, a code data output unit that outputs a copy code. f and a reference code amount F based on the magnitude discrimination result of the reference code amount F.
Is a reference code amount updating unit for updating the reference code amount.

The coding portion encoding unit 12 is intended for coding image data by ADCT method has a coding device having the same configuration of a conventional static image as shown in FIG. That is, the image data divided into blocks of 8 × 8 pixels is converted into a two-dimensional DCT.
Input to the converter 12a. Two-dimensional DCT converter 12a
Converts the input image signal into an orthogonal transform by DCT to convert to a DCT coefficient of a spatial frequency distribution,
Output to The linear quantization unit 12b receives the input DCT
A quantization threshold is obtained by multiplying each of the quantization matrices 12c having coefficients by a threshold determined by a visual experiment by a quantization control parameter, and a DC is determined by the quantization threshold.
Linear quantization for obtaining a quantized DCT coefficient by dividing the T coefficient is performed. The two-dimensionally arranged quantized DCT coefficients are converted into one dimension by zigzag scan and input to the variable length coding unit 12d. The variable-length encoding unit 12d uses DC
As for the components, the difference between the DC component at the head of each block and the DC component of the previous block is subjected to variable length coding. For the AC component, the value (index) of the effective coefficient which is not zero,
Variable length coding is performed on the length of the run of the invalid coefficient which is zero up to the index. Further, the DC and AC components are encoded based on the Huffman code table 12e, and the obtained code data is sequentially output.

Code Data Output Unit The code data output unit 17 includes a code data control unit 17a and a selection switch 17b. Code data control unit 17a
Controls the selection switch 17b based on the magnitude discrimination result between the code amount f and the reference code amount F, and instructs the reference code amount update unit 18 to update the reference code amount. That is, the code data control unit 17a sets the selection switch 17b to select and output the code data of the image when f ≦ F, and selects the copy code instructing the duplication of the previous image when f> F. Output. The code data string output from the selection switch 17b has a structure shown in FIG. 4, and a SOM (Start of Motion)
A code is placed, and an EOM (End of Motion) code is placed at the end, during which code data of each frame is continuously arranged. If the code data of the frame is an image, an SOI (Start of Image) code is inserted at the beginning, and an EOI (End of Image) code is inserted at the end, and one frame of code data is arranged therebetween. You.
When a copy is instructed, the code data of the frame is composed of only a COI code (Copy of Image). Therefore, the code data control unit 17a operates the selection switch 17b to select and output the SOM code at the start of the moving image. Thereafter, when f ≦ F for each frame, “SOI code + frame code data + EOI code” Is output, and when f> F, a “COI code” is output, and at the end of the moving image, an EOM code is output.

Reference code amount update unit The reference code amount update unit 18 calculates the reference code amount F according to the following equation F + fr when the code amount f is larger than the reference code amount F in accordance with the code amount update instruction from the code data control unit 17a. If the code amount f is smaller than the reference code amount F, the reference code amount F is updated by the following formula: F−f + fr → F (2).

[0019] Operation Figure 5 the whole is a flowchart of the encoding process of the present invention. The code amount (code rate) of one frame image to be obtained is determined in advance, and the code amount fr is set as the reference code amount F, and the reference code amount holding unit 15 is used.
(Step 101). In this state, when one frame of image data is stored in the image buffer 11, the encoding unit 12 encodes the input image data by the ADCT method, and stores the encoded data in the code buffer 13 (step 103). . After encoding one frame,
The code buffer 13 inputs the code amount f for one frame to the code amount determination unit 16. The code amount determination unit 16 determines the magnitude of the code amount f of one frame image and the reference code amount F, and inputs the determination result to the code data control unit 17a of the code data output unit 17 (step 105). Code data control unit 17a
If the code amount f is smaller than the reference code amount F, the selection switch 17b is controlled to first output the SOI code, and then sequentially output the code data for one frame stored in the code buffer 13. And finally EOI
Output the code. In the case of the first frame, SO
Before the I code, the SOM code is output, and in the case of the last frame, the EOM code is output (step 10).
7).

Next, the code data control unit 17a instructs the reference code amount update unit 18 to update the reference code amount where f ≦ F. Thereby, the reference code amount updating unit 18 calculates the reference code amount F according to the equation (2), and
Is updated (step 10).
9). On the other hand, if the code amount f is larger than the reference code amount F, the result is “NO” in step 105. Therefore, the code data control unit 17a sends the reference code amount update Instruct. As a result, the reference code amount update unit 18 calculates the reference code amount F according to the equation (1), and updates the reference code amount held in the reference code amount holding unit 15 (step 111). Thereafter, the selection switch 17b is controlled to output the code data (copy code) COI for instructing the duplication of the previous frame image (step 113). Thereafter, the above process is repeated until the last frame of the moving image (step 115). If the last frame, the EOM code is output and the encoding process ends.

The recovery apparatus 6 is an embodiment diagram of a recovery apparatus of a moving image according to the present invention. Reference numeral 21 denotes a code analysis unit that analyzes whether the code data of each frame is data obtained by coding an image or data instructing duplication (copy code), and 22 denotes original image data from the code data of the image. Restoration unit to restore,
Reference numeral 23 denotes an image memory for holding the restored image data for one frame, and 24, if the code data is data for instructing duplication, holds the restored image of the previous frame in the image memory as the restored image of the current frame and a display device. Is a frame display control unit to be displayed. Restoring unit 22
Has the same configuration as the conventional still image coding apparatus. The variable-length decoding unit 22a decodes the code data input from the code analysis unit 21 into fixed-length data of an index and a run using the Huffman decoding table 22b,
2c. The linear inverse quantization unit 22c inversely quantizes the input quantized DCT coefficient by multiplying each of the quantization matrices 22d by a value (quantization threshold value) multiplied by the quantization control parameter, thereby obtaining a DCT coefficient. And restore 2
Output to the dimensional inverse DCT transform unit 22e. Two-dimensional inverse DCT
The transform unit 22e orthogonally transforms the input DCT coefficients by inverse DCT transform, and transforms the DCT coefficients of the spatial frequency distribution into image signals.

The overall operation Figure 7 is a flow diagram of restoration process of the present invention. When the code data is input, the code analysis unit 21 checks whether the code data is an EOI code (step 201). If the code data is an EOI code, the restoration processing ends. On the other hand, if it is not an EOI code, it is checked whether or not it is a copy code COI for each frame (step 203).
Code data sandwiched between the I code and the EOI code is input to the restoration unit 22. The restoration unit 22 restores the code data to the original image and stores it in the image memory 23 (step 20).
4) Thereafter, display is made on a display device (not shown) (step 205). However, if it is a copy code in step 203, the code analysis unit 21 notifies the frame display control unit 24 of that fact. As a result, the frame display control unit 24 stores the restored image of the previous frame in the image memory 23 and displays it as the restored image of the current frame. As described above, the present invention has been described with reference to the embodiments. However, the present invention can be variously modified in accordance with the gist of the present invention described in the claims, and the present invention does not exclude these.

[0023]

According to the present invention, when encoding,
The code amount (reference value fr) to be obtained in advance is set as the reference code amount F. If the code data exceeds this value, the transfer is stopped and only the copy code is transmitted. If not, the code data is transferred. , The same code control is performed by updating the reference code amount F, and only the valid code data (code data of the image) is restored at the time of restoration. Since it is used as it is, the code rate can be controlled with a simple and small-scale configuration, and the moving image can be restored at a constant rate even when restoring as a storage medium. Further, unlike the first method of the related art, it is not necessary to determine the quantization threshold in consideration of the worst image, and high image quality can be obtained.

[0024] Further, according to the present invention, to update the reference code amount F based on the level decision result of the code amount f and the reference code amount F
You. For example, when the code amount f is larger than the reference code amount F, the reference code amount F is updated so as to be larger by the following equation: F + fr → F. When the code amount f is smaller than the reference code amount F, the reference code amount F is updated. The quantity F is updated by the following equation: F−f + fr → F. That is, when f> F, equation (1) is used.
Therefore, copy code is frequently used because F is increased.
Can be prevented from being sent,
Slow motion can be prevented. Also, f <F
At this time, F is updated according to the equation (2).
If the signal amount f is smaller than the reference code amount fr (f <fr),
F becomes large and image code data is sent as it is
The movement of the moving image becomes slow motion
(B) When the code amount f is equal to the reference code amount fr
If larger (f> fr), F becomes smaller and copies
The code is sent and the average transmission rate is
The transmission rate according to the signal volume fr can be prevented from exceeding.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

FIG. 2 is a block diagram of an embodiment of a moving picture coding apparatus according to the present invention.

FIG. 3 is a configuration diagram of an encoding unit.

FIG. 4 is an explanatory diagram of a code code of a moving image according to the present invention.

FIG. 5 is a flowchart of an encoding process according to the present invention.

FIG. 6 is a block diagram of an embodiment of a moving image restoration apparatus according to the present invention.

FIG. 7 is a flowchart of a restoration process according to the present invention.

FIG. 8 is a configuration diagram of a conventional still image encoding device.

FIG. 9 is an explanatory diagram of an original image signal.

FIG. 10 is an explanatory diagram of DCT coefficients.

FIG. 11 is an explanatory diagram of a threshold value for a DCT coefficient.

FIG. 12 is an explanatory diagram of DCT coefficients after quantization.

FIG. 13 is an explanatory diagram of a scanning order of quantization coefficients.

FIG. 14 is a configuration diagram of a conventional still image restoration device.

FIG. 15 is a configuration diagram of a conventional moving image encoding device.

FIG. 16 is a configuration diagram of a conventional moving image restoration device.

[Explanation of symbols]

 10. Image data encoding device 12 Encoding unit 15 Reference code amount holding unit 16 Code amount determination unit 17 Code data output unit 18 Reference code amount update unit 20 Image data composite Coder 21 Code analysis unit 22 Restoration unit 23 Image memory 24 Frame display control unit

────────────────────────────────────────────────── ─── Continuation of front page (56) References JP-A-62-272790 (JP, A) JP-A-3-6184 (JP, A) JP-A-4-222184 (JP, A) JP-A-6-272184 30396 (JP, A) (58) Field surveyed (Int. Cl. ⁷ , DB name) H04N ^7/ 24-7/68 G06T 9/00

Claims (4)

(57) [Claims] 1. An image data code for encoding image data.
In the encoding method , a reference code amount F of one frame image is set to a reference value fr, and a code amount f at the time of encoding one frame image is calculated, and the magnitude of the code amount f and the reference code amount F is calculated. To determine the code amount f
There is the case where the reference code amount F smaller and outputs the code data of one frame image, when the code amount f is greater than the reference code amount F outputs a code data indicating the replication of the previous frame image, the code amount If f is larger than the reference code amount F, the reference code
The amount F is updated by adding the reference value fr.
If the code amount F is smaller than the reference code amount F,
updated by subtraction of r additions and code amount f, the image data encoding method, characterized in that. 2. An image for encoding and restoring image data.
In the data encoding / restoring method, the reference code amount F of one frame image is set to a reference value fr
Every one frame image by calculating the amount of codes f when the encoding, to determine the magnitude of the said code amount f and the reference code amount F, the code amount f
Is smaller than the reference code amount F, the one-frame image
And the code amount f is larger than the reference code amount F.
Code data to instruct the previous frame image to be copied.
If the code amount f is larger than the reference code amount F, the reference code
The amount F is updated by adding the reference value fr.
If the code amount F is smaller than the reference code amount F,
The image data is encoded by updating by adding r and subtracting the code amount f , the code data is received, and the code data for each frame is converted.
Analyze and if the code data is data obtained by encoding an image,
To restore the original image, and if the code data is data that
Image data encoding / restoring method, wherein a restored image of a system is output as a restored image of a current frame . 3. An image data code for coding image data.
In the encoding device, a frame holding frame image data to be encoded is stored.
A frame image holding unit (11), and a frame image coding unit (1
2) and hold the code data obtained by encoding the frame image.
Code data holding unit (13), and a copy code for instructing the duplication of the image data of the previous frame.
And a reference code amount holding a reference code amount F per frame.
Holding unit (15), code amount f when one frame image is encoded, and reference code amount
The code amount f is smaller than the reference code amount F.
Output the code data of one frame image
When the amount f is larger than the reference code amount F, the copy code is
A code data output section (17) to be output, and a reference code when the code amount f is larger than the reference code amount F.
The amount F is updated by adding the reference value fr.
If the code amount F is smaller than the reference code amount F,
The reference code amount is updated by adding r and subtracting the code amount f.
An image data encoding device comprising a new unit (18) . 4. An image for encoding and restoring image data.
In a data encoding / decoding device, an image data encoding device is a frame encoding device that holds frame image data to be encoded.
Encoding the frame image data with the frame image holding unit (11)
A frame image encoding unit (12), which holds encoded data obtained by encoding the frame image.
Code data holding unit (13), and a copy code for instructing the duplication of the image data of the previous frame.
And a reference code amount holding a reference code amount F per frame.
Holding unit (15), code amount f when one frame image is encoded, and reference code amount
The code amount f is smaller than the reference code amount F.
Output the code data of one frame image
When the amount f is larger than the reference code amount F, the copy code is
A code data output section (17) to be output, and a reference code when the code amount f is larger than the reference code amount F.
The amount F is updated by the summing of the reference value fr, the code amount f the reference
If the code amount F is smaller than the reference code amount F,
The reference code amount is updated by adding r and subtracting the code amount f.
A new section (18), wherein the image data restoring apparatus is configured such that the reception code data for each frame is a data obtained by coding an image.
Solution that analyzes whether the data is
Analysis unit (21), if the code data is data obtained by coding an image,
Image data restoration unit that restores original image data from data
(22) and a frame image holding unit (23) for holding the restored image data
If the code data is data that instructs duplication,
Frame image as the restored image of the current frame.
Means (24) for holding the image data in the image holding unit.
Place.