JP2003125407A - Image decoding device, image decoding method, and image decoding program - Google Patents

Abstract

(57) [Problem] To provide an image decoding device capable of decoding code data having an image size larger than a specific image size without changing an image memory or a decoded image display unit. In an image decoding device, a motion vector is calculated by a motion vector calculation unit (S14) if the macroblock is an inter-frame encoded macroblock (S14).
The four-conversion unit 17 converts the data into one-fourth (S15). Quarter converted by reference image 1/4 capturing section 16
Based on the motion vector, a quarter of the predicted reference image is extracted from the reconstructed image (S16), and is magnified four times by the QCIF → CIF expansion unit 14 (S17). A reconstructed image is generated by adding the prediction reference image enlarged four times to the decoded image after the end of the inverse DCT transform (S18). The image data is again reduced from the CIF size to the QCIF size and stored in the image memory 11 (S10), and the reduced image is displayed on the decoded image display unit 13 (S11, S12).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image decoding device, an image decoding method and an image decoding program for decoding coded data in which a moving image is compressed.

[0002]

2. Description of the Related Art In mobile handheld terminals and videophones,
The transmission speed of the communication line is slow, and the encoding process is indispensable for the transmission of huge amount of image data. The encoding method that can be adopted is MPEG, which is an international standard for media-integrated video compression.
(Moving Picture Experts Group), which is an IS
O (International Organization for Standardizatio
n: International Organization for Standardization) is known as a coding system with an expert group name. In addition, ITU-T (In
ternational Telecommunication Union Telecommunicat
Ion Standardization Sector: International Telecommunication Union / Telecommunication Standardization Sector). 263 is also widely known. These recommendations are based on CIF (Common Intermediate
e Format: Common intermediate format) and QCIF (Qu
arter CIF).

A CIF image is composed of 352 × 288 pixels, and a QCIF image is composed of 176 × 144 pixels. In addition, as coding modes, there are an intra coding (INTRA: intra-frame coding) mode in which inter-picture prediction is not applied and an inter-coding (INTER: inter-frame coding) mode in which inter-picture prediction is applied. is there.

When decoding coded data, processing is performed for each macroblock (8 × 8 pixels) from the dequantization processing after the VLD processing. Then, the locally decoded image is temporarily stored in the image memory, and at the time of decoding the INTER block, the reference image is taken out from the image memory by motion compensation prediction to generate a reconstructed image.

A conventional image decoding method and apparatus will be specifically described with reference to the drawings. FIG. 6 is a block diagram showing the configuration of a conventional image decoding device. This image decoding apparatus includes a code input buffer 201, a variable length decoding (V
LD) unit 202, inverse quantization (IQ) unit 203, inverse DCT
(IDTC) unit 204, INTRA / INTER determination unit 205, motion compensation prediction unit 207, reconstructed image generation unit 21
3, an image memory 215, an image output buffer 216, and a decoded image display unit 217 that displays an image of QCIF size.
Equipped with. Further, the motion compensation prediction unit 207 includes a motion vector calculation unit 208 and a reference image acquisition unit 210.

In this image decoding device, an image memory 215 for storing a reference image for motion compensation prediction and a decoded image display unit 217 correspond to image data of QCIF size, and code data of QCIF size are stored. It is possible to decrypt.

Code input buffer 201 receives code data compressed by motion compensation predictive coding and DCT. The VLD unit 202 has a code input buffer 201.
Variable length decoding for decoding various parameters and DCT coefficient data is performed from the code data input to. The inverse quantization unit 203 inversely quantizes the quantized DCT coefficient. Reverse D
The CT unit 204 performs inverse DCT conversion on the inversely quantized DCT coefficient data. INTRA / INTER determination unit 20
5 determines whether each macroblock is INTRA (intra-frame coding) or INTER (inter-frame coding).

The operation of the conventional image decoding apparatus having the above configuration will be described. FIG. 7 is a flowchart showing a conventional image decoding operation processing procedure. First, the code data compressed by motion compensation predictive coding and DCT is input to the code input buffer 201 (step S101).

With respect to the code data input to the code input buffer 201, various parameters and DC are applied in the VLD section 202.
Variable length decoding for decoding the T coefficient data is performed (step S102). The inverse quantization unit 203 inversely quantizes the quantized DCT coefficient for the decoded data (step S103). Further, the inverse DCT unit 204 performs inverse DCT conversion of the inversely quantized DCT coefficient data (step S
104).

After this, the INTRA / INTER judgment unit 2
In 05, it is determined for each macroblock whether it is INTRA (intra-frame coding) or INTER (inter-frame coding) (step S105). When it is determined to be INTRA (intra-frame encoding), the image data after the inverse DCT conversion is finished is used as a decoded image in the image memory 2
15 (step S110). Then, the image data is output to the image output buffer 216 (step S111) and displayed on the decoded image display unit 217 (step S112). Then, the process returns to step S102.

On the other hand, when it is determined to be INTER (interframe coding) in step S105, the motion vector calculation unit 208 locally decodes the reconstructed image and the macroblock of the current image (the image that is temporally previous). ) And a macro vector are compared to calculate a motion vector (step S107). Based on the calculated motion vector, the reference image acquisition unit 210 extracts a predicted reference image from the reconstructed image (step S108).

Then, INTER after the end of the inverse DCT conversion is completed.
The predicted reference image extracted from the reconstructed image is added to the image data (decoded image) to generate a new reconstructed image (step S109), and the generated reconstructed image is stored in the image memory 215 in step S110. . further,
The image data is output to the image output buffer 216 and displayed on the decoded image display unit 217 (steps S111 and S11).
2). Then, the process returns to step S102.

[0013]

However, the above-mentioned conventional image decoding apparatus has the following problems, and improvement thereof is demanded. That is, when decoding code data having an image size (CIF size) larger than a specific image size (QCIF size), the capacity of the image memory for storing the reference image used for motion compensation prediction becomes insufficient, and However, there is a problem in that the decoding result (decoded image data) cannot be displayed as it is on the decoded image display unit.

The present invention has been made in view of the above circumstances, and is an image capable of decoding coded data having an image size larger than a specific image size without increasing the capacity of the image memory or changing the decoded image display unit. An object is to provide a decoding device, an image decoding method, and an image decoding program.

[0015]

An image decoding apparatus according to the present invention is an image decoding apparatus for decoding coded data in which a moving image is compressed and storing the decoded image data in an image memory. Input means for inputting the encoded code data, variable length decoding means for variable length decoding the input code data, dequantization means for dequantizing the data after the variable length decoding, and dequantization Inverse DCT conversion means for performing an inverse DCT conversion on the subsequent data, motion vector calculation means for calculating a motion vector when motion compensation prediction is applied to the coding of the coded data, and the calculated motion vector by a predetermined value. Motion vector reduction conversion means for reducing conversion to a ratio, and reference image data is fetched from the decoded image data stored in the image memory based on the reduced motion vector. Illuminated image capturing means, enlarging means for enlarging the captured reference image data to the original image size, and adding the enlarged reference image data to the data after the inverse DCT conversion, and the image data after decoding. Image generation means for generating, image reduction means for reducing the generated decoded image data to the predetermined ratio, and storage means for storing the reduced decoded image data in the image memory. It is characterized by having and.

Further, it is characterized by further comprising display means for displaying the decoded image data which is reduced to the predetermined ratio and stored in the image memory.

Further, the enlarging means, when enlarging to the original image size, with respect to the fetched reference image data,
It is characterized by performing linear interpolation and copying.

Further, there is provided motion compensation prediction determination means for determining whether or not motion compensation prediction is applied to the encoding of the input code data, and size determination means for determining the image size of the code data. The predetermined ratio may be determined according to the determined image size and the capacity of the image memory.

The present invention is also an image decoding method for decoding coded data in which a moving image is compressed, and storing the decoded image data in an image memory, which is an input for inputting the compressed coded data. Step, variable-length decoding step of variable-length decoding the input coded data, dequantization step of dequantizing the variable-length decoded data, and inverse DCT transform of the dequantized data. Reverse DC
T conversion step, and when motion compensation prediction is applied to the encoding of the coded data, a motion vector calculation step of calculating a motion vector, and a motion vector reduction conversion of reducing the calculated motion vector to a predetermined ratio. Step, a reference image capturing step of capturing reference image data from the decoded image data stored in the image memory based on the reduction-converted motion vector, and an original image of the captured reference image data An enlarging step for enlarging the size and the inverse D of the enlarged reference image data.
An image generation step of generating decoded image data in addition to the CT-converted data, an image reduction step of reducing the generated decoded image data to the predetermined ratio, and the reduced size. An image decoding method, comprising: storing the decoded image data in the image memory; and an image decoding program capable of being executed by a computer.

According to the present invention having the above-mentioned configuration, the motion vector is converted to control the amount of reference image data to be taken in, thereby increasing the capacity of the image memory and changing the decoded image display unit without changing the specific image. It is possible to decode coded data having an image size larger than the size. Therefore, the memory capacity used for image decoding can be reduced, the display size can be suppressed, and the circuit scale and cost can be reduced. Further, even if image size conversion such as image reduction or enlargement is performed, the decoded image display unit corresponds to a specific image size, and a reduced image is displayed, resulting in a large visual deterioration in image quality. Disappears.

[0021]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. Embodiments of an image decoding device, an image decoding method, and an image decoding program of the present invention will be described.
A description will be given with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the image decoding apparatus according to the embodiment. This image decoding device includes a code input buffer 1, a variable length decoding (V
LD) unit 2, inverse quantization (IQ) unit 3, inverse DCT (IDC)
T) section 4, reconstructed image generation section 5, INTRA / INTE
R determination unit 6, CIF / QCIF determination units 7, 8, CIF →
QCIF reduction unit 9, motion compensation prediction unit 10, image memory 1
1, an image output buffer 12, and a decoded image display unit 13 are provided.

Further, the motion compensation prediction unit 10 uses the QCIF →
CIF enlargement unit 14, reference image acquisition unit 15, reference image 1/4 acquisition unit 16, motion vector 1/4 conversion unit 17
And a motion vector calculation unit 18.

In this image decoding apparatus, the image memory 11 for storing the reference image for performing motion compensation prediction and the decoded image display unit 13 correspond to the image data of QCIF size, which is four times as large as the QCIF size. It is possible to decode CIF size coded data.

Image data (coded data) compressed by motion compensation predictive coding and DCT is input to the code input buffer 1. The VLD unit 2 receives various parameters and DCT from the code data input to the code input buffer 1.
Variable length decoding for decoding coefficient data is performed.

The dequantizer 3 dequantizes the quantized DCT coefficient. The inverse DCT unit 4 performs inverse DCT conversion on the inversely quantized DCT coefficient data. INTRA / INTE
The R determination unit 6 determines whether the macroblock is INTRA (intra-frame coding) or INTER (inter-frame coding).
Is determined.

The CIF / QCIF judging sections 7 and 8 judge whether the input code data is compressed CIF size image data or compressed QCIF size image data.

FIG. 2 is a block diagram showing the configuration of the QCIF → CIF expanding section 14. QCIF → CIF expansion unit 14
Is composed of a sub-scanning (Y) direction copying unit 21, a sub-scanning (Y) direction expanding unit 22, a main scanning (X) direction copying unit 23, and a main scanning (X) direction expanding unit 24.

In this embodiment, the variable length coding (VL
D) section 2, inverse quantization (IQ) section 3, inverse DCT section 4, reconstructed image generation section 5, INTRA / INTER section 6, CIF
The / QCIF units 7 and 9 and the motion compensation prediction unit 10 are realized by a CPU (not shown) executing a control program stored in a ROM (not shown). Each of these units is a dedicated hardware circuit. Of course, it may be realized by.

The operation of the image decoding apparatus having the above configuration will be described. FIG. 3 is a flowchart showing the image decoding operation processing procedure. This processing program, as described above,
It is stored in M and executed by the CPU. In addition, as described above, the image memory 11 storing the reference image for motion compensation prediction and the decoded image display unit 13 have Q
Although it corresponds to CIF size image data, in this image decoding process, C which is four times the QCIF size is used.
It is possible to decode IF size coded data.

First, image data (code data) compressed by motion compensation predictive coding and DCT is input to the code input buffer 1 (step S1). The VLD unit 2 performs variable length decoding for decoding various parameters and DCT coefficient data from the input code data (step S
2).

Further, the inverse quantization unit 3 inversely quantizes the quantized DCT coefficient (step S3), and the inverse DCT unit 104
Then, inverse DCT conversion of the inversely quantized DCT coefficient data is performed (step S4). Then, the CIF / QCIF determination unit 5 determines whether the input code data is compressed CIF size image data or compressed QCIF size image data (step S5).

In both cases of code data obtained by compressing image data of CIF size and code data obtained by compressing image data of QCIF size, the INTRA / INTER determination unit 6 performs INTRA ((intra-frame encoding) for each macroblock. Or the code data of INTER (interframe coding) is determined (steps S6 and S13).

First, the case where it is determined in step S5 that QCIF size code data has been input will be described. In step S6, the INTRA / INTER judgment unit 6 performs IN
When it is determined that the macro block is TRA (intra-frame encoding), the image data after the inverse DCT conversion is stored in the image memory 11 as a decoded image (reconstructed image) (step S10) and reconstructed. The image is output to the image output buffer 12 and displayed on the decoded image display unit 13 (steps S11 and S12). After this, step S2
Return to processing.

On the other hand, when it is determined that the macroblock input in step S6 is INTER (interframe coding), the motion vector calculation unit 18 locally reconstructs the reconstructed image with the macroblock of the current image. The motion vector is calculated by comparing it with the macro block of the previous image) (step S7). Then, the reference image acquisition unit 15 extracts a predicted reference image from the reconstructed image based on the calculated motion vector (step S8). Furthermore, after the inverse DCT conversion is completed (INTER)
A predicted reference image extracted from the reconstructed image is added to the decoded image to generate a reconstructed image (step S9). The generated reconstructed image is stored in the image memory 11 (step S10), this image is output to the image output buffer 12 and displayed on the decoded image display unit 13 (steps S11 and S1).
2). Then, the process returns to step S2.

On the other hand, the case where it is determined in step S5 that the code data of the CIF size is input is shown. QCIF
When it is determined in step S13 that the macro block is INTRA (intra-frame encoding), the CIF → QCIF reducing unit 9 is used as in the case where the code data of the size is input.
Reduce the image data to 1/4 (step S1
9) and store it in the image memory 11 (step S
10), this image is output to the image output buffer 12 and displayed on the decoded image display unit 13 (steps S11 and S1).
2). Then, the process returns to step S2.

On the other hand, when it is determined in step S13 that the macro block is INTER (interframe coding), the motion vector calculation unit 18 calculates a motion vector (step S14), and the motion vector 1/4 conversion unit 17
To set the motion vector to half in the X and Y directions,
That is, it is converted into a quarter (step S15).

Based on the quarter motion vector converted by the reference image 1/4 loading section 16, the reference image 1/4
One-half the predicted reference image is extracted (step S16).
The extracted predicted reference image is QCIF → CIF enlargement unit 14
Is expanded four times (step S17), and the predicted reference image expanded four times is added to the (INTER) decoded image after the inverse DCT conversion is completed to generate a reconstructed image (step S18).

Then, the CIF size is changed to the QC again so as to match the image memory 11 and the decoded image display unit 13.
The image data is reduced to the IF size and stored in the image memory 11 (step S10), and the reduced image is output to the image output buffer 12 and displayed on the decoded image display unit 13 (steps S11 and S12). Then, the process returns to step S2.

FIG. 4 shows the QCIF →
It is a flowchart which shows the CIF expansion process procedure. Figure 5
FIG. 6 is a diagram showing an enlargement process of image data. Since the CIF size is four times as large as the QCIF size, a macro block (8 × 8 pixels) is used when the predicted reference image is captured.
The image data of 1/4 of the above is captured. That is, the image 401 before enlargement is a 4 × 4 pixel block.

First, in the sub-scanning (Y) direction copying section 21,
The image 401 before enlarging is shifted by one pixel in the sub-scanning direction and copied (step S21). At this time, for the last line, the data of the last line of the image 401 before enlargement is copied, and the image after the copy becomes the image 402.

Then, the sub-scanning (Y) direction enlarging section 22 adds the image 401 before enlarging and the image 402 after copying in the sub-scanning direction and divides it by a value of 2, and the data of each line is enlarged before enlarging. The image 401 is inserted below the corresponding line of the image 401 to obtain the enlarged image 403 (step S22). In the figure, the numbers with the subscript "y" indicate the pixels inserted after the above-mentioned arithmetic processing. As a result, the image is magnified twice.

Further, the main scanning (X) direction copying section 23 shifts the image 403 enlarged in the sub scanning direction by one pixel in the main scanning direction and copies it (step S23). At this time, as for the last column, the image 40 enlarged in the sub-scanning direction is displayed.
The data of the last column of No. 3 is copied, and the image 404 after copying is obtained.

Finally, the main scanning (X) direction enlargement unit 24 adds the image 403 enlarged in the sub-scanning direction and the image 404 shifted by one pixel in the main scanning direction, and divides by 2 to obtain each column. Is inserted into the right side of the corresponding column of the enlarged image 403 in the sub-scanning direction to form the enlarged image 405 (step S24). In the figure, the numbers with subscripts “x” and “z” respectively represent the pixels inserted after the above-mentioned arithmetic processing. As a result, the image is magnified 4 times. After that, the process returns to the main process.

As described above, in the present embodiment, the code data of the CIF size which is four times larger than the QCIF size can be decoded without increasing the capacity of the image memory 11 or changing the display size of the decoded image display unit 13. Is possible. Therefore, the circuit scale and cost can be reduced.

Further, CIF → QCIF reduction, QCIF →
Even if the image size conversion such as CIF enlargement is performed, the decoded image display unit 13 corresponds to the QCIF size and displays the reduced image, so that the image quality does not deteriorate significantly in appearance.

The above is a description of the embodiments of the present invention, but the present invention is not limited to the configurations of these embodiments, and the functions shown in the claims or the configurations of the embodiments are not limited thereto. Any structure can be applied as long as it can achieve the function it has.

Further, when the present invention is achieved by supplying a program code of software for realizing the functions of the above-described embodiments to the apparatus, the program code itself realizes the novel function of the present invention. The program itself and the storage medium storing the program form the present invention.

In the above embodiment, the program code shown in the flowcharts of FIGS. 3 and 4 is a storage medium ROM.
It is stored in. The storage medium for supplying the program code is not limited to the ROM, but may be a flexible disk, a hard disk, a non-volatile memory card, or the like.

As described above, according to the present embodiment, the motion vector is converted to control the amount of reference image data to be taken in, so that it is possible to specify the image without increasing the capacity of the image memory or changing the decoded image display section. It is possible to decode coded data having an image size larger than the image size. Therefore,
The memory capacity used for image decoding can be reduced, the display size can be reduced, and the circuit scale and cost can be reduced.

Further, even if image size conversion such as image reduction or enlargement is performed, the decoded image display section corresponds to a specific image size, and the reduced image is displayed, so that the image quality is significantly deteriorated visually. There is an effect that it does not bring.

[0051]

As described above, according to the present invention, coded data having an image size larger than a specific image size can be decoded without increasing the capacity of the image memory or changing the decoded image display section. There is an effect that a possible image decoding device, image decoding method, and image decoding program can be provided.

[Brief description of drawings]

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

FIG. 2 is a block diagram showing a configuration of a QCIF → CIF expanding unit.

FIG. 3 is a flowchart showing an image decoding operation processing procedure.

FIG. 4 is a flowchart showing a QCIF → CIF enlargement processing procedure in FIG.

FIG. 5 is an operation explanatory diagram showing an enlargement process of image data.

FIG. 6 is a block diagram showing a configuration example of a conventional image decoding device.

FIG. 7 is a flowchart showing a conventional image decoding operation processing procedure.

[Explanation of symbols]

5 Reconstructed image generator 6 INTRA / INTER judgment section 7, 8 CIF / QCIF decision section 9 CIF → QCIF reduction unit 11 Image memory 13 Decoded image display section 14 QCIF → CIF expansion section 16 Reference image 1/4 capture section 17 Motion vector 1/4 converter

   ─────────────────────────────────────────────────── ─── Continued front page    F term (reference) 5C059 KK08 LB11 MA04 MA05 MA23                       MC11 ME01 NN21 TA01 TA08                       TB07 TC12 TC25 UA05 UA34                 5J064 AA04 BA09 BB03 BB05 BC01                       BC08 BC14 BC16 BC29 BD03                       BD04

Claims (12)

[Claims] 1. An image decoding device for decoding coded data in which a moving image is compressed, and storing the decoded image data in an image memory, the inputting means for inputting the compressed coded data, Variable-length decoding means for variable-length decoding the input code data, dequantizing means for dequantizing the data after variable-length decoding, and inverse DCT converting means for performing inverse DCT conversion on the dequantized data. When motion compensation prediction is applied to the encoding of the coded data, a motion vector calculation unit that calculates a motion vector, and a motion vector reduction conversion unit that reduces the calculated motion vector to a predetermined ratio, Reference image capturing means for capturing reference image data from the decoded image data stored in the image memory based on the reduced and converted motion vector; Enlarging means for enlarging the illuminated image data to the original image size; Image generating means for adding the enlarged reference image data to the data after the inverse DCT conversion to generate decoded image data; An image, comprising: an image reducing unit that reduces the decoded image data after decoding to a predetermined ratio; and a storage unit that stores the reduced decoded image data in the image memory. Decoding device. 2. The image decoding apparatus according to claim 1, further comprising a display unit that displays the decoded image data stored in the image memory and reduced to the predetermined ratio. 3. The image decoding apparatus according to claim 1, wherein the enlarging means performs linear interpolation and copying on the captured reference image data when enlarging to the original image size. 4. A motion compensation prediction determination unit that determines whether or not motion compensation prediction is applied to encoding of the input code data, and a size determination unit that determines an image size of the code data, The image decoding apparatus according to claim 1, wherein the predetermined ratio is determined according to the determined image size and the capacity of the image memory. 5. An image decoding method for decoding coded data in which a moving image is compressed, and storing the decoded image data in an image memory, the input step of inputting the compressed coded data, A variable length decoding step of variable length decoding the input code data, an inverse quantization step of inverse quantizing the data after the variable length decoding, and an inverse DCT conversion step of performing an inverse DCT conversion of the inversely quantized data. A motion vector calculation step of calculating a motion vector when motion compensation prediction is applied to the encoding of the code data, and a motion vector reduction conversion step of reducing the calculated motion vector to a predetermined ratio, Based on the reduced and converted motion vector, reference image data is acquired from the decoded image data stored in the image memory. A step of enlarging the captured reference image data to an original image size; an image for generating image data after decoding by adding the enlarged reference image data to the data after the inverse DCT conversion A generating step; an image reducing step of reducing the generated decoded image data to the predetermined ratio; and a storing step of storing the reduced decoded image data in the image memory. An image decoding method characterized by: 6. The image decoding method according to claim 5, further comprising a display step of displaying the decoded image data reduced to the predetermined ratio and stored in the image memory. 7. In the enlarging step, when enlarging to the original image size, with respect to the captured reference image data,
The image decoding method according to claim 5, wherein linear interpolation and copying are performed. 8. A motion compensation prediction determination step of determining whether or not motion compensation prediction is applied to encoding of the input code data, and a size determination step of determining an image size of the code data. The image decoding method according to claim 5, wherein the predetermined ratio is determined according to the determined image size and the capacity of the image memory. 9. An image decoding program executed by a computer for decoding coded data in which a moving image is compressed, and storing the decoded image data in an image memory, wherein the compressed coded data is input. An input step, a variable length decoding step for variable length decoding the input code data, an inverse quantization step for inverse quantizing the data after the variable length decoding, and an inverse DCT transform for the inversely quantized data An inverse DCT conversion step, and a motion vector calculation step for calculating a motion vector when motion compensation prediction is applied to the encoding of the code data, and a motion vector for reducing and converting the calculated motion vector into a predetermined ratio. Reduction conversion step, from the decoded image data stored in the image memory based on the reduction-converted motion vector A reference image capturing step of capturing the illumination image data; an enlarging step of enlarging the captured reference image data to an original image size; and a decoding of the enlarged reference image data in addition to the data after the inverse DCT conversion. An image generation step of generating image data after conversion, an image reduction step of reducing the generated image data after decoding to the predetermined ratio, and the reduced image data after decoding in the image memory An image decoding program, comprising: 10. The image decoding program according to claim 9, further comprising a display step of displaying the decoded image data which is reduced to the predetermined ratio and stored in the image memory. 11. The image decoding program according to claim 9, wherein, in the enlarging step, when the image is enlarged to an original image size, linear interpolation and copying are performed on the captured reference image data. 12. A motion compensation prediction determination step of determining whether motion compensation prediction is applied to encoding of the input code data, and a size determination step of determining an image size of the code data, The image decoding program according to claim 9, wherein the predetermined ratio is determined according to the determined image size and the capacity of the image memory.