JPH11234638A - Image coding device, its method and storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To allow a video telephone set to attain image transfer at a low bit rate and with high image quality. SOLUTION: A block conversion circuit 10 divides an input image into a plurality of blocks placed concentrically with respect to a center block in the middle of the screen. An orthogonal transform circuit 14 uses discrete cosine transform(DCT) when a motion is detected in an image in the middle within the screen and uses lapped orthogonal transform(LOT) in the case of a still image when no motion is detected. A quantization circuit 16 applies a recommended quantization table to a center block of the screen when no motion is detected at a peripheral part and the motion is comparatively concentrated on the screen center, and it applies a quantization table whose values are higher than those of the recommended quantization table as a variable quantization method. When any motion is detected at the peripheral part of the screen, non-variable quantization, where the same quantization table is applied to all the blocks within the screen, is used.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image encoding apparatus, method, and storage medium, and more particularly, to an image encoding apparatus, method, and storage medium suitable for a video conference, a video phone, and the like.

[0002]

2. Description of the Related Art In recent years, data networks such as ISDN and the Internet have been developed in Japan.
As a result, information services such as videoconferencing and videophone have become available. The image compression technology used for these services is JPEG for still image compression.
(Joint Photographic Expert
t Group) method is used, and in a moving image, H.264 is used. 2
61 system and MPGE (Moving Picture)
An Expert Group method is generally used. H. The H.261 method is 64 kbp of the ISDN line.
It is intended for moving image transmission at a low bit rate such as s. The MPEG system is based on H.264. This is an evolution from the H.261 system, taking into account commonality.

The MPEG system currently has three levels. Level 1 is mainly for videophone and video conference use, and the recommended rate is 15 Mbps.
It is. Level 2 is targeted for digital transmission of video signals, with a recommended rate of 10 Mbps. Level 3 is for next-generation television, that is, high-definition television signals.

The features of the H261 system and the MPEG system are as follows.
Discrete Cosine Transform DCT
NeTransform), which is a point that an inter-frame prediction process for compression in the time axis direction is added to a JPEG still image compression technology based on orthogonal transform represented by ne Transform. The effect of this inter-frame processing is basically enhanced by performing motion compensation prediction. The macroblock sizes for DCT and motion compensation are 8
× 8 pixels and 16 × 16 pixels.

[0005]

Problems to be Solved by the Invention In the H.261 system,
One screen is divided into 12 blocks as shown in FIG. 6, and one block is further divided into 16 blocks as shown in FIG.
It is divided into 33 (= 3 × 11) macroblocks of × 16 pixels. In the MPEG system, as shown in FIG.
A plurality of blocks obtained by dividing the image data of the screen at arbitrary positions are horizontally scanned from the upper left to the lower right of the screen.

[0006] Such a block arrangement is considered to be effective for an image such as a commercial television, but the movement of the speaker is concentrated in one part and the other part is almost stationary as in a videophone and a video conference. This is not an efficient block operation for the image that is being processed. In other words, if portions having similar properties belong to the same block, codes can be reduced in terms of motion compensation and quantization.

An object of the present invention is to provide an image encoding apparatus and method and a storage medium capable of achieving a low bit rate and / or high image quality.

[0008]

According to the present invention, there is provided an image coding apparatus comprising: a block converting means for dividing a screen into a plurality of blocks; and an encoding means for coding image data for each block by the block converting means. Wherein the block by the block converting means comprises a central block at the center of the screen when a person is photographed and a plurality of blocks located around the central block.

An image encoding method according to the present invention comprises a block conversion step of dividing a screen into a plurality of blocks, and an encoding step of encoding image data for each block by the block conversion step. The block in the conversion step includes a central block at the center of the screen when a person is photographed and a plurality of blocks located around the central block.

The image encoding method according to the present invention can be stored as program software in a storage medium that can be externally stored.

[0011]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a schematic block diagram of an image coding apparatus according to an embodiment of the present invention. FIG. 2 is a schematic block diagram of an image decoding apparatus corresponding to the image coding apparatus shown in FIG. Is shown.

The configuration of the image coding apparatus shown in FIG. 1 will be described. A block conversion circuit 10 divides image data input from an image memory (not shown) into coding blocks.
2 subtracts the predicted value of the predicted frame in the case of inter-frame prediction coding from the image data of each block output from the block conversion circuit 10, and outputs the output of the block conversion circuit 10 in the case of intra-frame coding. Subtractor for through output,
14 is an orthogonal transform circuit for orthogonally transforming the output of the subtracter 12, 16 is a quantization circuit for quantizing the output of the orthogonal transform circuit 16, 18 is an entropy encoding circuit for entropy encoding the output of the quantization circuit 16, Reference numeral 20 denotes a buffer for buffering output data of the entropy encoding circuit 18 and transmitting the data to a transmission path.

Reference numeral 22 denotes a motion detection circuit for detecting a motion from the output data of the block conversion circuit 10, and 24 locally decodes the output of the quantization circuit 16 and refers to the motion vector output from the motion detection circuit 22. This is a motion inter-frame prediction circuit that outputs a motion-compensated prediction value to the subtractor 12.

The motion inter-frame prediction circuit 24 includes an inverse quantization circuit 26 for inversely quantizing the output of the quantization circuit 16 and an inverse orthogonal transformation circuit 2 for inversely orthogonally transforming the output of the inverse quantization circuit 26.
8. In the case of inter-frame predictive coding, an inverse orthogonal transform circuit 2
8 in the case of intra-frame coding, an adder 30 that outputs the output of the inverse orthogonal transform circuit 28 through-output, and an adder 30 according to the motion vector detected by the motion detection circuit 22. And a prediction circuit 32 for calculating a prediction value from the output of

In the present embodiment, the orthogonal transform circuit 14 performs discrete cosine transform (DCT) and overlap orthogonal transform (Lapped).
Orthogonal Transform), and the quantization circuit 16 appropriately selects variable quantization and non-variable quantization. Corresponding to the orthogonal transformation circuit 14 and the quantization circuit 16, the inverse quantization circuit 26 of the motion inter-frame prediction circuit 24 has variable inverse quantization and non-variable inverse quantization, and the inverse orthogonal transformation circuit 28 has inverse DCT. It has an inverse LOT. Although the details will be described later, the combination of the orthogonal transformation and the quantization is determined according to the magnitude of the motion vector detected by the motion detection circuit 22.

Before describing the operation of the present embodiment, conditions necessary for solving the problems of the present application will be briefly described. In order to solve the above-described problems, first, efficient blocking is performed, second, an optimal quantization table is allocated (variable quantization), and third, block noise that is conspicuous when the screen is stationary. It is necessary to consider the three points of suppressing.

For efficient block formation, basically, assuming that the speaker is located at the center of the screen, the vicinity of the face where the movement is most concentrated may be formed into the same block. And
By allocating other blocks concentrically from that block, the lesser moving parts are accommodated in the same block. By doing so, image portions having similar properties can be included in the same block, and coding efficiency can be improved.

The reason why each block is arranged concentrically from the center block is that human visual characteristics are taken into consideration. That is, all the visual information shown in the eyes is not uniformly transmitted to the brain, but the information shown in the fovea in the eyeball is transmitted to the brain with higher accuracy. Since the fovea is not wide, humans supplement visual information with eye movements. Therefore, most of the visual information visible to the eyes is "visible", and the "viewing" portion is small. In a videophone or videoconference, the "viewing" area is usually near the face of the speaker at the center of the screen, and concentric blocks are formed so as to gradually move to the "viewing" area. From this point of view, the block conversion circuit 10 of the present embodiment uses the face image shown in FIG.
The image data is subjected to block conversion with a block configuration as shown in FIG.

For optimal allocation of the quantization table, the recommended quantization table is adopted for the center block corresponding to the speaker's face, and the recommended quantization table is used for the blocks concentrically extending around the center block. Use a table that is several times larger. That is, a variable quantization in which a value is changed for each block is applied. This is a result of focusing on the fact that the visual characteristics of a human are focused on the center of the screen as described above, and the outside is visible but not so conscious. However, when the speaker in the screen moves, the viewpoint of the other party who gazes at the screen follows the speaker and deviates from the center of the screen. In that case, the recommended quantization table is used also for the concentric peripheral block portion.

In the present embodiment, when the recommended quantization table is used for all blocks of the entire screen as described above, it is referred to as non-variable quantization in distinction from the above-described variable quantization.
Although details will be described later, whether to select variable quantization or non-variable quantization is determined by a motion vector in a peripheral block detected by the motion detection circuit 22. Usually, since the speaker in the screen is at a fixed position, the motion vector of the concentric peripheral block is considered to be extremely small, but when the speaker in the screen changes posture, the motion of the concentric peripheral block is This is because the vector becomes large. In other words, when the speaker in the screen moves at a substantially constant position and the movement concentrates on the central block of the face position, the motion is small in the peripheral blocks, and the motion vector amount is small, the variable quantization is used as the recommended quantization table for the peripheral blocks. Use a number obtained by multiplying If the speaker in the screen changes his posture and the motion vector amount of the peripheral block increases, the recommended quantization table is used for all blocks as non-variable quantization. And
When the speaker returns to the center of the screen and stops, the variable quantization is used again.

Block noise is suppressed as follows. That is, according to ITU-T Recommendation H.264. 261 and ISO /
IEC MPEG (Moving Picture E
All international standard video coding schemes, such as xparts Group, use DCT. However, DCT
As described above, since the conversion is completely closed within a macroblock (8 × 8 pixels) as described above, there is a drawback that discontinuity of the macroblock occurs in encoding at a low bit rate.

These block noises are not particularly noticeable from the viewpoint of human vision when the screen is constantly changing, but are particularly noticeable when the screen is stationary. In a videophone call, a video conference, or the like, a situation in which the screen is stopped halfway may be considered. For example, this is a case where a graph is shown in a presentation or a photograph is referred to.

In the present embodiment, in the still picture state, the overlapping orthogonal transform LOT (La
ppped Orthogonal Transfer
m). Here, the block noise reduction method and the LOT proposed previously have been briefly described. Methods of reducing block noise that have been proposed in the past include an overlap method in which blocks overlap each other and a filtering method in which a low-pass filter is applied to boundary pixels on the decoder side. The former has a disadvantage that the bit rate is high, and the latter has a drawback that the block boundaries are blurred.

[0025] Whereas the overlap scheme overlaps the image in the spatial domain, LOT not only reduces block noise because the transform basis overlaps adjacent blocks, but also reduces the specific transform. Power Concentration Ratio to Energy (Energy Compaction)
n) is better than DCT. Moreover, since the basis matrix is determined so that the number of samples for encoding is the same as that of DCT, additional bits are unnecessary. In other words, the conventional orthogonal transform DCT transform circuit can be replaced with LOT transform and used as it is. However, the calculation time required for the LOT is about 30% longer than that of the DCT, and the calculation load increases in the moving image processing state. Therefore, the calculation time is not applied to the moving image processing as much as possible.

Summarizing the above, in this embodiment, LOT conversion is mainly used in a still image state where the screen is stopped,
When movement occurs again due to a speaker in the screen speaking, DCT conversion is mainly used. The switching is determined with reference to the motion vector, particularly the motion vector of the central block. This is because the center block has the most movement (such as the movement of the mouth and the change of the facial expression). By referring to the motion vector amount of this block, it is determined whether the screen is stationary or the speaker has movement. Can be easily determined. However, when applying the LOT transform, the same quantization table is used for all blocks in the screen (non-variable quantization). This is to improve the image quality of a still image as described above.

The basic operation of this embodiment will be described. For the reason described above, the block conversion circuit 10 of the present embodiment is used.
Converts the image data into blocks in a block configuration as shown in FIG. 3B along the face image shown in FIG. 3A having a face at the center of the screen. The image data block-converted by the block conversion circuit 10 is applied to the subtractor 12 and the motion detection circuit 22. The motion detection circuit 22 obtains a motion vector from the block-converted image data from the block conversion circuit 10. The motion vector obtained by the motion detection circuit 22 is supplied to the prediction circuit 32 for motion compensation, and is supplied to the entropy encoding circuit 18 for transmission with the compressed image data. The motion vector obtained by the motion detection circuit 22 is also supplied to the orthogonal transformation circuit 14 and the quantization circuit 16 for selecting the orthogonal transformation method in the orthogonal transformation circuit 14 and the quantization method in the quantization circuit 16.

The subtractor 12 subtracts the motion-compensated predicted value from the inter-frame prediction circuit 24 from the image data from the block conversion circuit 10, and applies the subtracted value to the orthogonal transformation circuit 14. The orthogonal transform circuit 14 orthogonally transforms the difference image data output from the subtractor 12 and outputs the transform coefficients to the quantization circuit 16.

As described above, the orthogonal transform circuit 14
Has a DCT operation unit and a LOT operation unit, and determines which one to apply according to the amount of motion vector from the motion detection circuit 22. Specifically, a threshold Ta is set for the motion vector amount in the central block A shown in FIG. 3B, and when the vector amount Va of the block A exceeds the threshold Ta (Va> Ta), the screen Is determined to be a moving image state in which a person is moving, and DCT calculation is applied to increase the speed. Conversely, when the motion vector amount Va is equal to or less than the threshold value Ta, it is determined that the screen is in a stationary state, and the LOT calculation is applied to improve the image quality.

The quantization circuit 16 converts the transform coefficient data from the orthogonal transform circuit 14 into a quantization table ROM (not shown).
Are quantized according to the quantization table stored in. However, as described above, in the present embodiment, the quantization circuit 1
6 refers to the motion vector from the motion detection circuit 22;
If motion is relatively concentrated in the center of the screen without any movement in the periphery of the screen, apply variable quantization, and use bits obtained by multiplying the recommended quantization table by several times for surrounding blocks. Reduce rate. When motion is detected also in the peripheral portion and there is motion in the entire screen, non-variable quantization is applied, and the recommended quantization table is used for the peripheral blocks, so that H.264 is used. Image quality degradation is reduced using standard image coding methods such as H.261 and MPEG.

The quantization circuit 16 selects variable quantization or non-variable quantization according to the motion vector amount of a peripheral block among the motion vectors for each block obtained by the motion detection circuit 16. Specifically, the sum Vb of the vector amounts of the peripheral blocks is compared with the threshold value Ts, and the vector sum Vb is set to the threshold value Ts.
In the following cases, it is determined that there is no motion in the peripheral portion, and variable quantization is applied. When the value is larger than the threshold value Ts, it is determined that there is also motion in the peripheral portion, and non-variable quantization is applied. In the variable quantization, a quantization table having a value three times the recommended quantization table is applied to the peripheral blocks BF, and a quantization table having a value five times the recommended quantization table is applied to the peripheral blocks GI. Apply. However, this multiple may be arbitrarily determined according to environmental conditions. FIG. 4 is an example of a quantization table of the variable quantization system, and is a recommended quantization table applied to the central block A (FIG. 4A).
And a five-fold quantization table (FIG. 4B) applied to peripheral blocks GI.

FIG. 5 shows the relationship between the motion vector detected by the motion detection circuit 22, the transformation method in the orthogonal transformation circuit 14, and the selection of the quantization table in the quantization circuit 16.

If no motion is detected from the motion vector amount of the central block A (S1), a combination of LOT and non-variable quantization is selected (S2), and higher image quality is prioritized. When a motion is detected in the central block A (S
1) It is determined whether or not there is a motion in the peripheral portion based on the vector amount of the peripheral block (S2). If there is, a combination of DCT and variable quantization is selected (S4), and a lower bit rate is prioritized. When it is determined that there is movement in the peripheral part (S3), that is, when there is movement in the entire screen, a combination of DCT and non-variable quantization is selected, and H.264 is selected. The balance between the bit rate and the image quality according to standard methods such as H.261 and MPEG is selected in a well-balanced manner.

The output of the quantization circuit 16 is applied to an entropy coding circuit 18 and a motion inter-frame prediction circuit 24. The entropy encoding circuit 18 entropy-encodes the output of the quantization circuit 16 and the motion vector from the motion detection circuit 22 according to the value of the entropy encoding table ROM, and writes the encoded data into the data buffer 20. Specifically, the entropy encoding circuit 18
The matrix data output from is converted into a one-dimensional array by zigzag scanning in order from the coefficient in which energy is concentrated for each macroblock, and the Huffman coding or arithmetic coding is performed on the quantized data of the one-dimensional array. A code is assigned based on the statistical frequency of occurrence by using a method such as conversion. The buffer 20 smoothes the rate of the data output from the entropy encoding circuit 18 to the rate of the transmission path of the output destination.

The data transmitted to the transmission path propagates through the transmission path, is input to the receiver shown in FIG. 2, and is decoded.

The inter-motion-frame prediction circuit 24 first locally decodes the output data of the quantization circuit 16. That is, the inverse quantization circuit 26 inversely quantizes the output data of the quantization circuit 16, and the inverse orthogonal transform circuit 28 performs an inverse orthogonal transform on the output of the inverse quantization circuit 26. Of course, the inverse quantization circuit 26 uses the same quantization table as that used in the quantization circuit 16 with reference to the motion vector from the motion detection circuit 22,
The inverse orthogonal transform circuit 28 also adopts an inverse orthogonal transform corresponding to the orthogonal transform method adopted by the orthogonal transform circuit 14 with reference to the motion vector from the motion detection circuit 22. The output of the inverse orthogonal transform circuit 28 corresponds to difference image data in the case of the inter-frame difference encoding method, and the adder 30 adds the predicted value of the predicted frame to the output of the inverse orthogonal transform circuit 28 to perform prediction. Supply to the circuit 32. The prediction circuit 32 refers to the motion vector output from the motion detection circuit 22 and
Is calculated by motion-compensating the output of. However, the prediction circuit 32 calculates a predicted value in the case of inter-frame encoding, but outputs a zero value in the case of intra-frame encoding. The output of the prediction circuit 32 is supplied to the adder 30 and the subtractor 12
Is applied to

A controller (not shown) controls the entire apparatus shown in FIG. 1, in particular, the coding mode and the buffer 2.
Frame thinning or the like based on the stored data amount of 0 is controlled.

The decoding device shown in FIG. 2 will be described. The data from the transmission path is rate-adjusted by the buffer 40 and input to the entropy decoding circuit 42. The entropy decoding circuit 42 decodes the data input from the buffer 40 by a process reverse to that of the entropy coding circuit 18, separates the image information and the motion information, and converts the image information into the inverse quantization circuit 4.
4 and the motion information to the inverse quantization circuit 44, the inverse orthogonal transform circuit 46, and the prediction circuit 50. The inverse quantization circuit 44 operates in exactly the same way as the inverse quantization circuit 26, and inversely quantizes the image information from the entropy decoding circuit 42. The inverse orthogonal transform circuit 46 operates in exactly the same way as the inverse orthogonal transform circuit 28. In operation, the output data of the inverse quantization circuit 26 is subjected to inverse orthogonal transform. The output data of the inverse orthogonal transform circuit 46 is
In the case of intra-frame encoding, the image data itself is used. In the case of inter-frame encoding, the image data is difference image data.

The adder 48 adds a zero value to the output data of the inverse orthogonal transform circuit 46 in the case of intra-frame coding, and outputs the result. , A predicted value is added to the output of the inverse orthogonal transform circuit 46, and the result is output. The output wave reception image data of the adder 48 is output to the outside as well as applied to the prediction circuit 50. The prediction circuit 50 refers to the motion vector from the entropy decoding circuit 42 similarly to the prediction circuit 32,
A predicted value obtained by motion-compensating the output of the adder 48 is calculated, and the predicted value is applied to the adder 48 in the case of inter-frame coding and the zero value in the case of intra-frame coding.

The transmission path should include not only a communication line but also a recording / reproducing system using various recording media, for example, a magneto-optical disk, an optical disk, a magnetic tape, and a magnetic disk. This is because there is no technical difference.

Next, the operation of the characteristic portion of this embodiment will be described in detail. In the motion compensation predictive coding method, a coding device on the transmission side performs block conversion of an input image into a predetermined block format, performs motion compensation predictive coding for each block, and performs entropy coding. More specifically, the block conversion circuit 10 converts the input image data into blocks arranged in blocks along the face image as shown in FIG. That is, FIG.
As shown in (b), a block B, a block C, a block D,... Are arranged concentrically around a block A located substantially at the center of the screen.

As can be seen from FIGS. 3A and 3B, the face is almost covered by the block A. The blocks B to D around the block A include the hair and clothes of the subject person, and the person is symmetrical, so that the right and left portions of the screen of each block have substantially the same characteristics.
Blocks GI are backgrounds, and so are the blocks. Therefore, if limited to video conferencing, video telephony, and the like, the right and left portions of each block become images having almost the same properties, and high coding efficiency can be achieved.

The adder 12 subtracts the prediction result from the motion compensation inter-frame prediction processing from the motion inter-frame prediction circuit 24 from the image data from the block conversion circuit 10, and performs the orthogonal transformation on the subtraction result, ie, the difference image data. Circuit 1
4 The orthogonal transform circuit 14 orthogonally transforms the difference image data from the adder 12. In this embodiment, as described above, it is determined whether the block A is in the moving image state or the still image state with reference to the motion vector of the central block A from the motion detection circuit 22, and DCT processing or LOT processing is selected.

The quantization circuit 16 quantizes the transform coefficient data output from the orthogonal transform circuit 14 according to a quantization table value stored in a quantization table ROM (not shown). In the present embodiment, as described above, referring to the motion vector of the block around the screen from the motion detection circuit 22,
Judgment is made between the case where no movement has occurred in the peripheral portion and the movement is relatively concentrated in the central portion and the case where there is movement in the entire screen, and variable quantization or non-variable quantization is selected.

The motion detecting circuit 22 includes the block converting circuit 1
0 to obtain a motion vector from the image data block-converted by the
Output to 2. The motion inter-frame prediction circuit 24 performs motion compensation prediction based on the motion vector detected by the motion detection circuit 22.

The block arrangement in FIG. 3B is an example,
The number, position, shape, and the like of the blocks are not limited to the illustrated example.

It is apparent that the present embodiment can be realized not only by hardware but also partly or entirely by software.

The present invention can be applied to a system including a plurality of devices (for example, a host computer, an interface device, a reader and a printer, etc.), or to an apparatus including one device (for example, a copier and a facsimile machine). May be.

Further, in order to operate various devices so as to realize the functions of the above-described embodiment, a computer in an apparatus or a system connected to the various devices includes:
The present invention also provides a program code of software for realizing the functions of the above-described embodiments, which is implemented by operating a computer (CPU or MPU) of the device or system according to a stored program to operate the various devices. Included in the scope of the invention.

In this case, the program code itself of the software realizes the functions of the above-described embodiment, and the program code itself and means for supplying the program code to the computer, for example, the program The storage medium storing the code constitutes the present invention. Examples of storage media for storing such program codes include floppy disks, hard disks, optical disks, magneto-optical disks, and CD-ROs.
M, a magnetic tape, a nonvolatile memory card, a ROM, and the like can be used.

When the computer executes the supplied program code, not only the functions of the above-described embodiment are realized, but also the OS (operating system) or other operating system running on the computer. Even when the functions of the above-described embodiments are realized in cooperation with application software or the like, it goes without saying that such program codes are included in the embodiments of the invention according to the present application.

Further, the supplied program code is
After being stored in the memory provided in the function expansion board of the computer or the function expansion unit connected to the computer, the CPU or the like provided in the function expansion board or the function expansion unit performs one of the actual processing based on the instruction of the program code. It goes without saying that a case where the functions of the above-described embodiments are realized by performing the processing of all or part of the units is also included in the invention according to the present application.

[0053]

As can be easily understood from the above description, according to the present invention, high image quality can be obtained even at a low bit rate by performing blocking, orthogonal transformation, and quantization based on the characteristics of the subject image and the visual characteristics of a person. , And high image quality can be realized in a videophone or videoconference even with a low bit rate transmission path. At the time of a still image, by using the LOT operation, it is possible to reduce block noise which is a problem when the DCT operation is used.

[Brief description of the drawings]

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

FIG. 2 is a schematic block diagram of an image decoding device corresponding to the image encoding device shown in FIG.

FIG. 3 is a diagram illustrating an example of a videophone or videoconference screen (a) and a corresponding block arrangement (b) of the present embodiment.

FIG. 4 is an example of a recommended quantization table (a) used for a central block and a quantization table (b) used for peripheral blocks GI in this embodiment.

FIG. 5 is a flowchart of selection of an orthogonal transform method and a quantization method according to the present embodiment.

FIG. 261 is a block layout diagram. FIG.

FIG. 7 is an enlarged view of a block A in FIG. 6;

FIG. 8 is a block diagram layout in the MPEG system.

[Explanation of symbols]

 10: Block transformation circuit 12: Subtractor 14: Orthogonal transformation circuit 16: Quantization circuit 18: Entropy coding circuit 20: Buffer 22: Motion detection circuit 24: Motion inter-frame prediction circuit 26: Inverse quantization circuit 28: Inverse orthogonal Conversion circuit 30: Adder 32: Prediction circuit 40: Buffer 42: Entropy decoding circuit 44: Inverse quantization circuit 46: Inverse orthogonal transformation circuit 48: Adder 50: Prediction circuit

Claims (36)

[Claims] An image processing apparatus includes: a block conversion unit that divides a screen into a plurality of blocks; and an encoding unit that encodes image data for each block by the block conversion unit. An image encoding apparatus comprising a central block at the center of a screen at the time of shooting and a plurality of blocks located around the central block. 2. The image encoding device according to claim 1, wherein peripheral blocks adjacent to the central block are arranged concentrically with respect to the central block. 3. An encoding apparatus comprising: an orthogonal transformation unit; a quantization unit for quantizing an output of the orthogonal transformation unit; an entropy encoding unit for entropy encoding the output of the quantization unit; 2. The image encoding apparatus according to claim 1, further comprising a motion detecting unit configured to detect a motion of the image. 4. The image coding apparatus according to claim 3, wherein said orthogonal transform means comprises a discrete cosine transform means and an overlap orthogonal transform means. 5. The orthogonal transform means uses the discrete cosine transform means when a motion is detected in the image in the center of the screen, and when the motion is not detected and the still image state is reached, the orthogonal transform means uses the discrete cosine transform means. 5. The image coding apparatus according to claim 4, wherein an orthogonal transform unit is used. 6. The image coding apparatus according to claim 5, wherein said orthogonal transform means selects said discrete cosine transform means and said overlap orthogonal transform means based on a motion vector of said central block detected by said motion detecting means. . 7. The orthogonal transform means selects the discrete cosine transform means when the motion vector amount of the central block detected by the motion detecting means exceeds a predetermined threshold, and otherwise selects the overlapping orthogonal transform means. The image encoding device according to claim 6, wherein: 8. The variable quantization means for applying a recommended quantization table to the center block and applying a quantization table having a larger value than the recommended quantization table to peripheral blocks. 8. The image encoding apparatus according to claim 3, further comprising: a non-variable quantization unit that applies the same quantization table to all blocks in the screen. 9. The quantization means applies the non-variable quantization means when motion is detected in a peripheral portion of the screen, and no motion is detected in the peripheral portion, and the motion is relatively concentrated in the central portion of the screen. 9. The variable quantizing means is applied when the variable quantization is performed.
An image encoding device according to claim 1. 10. The variable quantization unit or the non-variable quantization unit according to claim 9, wherein the quantization unit selects the variable quantization unit or the non-variable quantization unit based on a sum of all motion vector amounts of peripheral blocks detected by the motion detection unit. Image coding device. 11. The variable quantizing means applies the variable quantizing means when the sum of all motion vector amounts of peripheral blocks detected by the motion detecting means exceeds a predetermined threshold, and applies the variable quantization means when the sum is equal to or smaller than the predetermined threshold. The image coding apparatus according to claim 10, wherein the non-variable quantization means is applied to the image coding apparatus. 12. When the overlapping orthogonal transform means is selected,
The image encoding device according to claim 8, wherein the non-variable quantization unit is selected. 13. A block transforming step of dividing a screen into a plurality of blocks, and an encoding step of encoding image data for each block in the block transforming step. An image coding method comprising a central block at the center of a screen at the time of shooting and a plurality of blocks located around the central block. 14. The image coding method according to claim 13, wherein peripheral blocks adjacent to the central block are arranged concentrically with respect to the central block. 15. The encoding step includes: a motion detection step of detecting a motion of an image for each block; an orthogonal transformation step; a quantization step of quantizing an output of the orthogonal transformation step; 14. The image encoding method according to claim 13, further comprising an entropy encoding step of entropy encoding the output. 16. The image encoding method according to claim 15, wherein the orthogonal transform step includes a discrete cosine transform step and an overlapping orthogonal transform step. 17. The orthogonal transformation step uses the discrete cosine transformation step when a motion is detected in the image in the center portion of the screen, and when the motion is not detected and the image is in a still image state, the overlapping is performed. 17. The image coding method according to claim 16, wherein an orthogonal transformation step is used. 18. The image coding method according to claim 17, wherein the orthogonal transforming step selects the discrete cosine transforming step and the overlapping orthogonal transforming step based on a motion vector of the central block detected by the motion detecting step. . 19. The orthogonal transforming step comprises selecting the discrete cosine transform step when the motion vector amount of the central block detected by the motion detecting step exceeds a predetermined threshold, and otherwise selecting the overlapping orthogonal transform step. 18. The image encoding method according to claim 17, wherein (i) is selected. 20. The variable quantization step of applying a recommended quantization table to the central block and applying a quantization table having a larger value than the recommended quantization table to peripheral blocks. 20. The image encoding apparatus according to claim 16, further comprising: a non-variable quantization step of applying the same quantization table to all blocks in the screen. 21. The quantization step applies the non-variable quantization step when motion is detected in the peripheral portion of the screen, and no motion is detected in the peripheral portion, and the motion is relatively concentrated in the central portion of the screen. 21. The image encoding method according to claim 20, wherein the variable quantization step is applied when the image coding is performed. 22. The quantization step according to claim 21, wherein the variable quantization step or the non-variable quantization step is selected based on a sum of all motion vector amounts of peripheral blocks detected by the motion detection step. Image coding method. 23. The quantization step, wherein the variable quantization step is applied when the sum of all motion vector amounts of peripheral blocks detected by the motion detection step exceeds a predetermined threshold, and when the sum is equal to or smaller than the predetermined threshold. 23. The image coding method according to claim 22, wherein the non-variable quantization step is applied to the image coding. 24. The non-variable quantization step is selected when the overlapping orthogonal transformation step is selected.
24. The image encoding method according to any one of claims 23 to 23. 25. A block conversion step of dividing the screen into a block including a central block at the center of the screen when a person is photographed and a plurality of blocks located around the block, and an image for each block by the block conversion step. A program comprising an encoding step for encoding data;
A storage medium for storing software so that it can be read externally. 26. The storage medium according to claim 25, wherein peripheral blocks adjacent to the central block are arranged concentrically with respect to the central block. 27. The encoding step, comprising: a motion detection step of detecting a motion of an image for each block; an orthogonal transformation step; a quantization step of quantizing an output of the orthogonal transformation step; 26. The storage medium of claim 25, further comprising: an entropy encoding step of entropy encoding the output. 28. The storage medium according to claim 27, wherein said orthogonal transformation step includes a discrete cosine transformation step and an overlap orthogonal transformation step. 29. The orthogonal transforming step uses the discrete cosine transform step when motion is detected in the image in the center portion of the screen. 29. The storage medium according to claim 28, wherein an orthogonal transformation step is used. 30. The storage medium according to claim 29, wherein in the orthogonal transformation step, the discrete cosine transformation step and the overlapping orthogonal transformation step are selected based on a motion vector of the central block detected in the motion detection step. 31. The orthogonal transforming step, wherein the discrete cosine transform step is selected when the motion vector amount of the center block detected by the motion detecting step exceeds a predetermined threshold, and otherwise the overlapping orthogonal transform step The storage medium according to claim 29, wherein is selected. 32. A variable quantization step of applying a recommended quantization table to the central block and applying a quantization table having a larger value than the recommended quantization table to peripheral blocks. 32. The storage medium according to claim 28, further comprising: a non-variable quantization step of applying the same quantization table to all blocks in the screen. 33. The quantization step applies the non-variable quantization step when motion is detected in a peripheral portion of the screen, and no motion is detected in the peripheral portion, and the motion is relatively concentrated in the central portion of the screen. 33. The storage medium according to claim 32, wherein the variable quantization step is applied when performing. 34. The variable quantization step or the non-variable quantization step according to claim 33, wherein the quantization step selects the variable quantization step or the non-variable quantization step based on a sum of all motion vector amounts of peripheral blocks detected by the motion detection step. Storage medium. 35. The variable quantization step, wherein the variable quantization step is applied when the sum of all motion vector amounts of peripheral blocks detected by the motion detection step exceeds a predetermined threshold, and when the sum is equal to or smaller than the predetermined threshold. 35. The storage medium according to claim 34, wherein said non-variable quantization step is applied. 36. The non-variable quantization step is selected when the overlapping orthogonal transformation step is selected.
36. The storage medium according to any one of items 35 to 35.