JPH06105301A - System and device for encoding moving image - Google Patents

Abstract

(57) [Abstract] [Purpose] To reduce the amount of processing in a moving image encoding system and apparatus by using motion vectors to make various determinations in encoding processing. [Structure] The frame / field determination circuit 15 performs determination using the output from the motion vector detection circuit 14,
Based on the result, the orthogonal transformation circuit 4 performs orthogonal transformation on the difference image data. Further, the encoding circuit 16 uses the output from the motion vector detection circuit 14 to optimize the scanning order within the small block, and the scanning within the small block of the quantized data according to the scanning order. Assigns a variable length code. [Effect] The processing amount in the DCT frame / field determination or the scanning order determination is reduced, the processing speed is improved, and the amount of encoded data generated is reduced.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to a moving image signal,
The present invention relates to a moving picture coding system and apparatus for detecting a motion between frames or fields and performing high efficiency coding by orthogonal transformation.

[0002]

2. Description of the Related Art As an encoding method for moving image signals, an encoding method for videophones and videoconferences has been standardized mainly by the International Telegraph and Telephone Consultative Committee (CCITT), and CD-ROM and the like. The encoding system for the digital storage medium is being standardized mainly by the International Standards Organization (ISO). In those moving image coding systems, band compression is performed on an input signal by orthogonal transformation. Recently, one of them is the discrete cosine transform (hereinafter DCT [Discre
te Cosine Transform]) is the mainstream method.

FIG. 5 shows a block diagram of a moving picture coding method using a DCT for orthogonal transformation as an example of a conventional moving picture coding method, and the operation of each part will be described below.

In the figure, first, original image data of a moving image signal is input from a terminal 1 in an interlaced form,
The adder circuit 2 calculates difference data from the reference image data after the motion compensation processing. The frame / field determination circuit 3 determines, based on the difference data, whether the DCT should be performed in an interlaced state (frame unit) or in a non-interlaced state (field unit), and outputs the determination result. Output.

The DCT circuit 4 performs DCT on the difference data in accordance with the judgment output from the frame / field judgment circuit 3, and the obtained transform coefficient data is quantized in the quantization circuit 5. A code is assigned to the quantized data after the quantization processing together with other data in the encoding circuit 6, and the obtained encoded data is used to equalize the generation amount of the encoded data that fluctuates with time. It is temporarily stored in the buffer memory 7 and sent from the terminal 8 at a predetermined transmission rate.

The inverse quantization circuit 9 and the inverse DCT circuit 10 are the parts for performing the inverse processing of the quantization circuit 5 and the DCT circuit 4, respectively, and restore the original difference data from the quantized data. In the adder circuit 11, the difference data and the reference image data after the motion compensation processing are added, and as a result, locally decoded image data is obtained.

The motion vector detection circuit 14 detects a motion between the reference frame or the reference field based on the original image data of the moving image signal input from the terminal 1,
A motion vector is generated for each unit block described later. The motion compensation circuit 12 performs a motion compensation process by referring to the motion vector based on the locally decoded image data, and thus the reference image data is output. The switch 13 is for selecting '0' when the input signal is encoded within a frame or field to generate reference image data for which motion compensation processing is not performed.

Now, the operation of the frame / field determination circuit 3 will be described in more detail.

First, as shown in FIG. 2, the input difference data for one frame from the adder circuit 2 is made up of several small blocks (for example, square blocks of horizontal 8 pixels × vertical 8 lines) for which DCT is performed. And is divided into unit blocks (for example, a square block of horizontal 16 pixels × vertical 16 lines composed of the four small blocks).

Inside the unit block, as shown in FIG.
Field (hatched area) and second field (white area)
, Are alternately folded (interlaced). Whether the DCT should be performed in frame units or field units means that two fields are interlaced with respect to four small blocks included in the unit block as shown in FIG. Is it more efficient to handle in frame units, or is it more efficient to handle in a non-interlaced state (in field units)?

Therefore, the frame / field determination circuit 3
In the above determination, the highest frequency power (Var1) in the vertical direction of the unit block and a frequency power (V
ar2).

When the data value of the pixel position of the i-th pixel on the right and the pixel position of the j-th line on the bottom is X _ij with reference to the pixel position at the upper left corner in the unit block, specifically, the maximum value in the vertical direction of the unit block is obtained. The frequency power is calculated by the following formula.

[0013]

[Equation 1]

The frequency power of 1/2 is calculated by the following equation.

[0015]

[Equation 2]

From these values, Var1 ≧ Var2 + 4
In the case of 096, DCT is performed in field units and Var
If 1 <Var2 + 4096, DCT in frame units
Is performed.

Further, the above judgment result is output to the encoding circuit 6, a code is assigned, and added to the encoded data as a part of the additional information.

Next, the operation of the above-described encoding circuit 6 will be described in more detail with reference to FIG. 6 in the case where the scanning order within a small block is the two types shown in FIG.

In FIG. 6, when the quantized data is input from the terminal 601, the code assignment circuit 602 is smaller than the scan order 1 according to the scan order table 603, and the code assignment circuit 604 is smaller than the scan order 2 according to the scan order table 605. The blocks are scanned and a variable length code is assigned to each. The comparison / determination circuit 606 compares the generated amounts of the two types of encoded data for each unit block, and outputs the encoded data having the smaller generated amount.

The code assignment circuit 607 receives the judgment output from the frame / field judgment circuit 3 input from the terminal 608, the motion vector from the motion vector detection circuit 14 input from the terminal 609, and the comparison judgment circuit 6 described above.
The additional information such as the determination output from 06 is received, the respective codes are assigned, and output as encoded data. The multiplexing circuit 610 multiplexes the coded data output from the comparison / determination circuit 606 and the coded data output from the code allocation circuit 607, and then outputs the multiplexed data from the terminal 611 to the buffer memory 7.

[0021]

In the above-described conventional method, it is determined whether orthogonal transformation is performed in frame units or field units by using the maximum vertical frequency power of the unit block and 1/2 of the maximum frequency power. Since it is performed by comparing with the frequency power, it is necessary to perform the operation several hundred times for each unit block. Further, it is necessary to add a flag indicating the determination result to the encoded data as a code.

An object of the present invention is to solve the above-mentioned conventional problems and to reduce the number of calculations for each processing unit block at the time of the determination, and also to use the flag indicating the determination result as a code. Another object of the present invention is to provide a moving picture coding system and apparatus provided with a frame / field determination means that does not need to be added to encoded data.

Further, in the above-mentioned conventional method, when a variable length code is assigned to the quantized data after the quantization processing, the scanning order within the small block is once encoded by several kinds of scanning methods, and then the code is encoded. It is necessary to perform a process that is not actually needed (not selected), such as selecting a method that generates less generated data. Further, it is necessary to add a flag indicating which method is selected as the scanning order within the small block to the encoded data as a code.

Therefore, an object of the present invention is to solve the above-mentioned conventional problems, and when assigning a variable-length code to the quantized data, the scanning order within a small block is optimal without performing unnecessary processing. A moving image provided with a scanning order determining means that can determine any of the methods, and does not need to add a flag indicating which method is selected as the scanning order in the small block to the encoded data as a code. An object is to provide an image coding system and apparatus.

[0025]

According to the present invention, a motion vector is used to compare the magnitude of the motion vector with an appropriate threshold value to determine whether orthogonal transformation is performed in frame units or in field units. A frame / field determination means for obtaining the above is provided.

Further, the present invention uses motion vectors to
By comparing its size with an appropriate threshold value, the scan order that can determine the optimum scan order within a small block when assigning a variable-length code to the quantized data after quantization processing Judgment means is provided.

[0027]

The frame / field determination means should receive the motion vector from the motion vector detection means, and use that value to compare it with a certain appropriate threshold value to perform the orthogonal transformation in frame units or in field units. Whether or not it is determined. The orthogonal transformation means is based on the determination output,
Orthogonal transformation is performed in frame units or field units. Since it is sufficient to make a determination using one set of motion vectors for each processing unit block, the amount of calculation is small,
The processing speed can be improved.

Furthermore, since the determination can be performed on the decoding side, it is not necessary to add a flag indicating the determination result to the encoded data, so that the amount of encoded data generated can be reduced.

Further, the scanning order determination means receives the motion vector from the motion vector detection means, and uses the value thereof,
By comparing with a certain suitable threshold value, the optimum method for the scanning order within the small block when allocating the variable length code to the quantized data after the quantization processing is determined. The encoding means scans a small block in the determined scanning order and allocates a variable length code to generate encoded data. Since the scanning order is uniquely determined, it is not necessary to perform encoding in other scanning orders, and the processing speed can be improved.

Furthermore, since the decoding side can determine the scanning order in which the scanning within the small blocks is performed, it is necessary to add a flag indicating the determination result to the encoded data. Since there is no problem, the amount of encoded data generated can be reduced.

[0031]

First, a first embodiment of the present invention will be described.

FIG. 1 is a block diagram showing a moving picture coding system and apparatus as a first embodiment of the present invention. The same parts as those in FIG. To do.

As shown in FIG. 1, in comparison with the conventional configuration shown in FIG. 5, the frame / field determination circuit 15 is a new component in this embodiment. The frame / field determination circuit 15 outputs the determination result as to whether the DCT should be performed in frame units or in field units.
This is similar to the field determination circuit 3, but differs in that it receives not the difference image data but the motion vector for the unit block from the motion vector detection circuit 14 as an input.

Considering a moving image signal, if the pattern is stationary or hardly moves due to the interlaced structure, it is more efficient to perform DCT in frame units, and there is a certain amount of motion. If DCT
It is conceivable that it is more efficient to carry out in units of fields.

Therefore, the frame / field determination circuit 1
In 5, the magnitude of the motion vector input to the same circuit is used. First, in the motion vector detection circuit 14, since the maximum search range of motion vectors is set in advance,
The value is (horizontal component, vertical component) = (MVX, MVY)
Then, the maximum value M of the magnitude (absolute value) of the motion vector
V is given by the following equation.

[0036]

[Equation 3]

By dividing this value into 20 steps, for example, 21 normal values (0 to 20) including 0 can be obtained. Then, when the absolute value of the motion vector actually input to the frame / field determination circuit 15 is, for example, the normal value 5 or less, the DC value is increased in units of frames.
T is performed, and if it is greater than the normal value 5, DCT is performed in field units. By such processing, it is possible to obtain a determination result that is substantially the same as that of the frame / field determination circuit 3.

Further, in the above-mentioned determination, since it is necessary to obtain only the magnitude of one unit block from a set of motion vectors (horizontal component, vertical component) corresponding thereto,
Compared with the above-mentioned conventional determination method, the amount of calculation is greatly reduced. Further, since the value of the motion vector is included in the encoded data as additional information, the determination can be easily performed on the decoding side. That is, it is not necessary to add the flag indicating the determination result, which is required to be added for each unit block in the conventional method, to the encoded data as a code.

Next, a second embodiment of the present invention will be described.

The overall configuration of this embodiment is basically the same as the configuration shown in FIG. 1 described above, and is represented by replacing the encoding circuit 6 with an encoding circuit 16 as a new component. Therefore, referring to FIG. 1 below,
The operation of the encoding circuit 16 will be described with reference to FIG. It should be noted that parts corresponding to those in FIG.

In FIG. 7, the comparison / determination circuit 612 receives the motion vector output from the motion vector detection circuit 14 from the terminal 609, and switches the switch 613 according to its magnitude (absolute value). Specifically, a certain threshold value T is set with the magnitude of the motion vector as mv, and m
When v ≦ T, the scan order table 603 is selected, and mv>
When it is T, a determination result for selecting the scan order table 604 is output. The switch 613 is switched according to the determination output, and the code is assigned in the code assignment circuit 614 according to the determined scanning order. However, in the above determination, since the motion vector is obtained for each unit block, all the small blocks included in the unit block are assigned codes in the same scanning order.

By performing the above determination, it is not necessary to assign a code in a scanning order that is not actually selected, so that the processing amount can be reduced as compared with the above-described conventional determination method. Further, as in the case of the above-described first embodiment, the value of the motion vector is included in the encoded data as additional information, so the determination can be easily performed on the decoding side. That is, it is not necessary to add the flag indicating the determination result to the encoded data as a code.

Next, a third embodiment of the present invention will be described.

The overall configuration of this embodiment is basically the same as the configuration shown in FIG. 1 described above, and the frame / field determination circuit 15 and the determination circuit 1 are added as new components.
It is represented by being replaced by 7. Therefore, this embodiment will be described below with reference to FIG.

The determination circuit 17 is the motion vector detection circuit 1
4 is the same as the above-described frame / field determination circuit 15 in that it receives a motion vector for the unit block from 4 and outputs a certain determination result, but the determination content is different.

For example, the determination circuit 17 receives the motion vector from the motion vector detection circuit 14, uses the magnitude (absolute value) of mv as a certain threshold value T, and mv≤
When T, the above-mentioned DCT in frame units is performed, and mv>
When T, as shown in FIG. 8, the small block is divided and a one-dimensional (8 pixels × 1 line) DCT in the horizontal direction is performed to output a determination result. The DCT circuit 4 can switch the DCT method for each unit block by performing DCT according to the determination output. However, in the above determination, since the motion vector is obtained for each unit block, DCT is performed by the same method for all the small blocks included in the unit block.

As in the case of the first and second embodiments described above, the value of the motion vector is included in the encoded data as additional information, so that the determination can be easily performed on the decoding side. That is, the flag indicating the determination result is
It does not need to be added to the encoded data as a code.

As described above, according to the present embodiment, the processing amount can be reduced by obtaining the determination result of switching the DCT using the motion vector for each unit block, and the determination result is added to the encoded data. Since there is no need, the coding efficiency can be improved.

In the three embodiments of the present invention described above,
The case in which there is always a motion vector for the unit block to be processed has been described. However, in the case of a mode in which there is no motion vector such that the unit block is coded in a frame or in a field, various methods can be used by conventional methods. Is determined.

In the second embodiment described above, the case where the scanning order in the small block is two types has been described. However, by setting two threshold values and comparing them with the absolute value of the motion vector, 3 It is possible to make a determination such that an appropriate scanning order is selected from the scanning orders of the types, and the scanning order is not limited to the above two types.

Further, in the above-described third embodiment, the DCT in the frame unit and the horizontal direction 1 are determined by the absolute value of the motion vector.
Although the case of switching between the three-dimensional DCT and the one-dimensional DCT has been described, for example, it is possible to perform not the one-dimensional DCT in the horizontal direction but the one-dimensional DCT in the vertical direction (1 pixel × 8 lines). The size of the small block to be performed is not limited to a specific size.

[0052]

As described above, according to the present invention, it is determined whether or not orthogonal transformation should be performed on a frame-by-frame basis or on a field-by-field basis using a motion vector, and the orthogonal transformation should be performed accordingly. As a result, the amount of calculation for the determination can be reduced, and the processing speed can be improved. Furthermore, since the determination can be performed on the decoding side as well, it is not necessary to add a flag indicating the determination result to the encoded data, and the amount of encoded data generated can be reduced.

Further, according to the present invention, the optimum method for the scanning order within the small block is determined using the motion vector, the scanning within the small block is performed according to the scanning order, and the variable length code is assigned. It is not necessary to perform encoding other than the scanning order, and the processing speed can be improved. Furthermore, since the determination can be performed on the decoding side, it is not necessary to add the flag indicating the determination result to the encoded data, and the amount of encoded data generated can be reduced.

In addition, according to the present invention, it is possible to reduce the processing amount and improve the coding efficiency by sequentially switching the sizes of the small blocks to which DCT is applied by using the motion vector.

[Brief description of drawings]

FIG. 1 is a block diagram showing a moving picture coding system and apparatus as a first embodiment of the present invention.

FIG. 2 is a diagram for explaining a division structure within one frame.

FIG. 3 DCT in frame units and DC in field units
It is a figure for demonstrating the difference of T.

FIG. 4 is a diagram showing an example of a scanning order within a small block.

FIG. 5 is a block diagram showing a conventional moving image coding system.

FIG. 6 is a block diagram showing an operation of a conventional encoding circuit.

FIG. 7 is a block diagram showing an operation of an encoding circuit as a second embodiment of the present invention.

FIG. 8 is a diagram for describing a horizontal one-dimensional DCT.

[Explanation of symbols]

3, 15 ... Frame / field determination circuit, 4 ... DCT
Circuits, 5 ... Quantization circuits, 6, 16 ... Encoding circuits, 7 ... Buffer memories, 12 ... Motion compensation circuits, 14 ... Motion vector detection circuits, 17 ... Judgment circuits, 602, 604, 60
7, 614 ... Code assignment circuit, 603, 605 ... Scan order table, 606, 612 ... Comparison determination circuit, 610 ... Multiplexing circuit, 613 ... Switch.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Masahiro Yasuoka 4, 6 Kanda Surugadai, Chiyoda-ku, Tokyo Inside Hitachi, Ltd.

Claims (5)

[Claims] 1. A motion vector of a moving image signal is detected for each unit block between frames or fields, and motion compensation is performed by shifting the unit block by the motion vector to generate reference image data. In the moving picture coding system, which is configured to obtain at least highly efficient coded data by performing at least orthogonal transformation on the difference data between the image data and the reference image data and assigning a variable length code, A moving image coding method, wherein various determinations in a process for obtaining encoded data are performed based on the magnitude of the motion vector. 2. A small block which is subjected to the orthogonal transformation in a unit block is subjected to the orthogonal transformation in an interlaced state (frame unit) or a non-interlaced state (field unit). The moving picture coding method according to claim 1, wherein the moving picture coding method is determined by the magnitude of the motion vector relating to. 3. The moving picture coding system according to claim 1, wherein the scanning order within the small block when the variable length code is assigned is determined by the magnitude of the motion vector relating to the small block. 4. The moving picture coding system according to claim 1, wherein the sizes of the small blocks to which the orthogonal transformation is applied are sequentially switched according to the magnitude of the motion vector relating to the small blocks. 5. A motion vector detecting means for detecting a motion amount for each unit block between frames or fields of a moving image signal, and motion compensation for generating reference image data by shifting the unit block by the motion vector. Means, orthogonal transformation means for performing an orthogonal transformation on the difference data between the original image data and the reference image data, a quantization means for quantizing the transform coefficient data after the orthogonal transformation, and a quantization after the quantization. A coding means for allocating a variable-length code to the data to generate high-efficiency coded data, wherein the high-efficiency coded data is once written in a buffer memory and is transmitted at a predetermined transmission rate. In a moving picture coding device, a coding process is performed according to the magnitude of a motion vector, using the method according to any one of claims 1 to 4. 2. A moving picture coding device, comprising: a judging means for making various judgments in 1.