JPH0292180A - Three-dimensional orthogonal transformation encoding system for moving image signal - Google Patents

Abstract

PURPOSE:To highly efficiently compress picture data by supplying a read signal line to the prescribed terminal of the memory of a memory device according to the read sequence of a moving image signal. CONSTITUTION:In a memory read control circuit 13, by the signal of a decision result sent from a deciding circuit 20 of the picture content change between screens, to which part of continuous screens in an N number on a time base the signal of the decision result corresponds is checked. Further, read control signals are supplied to prescribed terminals 3 to 10 in the memories of the memory device in the N number through a line 22 according to the pattern of the read sequence of the moving image signal from the memories of the memory device in the N number fixed beforehand corresponding to the part. Thus, the deterioration of a picture quality is reduced, the discontinuous motion is not generated in the moving image, and the picture data can be highly efficiently compressed.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a three-dimensional orthogonal transform encoding system for moving image signals.

(Prior art) As a high-efficiency encoding method that enables signals to be efficiently encoded with a smaller amount of information in recording, transmission equipment, and other various devices that perform signal processing of digital signals, human Each sample is It is well known that various high-efficiency encoding methods have been proposed in the past that reduce the amount of information per signal. A method of encoding the orthogonal transform coefficients obtained by performing the following has also been known as one of the high-efficiency encoding methods.

By the way, when a three-dimensional orthogonal transformation is applied to a moving image signal, the respective orders are changed depending on the level of definition of the original image subjected to the three-dimensional orthogonal transformation, the magnitude of the movement of the original image, etc. The state of the energy distribution of the orthogonal transform coefficients is different for each, but let's look at the energy distribution of the three-dimensional orthogonal transform coefficients obtained when three-dimensional orthogonal transform is applied to a normal video image and the corresponding video signal. , the low-order orthogonal transform coefficients exhibit large energy, but the energy of the high-order orthogonal transform coefficients in the time axis direction, that is, the absolute value of the high-order orthogonal transform coefficients in the time axis direction, is small. Therefore, by utilizing such characteristics and reducing the number of bits allocated to high-order orthogonal transform coefficients in the time axis direction, it is possible to efficiently compress the amount of information and encode a moving image.

(Problem to be Solved by the Invention) However, the screen of the moving image corresponding to the moving image signal subjected to the three-dimensional orthogonal transformation may be affected by, for example, a scene change, sudden melting, or Fade
When the image content is suddenly changed by a wide range (in, fade out) or other operations, it is natural to use the time obtained by applying three-dimensional orthogonal transformation to the normal moving image and the corresponding moving image signal. Unlike the energy distribution of orthogonal transformation coefficients in the axial direction, high-order orthogonal transformation coefficients also have a large energy distribution, so it is necessary to increase the number of bits allocated to high-order orthogonality and transformation coefficients. is required,
Therefore, a problem arises in that the compression efficiency decreases.

As mentioned above, various solutions have been proposed as solutions for the case where the image content changes suddenly and widely due to an operation such as a scene change. For example, inter-frame coding using inter-image correlation has been proposed. Regarding the method, when the image content suddenly changes drastically due to an operation such as a scene change, and the rebuffer memory, which generates a large amount of information in the interframe encoder, is about to overflow. A solution was proposed in which a stop signal is given to the interframe encoder to stop encoding, but depending on the timing of stopping encoding, the screen before stopping and the new screen may be the same. However, there are drawbacks such as unnatural overlapping on the screen and discontinuous movement of moving images. In other words, the dynamic range of the quantizer's output is limited to a small value when the image content changes rapidly and widely, and the limited dynamic range is gradually expanded over time. In predictive coding or transform coding methods for video signals that return to a predetermined dynamic range, the dynamic range of the quantizer output is limited to a small value, and then the dynamic range gradually expands over time. The problem is that the image deteriorates during the period from which the dynamic range is restored to the predetermined dynamic range.

(Means for Solving the Problem) The present invention divides image signals corresponding to consecutive screens on the time axis into image signals for each screen group consisting of a predetermined number of screens. A means for obtaining image signals corresponding to individual screen groups, and a state of change in image content between adjacent screens on the time axis in one group of screens with respect to the image signals for each individual screen group described above. a means for detecting a state of change in the image content, a means for determining whether or not the state of change in the image content is a wide range; The screen that is earlier on the time axis among the two screens to be used is the first screen, and the other screen among the two screens is the second screen.
When the screen of means for generating an image signal corresponding to any state in which the entire screen is replaced with a second screen; means for three-dimensional orthogonal transformation of the image signal; A three-dimensional orthogonal transform encoding method for a moving image signal comprising a means for encoding three-dimensional orthogonal transform coefficients, and a means for encoding a three-dimensional orthogonal transform coefficient, and a means for encoding a predetermined number of screens of continuous screens on the time axis and corresponding image signals. Means for obtaining image signals corresponding to each individual screen group by dividing the image signals for each screen group consisting of;
means for detecting a state of change in image content between screens adjacent on the time axis in a group of screens; means for determining whether the above-mentioned state of change in image content is wide; The screen that is earlier on the time axis in the two consecutive screens on the time axis that gave rise to the judgment result that the state of change is large is the first screen, and the screen in the two screens mentioned above is The other screen is the second screen, the number of screens from the beginning of the first group of screens to the first screen is A, and the number of screens from the second screen to the last screen of the first group of screens is A. B, and furthermore, the state in which the second screen and all subsequent screens in the first group of screens are replaced with the first screen is the first screen replacement state, and the state in which the first screen of the first group of screens is replaced with the first screen is the first screen replacement state. When the state in which all screens up to the first screen are replaced with the second screen is defined as the second screen replacement state, the number of screens A from the beginning of the first group of screens described above to the first screen described above, and the The relationship with the number of screens B from screen 2 to the last screen of group 1 is A
In the case of ≦B, the A
> In the case of B, means for generating an image signal in a first screen replacement state, means for three-dimensional orthogonal transformation of the above-mentioned image signal, and three-dimensional orthogonal transformation obtained by the three-dimensional orthogonal transformation. A three-dimensional orthogonal transform encoding method for a moving image signal comprising means for encoding coefficients, and an image signal corresponding to consecutive screens on the time axis for each screen group consisting of a predetermined number of screens. Means for obtaining image signals corresponding to individual screen groups by dividing the image signals into image signals corresponding to each individual screen group; means for detecting the state of change in image content between the two; means for determining whether the state of change in image content is large [1]; and means for determining whether the state of change in image content is large [1]; Of the two consecutive screens on the time axis that caused the determination result, the screen that is earlier on the time axis is the first screen, the other screen of the two screens is the second screen, and , the number of screens from the beginning of the first group of screens to the above-mentioned first screen is A, the number of screens from the second screen to the last screen of the first group of screens is B, and furthermore, in the first group of screens, The state where the second screen and all subsequent screens are replaced with the first screen is called the first screen replacement state, and the entire screen from the first screen of the first group of screens to the first screen is called the second screen. When the screen replacement state 7 is set as the second screen replacement state, the number of screens A from the beginning of the first group of screens described above to the first screen described above, and the number of screens of the first group from the second screen. The relationship with the number of screens B up to the last screen is
In the case of A<B, the second screen replacement state is established.
In the case of A≧B, means for generating an image signal in a first screen replacement state;
Means for dimensional orthogonal transformation and 3 obtained by 3-dimensional orthogonal transformation
The present invention provides a three-dimensional orthogonal transform encoding method for a moving image signal, which comprises means for encoding three-dimensional orthogonal transform coefficients.

(Example) Hereinafter, specific contents of the three-dimensional orthogonal transform encoding method for moving image signals of the present invention will be explained in detail with reference to the accompanying drawings.

FIG. 1 is a block diagram of an embodiment of the three-dimensional orthogonal transform encoding method for moving image signals of the present invention. In FIG. This is an input terminal for a digital signal (hereinafter simply referred to as a moving image signal), and the moving image signal supplied to the input terminal 1 is connected to the movable contact C of the changeover switch SWI and the change in image content between screens. The signal is supplied to a digital subtracter 14 provided in the determination circuit 20.

The changeover switch SW1 is a changeover switch that performs a switching operation in conjunction with the changeover switch SW2 and the changeover switch SW3. The three changeover switches SW1 to SW3 described above
The mode of interlocking switching operation is as follows for each changeover switch SWI~S
Each movable contact C in W3 is connected to each changeover switch SWI~
In SW3, the fixed contacts having the same reference numerals in the drawings are switched simultaneously.

Further, the timing at which the movable contact C of each of the changeover switches SWI to SW3 described above is sequentially and alternately switched between the two fixed contacts a and b is as follows.
The period of the switching control signal is determined by the switching control signal of a predetermined period supplied from the control signal generation circuit 12 to the switching control signal supply terminals 24 to 26 in the storage device MA, which will be described later. The length of the moving image signal that should be stored in all of the multiple memories Mal, Mg2... (The length of the moving image signal that should be stored sequentially in all of the multiple memories Mbl, Mb2... (same length as the video signal length).

Therefore, the number of memories M a 1 , M a 2 . . . used to configure the storage device MA described above, and the memories Mbl, Mb2 . . . used to configure the storage device MB are as follows. and the number of N pieces (the first
In the illustrated embodiment, N=4), and each of the above-mentioned memories Mal, Mg2-1Mbl, Mb2...
When the length of the moving image signal to be sequentially stored in each one of the above is an image signal corresponding to each one screen in the moving image (image signal of one frame period) In this case, the cycle of the switching control signal described above is N frame periods (in the embodiment shown in the first figure, it is four frame periods).

Note that each memory Ma in the storage devices MA and MB
1, Ma2..., Mbl, Mb2..., the length of the moving image signal is the length of the image signal corresponding to the field period in the moving image signal (image signal of one field period). ), and it goes without saying that the cycle of the switching control signal in this case will be the N field period.

In the following description of the embodiment shown in FIG. 1, it is assumed that the period of the switching control signal described above is a period of four frames. Therefore, the moving image signals that are sequentially supplied to the input terminal 1 on the time axis are controlled by the changeover switch SW.
4. The movable contact C of I is switched to the fixed contact to the a side.
During the frame period, each memory Ma in the storage device MA
The video signals of four consecutive frame periods on the time axis are supplied to the storage device MA so that the video signals of one frame period can be sequentially stored in Ma2, . During the 4-frame period when the movable contact C of the changeover switch SW1 is switched to the fixed contact on the b side, the video signal of the 1-frame period is sequentially stored in each memory Mbl, Mb2, etc. in the storage device MB. A moving picture signal of four consecutive frame periods on the time axis is supplied to the storage device MB so that it can be stored.

As described above, the two storage devices MA are supplied with video signals of four frame periods sequentially and alternately on the time axis.

Each memory in MB Mal-Ma4, Mbl ~ Mb
4..., from the control signal generation circuit 12 and the memory read control circuit 13 to the respective terminals 3 to 10,
A write control signal and a read control signal are supplied at predetermined timings, and during a period when a moving image signal of 4 frame period is supplied to the storage device MA, each memory Mal in the storage device MA ~Ma4 sequentially stores one frame period of video signals, and during the period when four frame periods of video signals are supplied to the storage device MB, each memory in the storage device MB A moving picture signal for one frame period is sequentially stored in Mb1 to Mb4, and each of the memories M a 1 to Ma4 in the two storage devices MA and MB described above is
The read operation from Mbl to Mb4, . . . is performed by read control signals sent from the memory read control circuit 13 at predetermined timings, as will be described later.

In FIG. 1, reference numeral 11 denotes a three-dimensional orthogonal transform circuit, and in this three-dimensional orthogonal transform circuit 11, the aforementioned changeover switch S
Via the movable contact C of W2, the moving image signal supplied from the storage device MA or the storage device MB described above is subjected to three-dimensional orthogonal transformation to generate three-dimensional orthogonal transformation coefficients.
The dimensional orthogonal transform coefficients are encoded by the encoding circuit 23 and sent to the output terminal 2.
The dimensional orthogonal transform circuit 11 and the encoding circuit 23 each perform the above-described predetermined operations, that is, the three-dimensional orthogonal transform coefficient generation operation and predetermined encoding, in response to required control signals supplied from the control signal generation circuit 12, respectively. It takes action. Furthermore, any configuration may be used as the three-dimensional orthogonal transform circuit 11 and the code converter circuit 23 described above.

The component indicated by a dashed-dotted line frame 20 in FIG. 1 is a circuit 20 for determining changes in image content between screens. Regarding the image signals for each screen group consisting of a predetermined number of screens N (N=4 in the embodiment shown in the first figure), the image signals corresponding to the screens shown in FIG. The system is configured as follows: a function to detect/detect the state of change in image content between adjacent screens; and a function to determine whether the state of change in image content is wide. Then, the judgment result by the above-described judgment function is supplied to the memory read control circuit 13.

The image content change determination circuit 20 between screens shown in FIG. 1 includes a digital subtracter 14, an absolute value circuit 15,
It is composed of an accumulator 16, a comparator 17, and a threshold generation circuit 18, and is digital; the subtracter 14 calculates and outputs the difference between image signals of adjacent screens on the time axis in one group of screens. .

Even if the calculation of the difference between the image signals of the screens adjacent on the time axis in one group of screens performed by the subtractor 14 described above is performed on all the image signals of the screens adjacent on the time axis, Alternatively, the process may be performed for corresponding discrete image signals of image signals of screens that are adjacent on the time axis.

The output signal from the digital subtracter 14 is converted into the absolute value of the output signal from the digital subtracter 14 in the absolute value circuit 15 and is applied to the accumulator 16. Adjacent on the time axis that appear sequentially within the period of the length of the video signal corresponding to the screen (in the case of the explanation of the embodiment, one frame period corresponding to one complete screen) The sum of the absolute values of the difference signals of the image signals on the screen is calculated and supplied to the comparator 17. The accumulator 16 stores a period of the length of the video signal corresponding to one complete screen (
In the case of the explanation of the embodiment, the calculation result of the sum of the absolute values of the difference signals of the image signals of adjacent screens on the time axis that appear sequentially within one frame period (corresponding to one complete screen) is supplied to the comparator 17, and then reset by a reset signal supplied from the control signal generation circuit 12 to the terminal 19.

The comparator 17 calculates the sum of the absolute values of the difference signals of the image signals of adjacent screens on the time axis supplied from the accumulator 16 as described above, and the dark value generation circuit (threshold value setting circuit) 18
A signal indicating the comparison result is supplied to the memory read control circuit 13.

The threshold value supplied from the threshold generation circuit 18 to the comparator 17 is determined when there is a large change in image content between adjacent screens on the time axis, that is, when there is a large change in image content between adjacent screens on the time axis, that is, when there is a large change in image content between adjacent screens on the time axis, The above-mentioned situation occurs when the image screen changes suddenly and widely, for example, due to a scene change, rapid brightening, fading, fading, or other operations. When the calculation result of the sum of the absolute values of the difference signals of the two image signals is supplied to the comparator 17, the comparison 17 can output a determination result that the state of change in the image content is large. The threshold value is set in advance.

By the way, the three-dimensional orthogonal transform operation performed in the three-dimensional orthogonal transform circuit divides the image signals corresponding to consecutive screens on the time axis into image signals for each screen group consisting of a predetermined number of screens. It is carried out in relation to
In the three-dimensional orthogonal transform encoding method of the moving image signal shown in FIG. A predetermined number N of screens supplied from MA or storage device MB (in the example shown in the first figure, N=
4) is carried out using image signals corresponding to one group of screens.

Therefore, in a moving image screen corresponding to a moving image signal corresponding to a group of screens divided into image signals for each screen group consisting of a predetermined number N of screens as described above, for example, When the image content suddenly changes drastically due to a scene change, rapid brightening or fading (fade in, fade out), or other operations,
What is the energy distribution of orthogonal transformation coefficients in the time axis direction obtained by three-dimensional orthogonal transformation operation?What is the energy distribution of orthogonal transformation coefficients in the time axis direction obtained by three-dimensional orthogonal transformation of a normal video and the corresponding video signal? It is understandable that things will be drastically different.

Figure 2 (a) and Figure 3 (a) are each image signal obtained by dividing a continuous image signal on the time axis into image signals for each screen group consisting of a predetermined number N of screens. For example, a scene change or sudden brightening or darkening (fade in, fade out) in any of the video signals corresponding to the first group of screens and the corresponding video image screen. FIG. 7 is a diagram illustrating a case where the image content suddenly changes widely due to other operations.

In Figures 2 to 4, N is the number of screens in a group of images that correspond to image signals obtained by dividing continuous image signals on the time axis, and α is the number of screens in a group of images obtained by dividing continuous image signals on the time axis. This is a symbol to indicate that it is a moving image signal in
Furthermore, β is a symbol to indicate that the video signal is in a different scene from the scene α described above, and the subscript added to the symbols α and β described above is the number of the frame in the video signal in each scene. It is.

(,) in FIG. 2 and (,) in FIG. 3 are moving image signals α1 to αm of a certain scene, and moving image signals β1 . β
2. When β3... exists continuously on the time axis, the moving image signals of the two scenes are the time axis that is subject to three-dimensional orthogonal transformation in the three-dimensional orthogonal transformation circuit. If it is included in the image signal of a group of screens consisting of N consecutive screens, that is, a group divided into image signals for each group of screens consisting of a predetermined number of N screens. During the video screen corresponding to the video signal corresponding to the screen of Figure 2 shows an example of a case where a moving image signal whose contents are rapidly changing over a wide range is subjected to three-dimensional orthogonal transformation in a three-dimensional orthogonal transformation circuit. As illustrated in (a), the moving image signals of two scenes are converted into a group of N consecutive screens on the time axis that are subjected to three-dimensional orthogonal transformation in a three-dimensional orthogonal transformation circuit. If it is included in the image signal of the screen group, the energy distribution of the orthogonal transformation coefficient in the time axis direction obtained by the three-dimensional orthogonal transformation operation is the same as the one obtained by three-dimensional orthogonal transformation of the normal moving image and the corresponding moving image signal. The energy distribution of orthogonal transformation coefficients in the time axis direction obtained by
As a result, high-efficiency encoding cannot be achieved, or image quality may deteriorate.

(b), (C) in Figure 2 and (b), (Q) in Figure 3 are explained with reference to (a) in Figure 2 and (a) in Figure 3 above. FIG. 4 is a diagram for explaining the configuration principle and operation principle of the three-dimensional orthogonal transform encoding method for moving image signals of the present invention, which can solve the above problems;
This figure is a diagram used to explain the operation of the embodiment shown in FIG. 1.

First, (b) and (c) in Fig. 2 are
), the moving image signal α of two scenes is
When 3 to αm and β1 to β3 are included in the image signal of a group of screens consisting of N screens that are continuous on the time axis that is the target of three-dimensional orthogonal transformation in a three-dimensional orthogonal transformation circuit, , 3D orthogonal transformation 1, unify the image signals of a group of screens consisting of N consecutive screens on the time axis, which are the targets of 3D orthogonal transformation, into the image signal of one of the scenes. It was designed to let you do so.

FIG. 2(b) shows two scene moving image signals α3 to αm, β1 as illustrated in FIG. 2(,) above.
~ If β3 is included in the image signal of a group of screens consisting of N consecutive screens on the time axis that is the target of three-dimensional orthogonal transformation in a three-dimensional orthogonal transformation circuit, The moving image signals β1 to β3 are the moving image signal αm of the last frame of the moving image signal of the scene before the scene is changed.
The image signals of a group of N screens that are continuous on the time axis that are subject to three-dimensional orthogonal transformation are converted into moving image signals α3. α■ζm, αm.

This is an example where αm and αm are unified to one value, and (c) in Figure 2 is a moving image of two scenes as illustrated in (,) in Figure 2 above. Signal α3~
When αm, β1 to β3 are included in the image signal of a group of screens consisting of N consecutive screens on the time axis that are subject to three-dimensional orthogonal transformation in a three-dimensional orthogonal transformation circuit,
The moving image signals α3 to αm of one of the scenes are replaced by the moving image signal β1 of the first frame of the moving image signal of the scene after the scene conversion, and are subjected to three-dimensional orthogonal transformation 1, β1, . β2. This is an example of a case where one is unified like β3.

Next, (b) and (c) in FIG. 3 are shown in (a) in FIG.
), the moving image signal α of two scenes is
3 to αm, +1 to β1 to β is included in the image/image signal of a group of screens consisting of N screens that are continuous on the time axis, which is the target of three-dimensional orthogonal transformation in the three-dimensional orthogonal transformation circuit. In this case, a 3D orthogonal transformation circuit transforms the image signals of a group of N screens that are continuous on the time axis into the image signals of one of the scenes. In the case of unification, the screen αm that is earlier on the time axis of two consecutive screens on the time axis that caused the screen change is set as the first screen, and the screen αm that is earlier on the time axis is the first screen, and Let the screen β1 be the second screen, and let the number of screens from the beginning of the first group of screens to the above-mentioned first screen αm be A, and the screens from the second screen β1 to the last screen of the first group of screens. The number is B, and the state in which the second screen and all subsequent screens in the first group of screens are replaced with the first screen αm is defined as the first screen replacement state, and the first screen of the first group of screens is The entire screen from the first screen αm to the second screen
When the state where screen β1 is replaced is set as the second screen replacement state, the first screen from the top of the first group of screens described above is
If the relationship between the number A of screens up to the screen A and the number B of screens from the second screen to the last screen of the first group of screens is A≦B, the second screen is In addition, if A>B, the screen replacement state is changed to the first screen replacement state as shown in FIG. 3(b), or the The relationship between the number of screens A up to the first screen and the number B of screens from the second screen to the last screen of the first group of screens is A (in the case of B, as shown in Figure 3 (c)). When A≧B, the screen is replaced in the second screen replacement state, and in the case of A≧B, the screen is replaced in the first screen replacement state as shown in FIG. 3(b). That is, if the number of screens to be replaced is different for the image signals of two scenes, the image signal of the scene with the smaller number of replaced screens is selected and replaced.
If the number of screens to be replaced is the same for two scenes, the image signal of an arbitrary scene is selected for replacement, thereby replacing the image signal as described above. This is to prevent large unnaturalness from appearing in the motion of the moving image even if the motion is performed.

The image signal replacement as explained with reference to (bL (c) in FIG. 2 and (b) and (Q) in FIG. In the encoding method, the memories Mal in the storage devices MA and MB,
This can be done by setting the timing of reading from M a 2-1 Mbl, Mb 2-.

In (a), (b), and (c) of FIG. 4, the square frames written above each indicate the storage devices MA,
Memory Mal in MB, M a 2-1 Mbl.

This is a diagram for explaining the state of writing an image signal to Mb2..., and also includes (a), (b), and (b) in FIG.
In (c), the square frames below each indicate the memories Ma in the storage devices MA and MB.
1, Ma2 . . . , Mbl, Mb2 . . . FIG.

The square frame shown in Figure 4 is the storage device M in Figure 1.
It is said to represent either A or MB.

Also, in FIG. 4, #1. #2. #3. β
4 are each memory Mal, Ma2.4 in the storage device MA (MB) shown in FIG. Ma3. M
Indicates the memories with the same numbers as the subscripts 1, 2, 3.4 in a4 (Mbl, Mb2.Mb3.Mb4), and are written in the squares indicating each memory. Letters and numbers indicate signals in sequential frame periods of a moving image signal that is continuous on the time axis.

In the two storage devices MA and MB shown in FIG. 1, the moving image signals that are sequentially supplied to the input terminals 1 on the time axis are generated when the movable contact C of the changeover switch SWI is switched to the fixed contact A side. During the 4th frame period, the storage device MA
The moving image signals for one frame period are sequentially stored in each memory Mal, Ma2, etc., and the switching switch +
During the 4-frame period when the movable contact C of SW1 is switched to the fixed contact on the b side, the moving image signals of the 1-frame period are sequentially stored in each memory Mbl, Mb2, etc. in the storage device MB. The moving image signals of four frame periods are sequentially and alternately stored on the time axis in two storage devices M.
Each memory Mal to Ma4, Mbl to M in A, MB
It is stored in b4...

As described above, each memory M in the storage device MA (MB)
The moving image signal stored in al-Ma4 (Mbl'' Mb4) is stored in N memories (4 memories in the embodiment shown in Figure 1) in the square frame above it in (,) in Figure 4. αm → β1 → β2 → for #1 → #2 → #3 → #4
It is exemplified as something stored like β3,
In addition, in (b) of FIG. 4, N memories (four memories in the embodiment shown in FIG. 1) in the square frame above it are shown.
αm → α set +1 → β1 for #l → #2 → #3 → #4
→It is exemplified as stored as β2,
Furthermore, in (C) of FIG. 4, αm for N memories (four memories in the embodiment shown in FIG. 1) #1 → #2 → #3 → #4 in the square frame above →αm+1→α
It is illustrated as being stored as m+2→β1.

Two storage devices MA are shown in FIG.

As mentioned above, when one storage device operates in the storage mode, the other storage device operates in the readout mode. The mode of memory is shown in Figure 4 (,
) to (c) in the storage device.
The manner in which image signals are read from the storage device in the readout mode when moving image signals of two scenes are stored in the memory is shown below in each of FIGS. 4(a) to (C). As shown in the square frames, the read operations are controlled by read control signals sent from the memory read control circuit 13 at predetermined timings, so that different operations are performed.

That is, the moving image signals stored in each of the memories Mal to Ma4 (Mbl-Mb4) in the storage device MA (MB) 7 as described above are N in the square frame at the top of (a) in FIG. memories (four memories in the embodiment shown in the first diagram) #1 → #2 → #3 → #4 αm → β1 → β2
→ In the case of the memory contents like β3, the period of one frame from the time when the moving image signal β1 begins to be input to the input terminal 1 is indicated by the dot-dash chain line frame 20 in FIG. The image content change determination circuit 20 performs an operation to determine the state of change in image content between the moving image signal αm and the moving image signal β1, thereby determining the change in image content between the two adjacent screens described above. A signal indicating that the state of change is wide is supplied to the memory read control circuit 13.

The memory read control circuit 13 uses the determination result signal sent from the determination circuit 20 for =32 image content change between screens as described above to determine whether the determination result signal is continuous on the time axis. By checking which part of the N screens corresponds to the corresponding part, a readout control signal is stored via the line 22 in accordance with a predetermined pattern of the readout order of moving image signals from the N memories of the storage device. The terminals (3 to 10) of a predetermined one of the N memories of the device are supplied.

Reading from the memory of the storage device illustrated in FIG. 4(a) is performed from N memories (in the embodiment shown in FIG. memory) in the order of #2 → #2 → #3 → #4, and the three-dimensional orthogonal transformation circuit 11 is supplied with the moving image signal β1 → β1 → β2 → β3 in series on the time axis. be done.

Further, as described above, the moving image signals stored in each of the memories Mal to Ma4 (Mbl - Mb4) in the storage device MA (MB) are stored in the N memories in the upper square frame of (b) in FIG. (Four memories in the embodiment shown in the first diagram)
αm → αm÷1 → β1 like #1 → #2 → #3 → #4
→ In the case of the stored content as β2, the screen shown by the dashed-dotted line frame 20 in FIG. The change in image content between the two adjacent screens is determined by the operation for determining the state of change in image content between the moving image signal αm+1 and the moving image signal β1, which is performed by the judgment circuit 20 for determining a change in image content between the two adjacent screens. A signal indicating that the state is wide is supplied to the memory read control circuit 13.

The memory read control circuit 13 uses the judgment result signals sent from the inter-screen image content change judgment circuit 20 as described above, so that the seven judgment result signals are continuous on the time axis. Look at which part of the N screens is located, and set the predetermined storage device N accordingly.
A readout control signal is supplied via line 22 to terminals (3 to 10) of predetermined ones of the N memories of the storage device according to the pattern of readout order of moving picture signals from the N memories.

Reading from the memory of the storage device illustrated in FIG. 4(b) is performed from N memories (four in the embodiment shown in FIG. 1) in the lower rectangular frame of FIG. 4(b). Memory) is performed in the order of #1 → #2 → #2 → #2, and the three-dimensional orthogonal transformation circuit 11 is supplied with the moving image signal αm → αm+1 → αm+1 → αm+1 in series on the time axis. be done.

Furthermore, as described above, the moving image signals stored in each of the memories Mal to Ma4 (Mbl - Mb4) in the storage device M A (M B ) are divided into N pieces in the upper rectangular frame of (c) in FIG. memories (four memories in the embodiment shown in the first diagram) αm → αm like #1 → #2 → #3 → #4
In the case of the memory contents such as +1 → αm+2 → β1, the signal shown by the dashed-dotted line frame 20 in FIG. Circuit 2 for determining image content changes between screens
The operation of determining the state of change in image content between the moving image signal αm+2 and the moving image signal βl performed in step 0 determines that the state of change in image content between the two adjacent screens described above is wide. A signal of the determination result is supplied to the memory read control circuit 13.

The memory read control circuit 13 uses the judgment result signal sent from the inter-screen image content change judgment circuit 20 as described above to detect N images that are continuous on the time axis. A read control signal is sent to the storage device via a line 22 in accordance with a predetermined readout order pattern of moving image signals from N memories of the storage device corresponding to the portion of the screen of the storage device. is supplied to terminals (3 to 10) of predetermined ones in the N memories.

Reading from the memory of the storage device illustrated in FIG. 4(c) is performed from N memories (four in the embodiment shown in FIG. 1) in the lower rectangular frame of FIG. 4(c). memory) in the order of #1 → #2 → #3 → #3, and the three-dimensional orthogonal transformation circuit 11 is supplied with the moving image signal αm → αm+1 → αm+2 → αm+2 in series on the time axis. be done.

As mentioned above, if it is known in which part of the N consecutive screens on the time axis the judgment result signal is located, screens like the ones shown in (b) and (c) in Fig. 3 mentioned above can be obtained. Even when the screen replacement is performed, the pattern of screen replacement is predetermined, so that the signal indicating that the state of change in image content is large between two adjacent screens is transmitted between the screens. When this occurs in the image content change determination circuit 20,
The circuit 20 for determining changes in image content between screens supplies the above-mentioned signal to the memory read control circuit 13 as an address signal indicating which part of the N screens that are continuous on the time axis, and performs memory read control. The circuit 13 may be configured to use the address signal described above to output a read control signal to be applied from the table to the memory targeted for read in the storage device.

(Effects of the Invention) As is clear from the detailed explanation above, the three-dimensional orthogonal transform encoding method for moving image signals of the present invention predetermines image signals corresponding to continuous screens on the time axis. means for obtaining image signals corresponding to each individual screen group by dividing the image signals into each screen group consisting of a number of screens, and dividing the image signals for each screen group into one group. means for detecting the state of change in image content between screens adjacent on the time axis in the screen; means for determining whether the state of change in the image content is large; The screen that is earlier on the time axis in two consecutive screens on the time axis that gave rise to the judgment result that the state of change is wide is defined as the first screen, and When the other screen is set as the second screen, the second screen and all subsequent screens in the first group of screens are replaced with the first screen, and the above from the first screen of the first group of screens. means for generating an image signal corresponding to a state in which the entire screen up to the first screen is replaced with a second screen; and means for three-dimensional orthogonal transformation of the image signal; A three-dimensional orthogonal transform encoding method for a moving image signal consisting of means for encoding three-dimensional orthogonal transform coefficients obtained by dimensional orthogonal transform, and image signals corresponding to continuous screens on the time axis are predetermined. means for obtaining image signals corresponding to each individual screen group by dividing the image signals into each screen group consisting of a number of screens, and dividing the image signals for each screen group into one group. means for detecting a state of change in image content between screens adjacent on a time axis in a screen of , means for determining whether the state of change in image content is wide or not, and a change in image content. The screen that is earlier on the time axis in two consecutive screens on the time axis that gave rise to the judgment result that the state is wide is the first screen, and the other screen in the two screens The screen is the second screen, and the number of screens from the beginning of the first group of screens to the first screen is A, and the number of screens from the second screen is 1.
Let B be the number of screens up to the last screen in the group, and further,
The state in which the second screen and all subsequent screens in a group of screens are replaced with the first screen is called the first screen replacement state, 1.
When the second screen replacement state is a state in which all the screens from the first screen of the group of screens to the first screen are replaced with the second screen, If the relationship between the number of screens A to the first screen and the number B of screens from the second screen to the last screen of the first group of screens is A≦B, the second screen replacement state is established. , means for generating an image signal in a first screen replacement state when A>B, means for performing three-dimensional orthogonal transformation on the image signal, and three-dimensional orthogonal transformation obtained by the three-dimensional orthogonal transformation. A three-dimensional orthogonal transform encoding method for a moving image signal comprising means for encoding orthogonal transform coefficients, and a screen group comprising a predetermined number of screens of continuous screens on the time axis and corresponding image signals. A means for obtaining image signals corresponding to individual screen groups by dividing the image signals into each image signal, and a means for obtaining image signals corresponding to each individual screen group, and a means for obtaining image signals corresponding to each individual screen group. a means for detecting a state of change in image content between matching screens; a means for determining whether the state of change in image content is wide-width; and a means for determining whether the state of change in image content is wide-width; Of the two consecutive screens on the time axis that caused the judgment result, the screen that is earlier on the time axis is the first screen, and the other screen of the two screens is the second screen. , and the number of screens from the beginning of the first group of screens to the first screen described above is A, and the number of screens from the second screen is 1.
Let B be the number of screens up to the last screen in the group, and further,
The state in which the second screen and all subsequent screens in a group of screens are replaced with the first screen is called the first screen replacement state, 1.
When the second screen replacement state is a state in which all the screens from the first screen of the group of screens to the first screen are replaced with the second screen, If the relationship between the number of screens A to the first screen and the number B of screens from the second screen to the last screen of the first group of screens is A<B, the second screen replacement state is established. , a means for generating an image signal so as to be in a first screen replacement state when A≧B, a means for three-dimensional orthogonal transformation of the image signal, and a three-dimensional image signal obtained by the three-dimensional orthogonal transformation. Since this is a three-dimensional orthogonal transform encoding method for a moving image signal, which comprises a means for encoding orthogonal transform coefficients, the three-dimensional orthogonal transform encoding method for a moving image signal of the present invention has little deterioration in image quality.
Moreover, it does not cause discontinuous movement in moving images,
Since image data can be compressed with high efficiency, the present invention can satisfactorily solve the various drawbacks of the conventional methods described above.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of an embodiment of the three-dimensional orthogonal transformation and encoding method for moving image signals of the present invention, and FIGS. 2 to 4 are signal arrangement diagrams for explanation. SW1~SW3...Choice switch, MA, MB...
Storage device, Ma 1. Ma2. Mb1. Mb2-Memory, 1... Targeted for three-dimensional orthogonal transform encoding.
Input terminal for digital signal of moving image, 2... Output child, 3 to 10... Terminal, 11... Three-dimensional orthogonal transformation circuit, 12... Control signal generation circuit, 13... Memory Read control circuit, 14... Digital subtracter, 15... Absolute value port 2 path, 16... Accumulator, 17... Comparator, 1
8... Threshold generation circuit, 19... Terminal, 20... Image content change judgment circuit between screens, 23... Encoding circuit, 24-26 switching, switching control signal for changeover switches SWI-SW3. Supply terminal, patent applicant: Victor Japan Co., Ltd.

Claims (1)

[Claims] 1. By dividing the image signals corresponding to consecutive screens on the time axis into image signals for each screen group consisting of a predetermined number of screens, each screen can be divided into individual screen groups. means for obtaining corresponding image signals; and means for detecting the state of change in image content between adjacent screens on the time axis in one group of screens with respect to the image signals for each individual screen group. , a means for determining whether or not the state of change in image content is wide-width, and a means for determining whether the state of change in image content is wide-width; When the screen that is earlier on the time axis is the first screen, and the other of the two screens is the second screen, the second screen and the subsequent screen in a group of screens All screens are replaced with the first screen, and all screens from the first screen of the first group to the first screen are replaced with the second screen.
means for generating an image signal corresponding to any of the states replaced by the screen; means for three-dimensional orthogonal transformation of the image signal; Three-dimensional orthogonal transform encoding method 2 for moving image signals consisting of means for encoding coefficients, in which image signals corresponding to consecutive screens on the time axis are encoded for each screen group consisting of a predetermined number of screens. Means for obtaining image signals corresponding to individual screen groups by dividing the image signals into image signals corresponding to each individual screen group; means for detecting the state of change in image content between the two; means for determining whether the state of change in image content is wide-width; and a determination result that the state of change in image content is wide-width. Of the two consecutive screens on the time axis that caused this, the screen that is earlier on the time axis is the first screen, the other screen of the two screens is the second screen, and 1 Let A be the number of screens from the beginning of the group of screens to the first screen mentioned above, B be the number of screens from the second screen to the last screen of the first group of screens, and furthermore, The first screen replacement state is a state where the screen and all subsequent screens are replaced with the first screen, and the entire screen from the first screen of a group of screens to the first screen is the second screen. When the replaced state is the second screen replacement state, the number of screens A from the beginning of the first group of screens described above to the first screen described above and from the second screen to the last screen of the first group of screens. The image signal is generated in such a way that the relationship with the number of screens B is such that when A≦B, the second screen replacement state is established, and when A>B, the first screen replacement state is established. A three-dimensional orthogonal transform encoding method 3 for a moving image signal, comprising: a means for performing three-dimensional orthogonal transform on the above-mentioned image signal; and a means for encoding three-dimensional orthogonal transform coefficients obtained by the three-dimensional orthogonal transform; By dividing the image signals corresponding to continuous screens on the time axis into image signals for each screen group consisting of a predetermined number of screens, image signals corresponding to each individual screen group are obtained. means for detecting the state of change in image content between adjacent screens on the time axis in one group of screens with respect to the image signal for each individual screen group; A means for determining whether the state is wide-width, and a means for determining whether the state of the change in image content is wide-width, and a means for determining whether the state of the change in image content is ahead on the time axis in two consecutive screens on the time axis that have produced the determination result that the state is wide. the other screen as the first screen,
The other of the above two screens is the second screen, the number of screens from the beginning of the first group of screens to the first screen is A, and the number of screens from the second screen to the last of the first group of screens is A. The number of screens up to the screen is B, and the state in which the second screen and all subsequent screens in the first group of screens are replaced with the first screen is the first screen replacement state, and the beginning of the first screen in the first group When the second screen replacement state is a state in which all screens from the screen to the first screen are replaced with the second screen, from the beginning of the first group of screens to the first screen. The relationship between the number of screens A and the number B of screens from the second screen to the last screen of the first group of screens is such that when A<B, the second screen replacement state is established. In this case, a means for generating an image signal in a first screen replacement state, a means for three-dimensional orthogonal transformation of the image signal, and a code for three-dimensional orthogonal transformation coefficients obtained by the three-dimensional orthogonal transformation. A three-dimensional orthogonal transform encoding method for moving image signals comprising means for converting images into three-dimensional images. As a means for determining whether the state of change is large or not, a means for comparing the sum of absolute values of differences between image signals of adjacent screens on the time axis in one group of screens with a predetermined threshold value is used. A three-dimensional orthogonal transform encoding method for a moving image signal according to any one of claims 1 to 3 to be used.