JP3159365B2 - Video signal transmission method and video signal transmission device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[Table of Contents] The present invention will be described in the following order. BACKGROUND OF THE INVENTION Problems to be Solved by the Invention Means for Solving the Problems Embodiments of the Invention (FIGS. 1 to 8) (1) Principle of Video Signal Coding Method (FIGS. 1 to 3) (2) Configuration of Embodiment (2-1) Image Signal Transmission System, Transmitter (FIGS. 4 to 6) (2-2) Adaptive Prediction Data Forming Circuit (FIGS. 5 to 7) (2-3) Receiver (FIGS. 6 and 8) (3) Operation and Effect of Embodiment (FIGS. 5 and 6) Effect of the Invention

[0002]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video signal transmission method and a video signal transmission apparatus, and more particularly, to a video signal transmission method suitable for transmitting a video signal.

[0003]

2. Description of the Related Art Conventionally, in a so-called moving picture video communication system for transmitting a video signal composed of a moving picture to a remote place such as a video conference system or a video telephone system, it is necessary to efficiently use the transmission capacity of a transmission line. As a method of increasing the transmission efficiency of significant information, without transmitting the entire successive frame images, a so-called frame dropping process of thinning out a predetermined frame is transmitted to the receiving device side, and the transmitting device transmits the frame image. Using the motion vector transmitted from the device side in place of the video signal of the dropped frame, performing an interpolation operation on the frame image subjected to the frame drop processing on the transmission device side based on the frame image information before and after the frame image. Has been proposed (Japanese Patent Laid-Open No. 60-28392).

[0004]

According to this method, instead of transmitting frame image information which has been theoretically subjected to frame drop processing, it is only necessary to transmit motion vector information having an information amount smaller than the information amount. It is considered that the significant information of the moving image can be transmitted efficiently.

However, since it is not always possible to form accurate motion vector data as motion vector information formed when executing a frame dropping process in a transmitting device, the receiving device is formed by interpolation calculation. There is a possibility that the contents of the frame image may be degraded to a degree that is not practically visible.

When moving picture information is obtained in a practical transmission device, a video signal to be transmitted is encoded using, for example, pixel information of 8 pixels × 8 lines per line as a transmission unit block. If the method is adopted, if the motion vector for the transmission unit block is incorrect, the phenomenon that the content of the frame image formed by the interpolation operation on the receiving device side varies for each transmission unit block. As a result, only low-quality images that are unsightly in practical use can be reproduced.

The present invention has been made in view of the above points, and is intended to propose a video signal transmission method and a video signal transmission apparatus capable of reproducing a compression-encoded video signal with high quality. is there.

[0008]

According to the present invention, there is provided a video signal transmission method for encoding and transmitting a digital video signal, wherein a first first frame (F1X) is encoded by intra-frame encoding. , And a plurality of subsequent second frames (F3X, F5X) are represented by at least a motion vector and difference data from a frame to be intra-coded. It is composed of frames encoded by inter prediction coding, and is provided between a first frame (F1X) coded intra-frame and a second frame (F3X) coded by inter prediction. A pixel to be predictively coded by at least the difference data and the motion vector from the first frame to be coded in the previous frame;
At least the difference data from the second frame to be predictively coded by the next frame and the pixel to be predictively coded by the motion vector, at least the difference data and the motion vector from the first frame to be coded by the previous intra-frame, A pixel predicted and coded by the difference data and the motion vector from the next frame to be predictively coded by the next inter-frame, or a third predictively coded frame composed of pixels selected from the pixels. Frame (F
2X) between the second frames (F3X, F5X) that have been subjected to the inter-frame predictive coding, the predictive coding based on at least the difference data and the motion vector from the second frame that has been subjected to the previous inter-frame predictive coding. Pixels,
At least pixels predicted and coded by difference data and motion vectors from the next inter-frame predictive coded second frame, at least difference data and motion vectors from the previous inter-frame predictive coded second frame And a fourth frame predictively coded as a frame composed of pixels predicted and coded by the difference data and the motion vector from the second frame to be coded next and the next frame, or a pixel selected from the pixels (F4
X), a first frame (F1X), a plurality of second frames (F3X, F5X), and a third frame (F1X).
2X) and the fourth frame (F4X) are divided into a first frame (F1X), a second frame (F3X), a third frame (F2X), a second frame (F5X), and a fourth frame (F4X). ).

[0009]

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the drawings.

(1) Principle of Video Signal Coding Method When the video signal coding method according to the present invention is applied to a moving image signal transmission system, image information of an interpolated frame is formed in a transmitting device by a method as shown in FIG. And transmit it to the receiving device.

That is, as shown in FIG. 1A, the 0th, 1st, 2nd, and 3rd video signals VD to be transmitted are transmitted.
... Frame data F0, F1, F of the th frame
At 2, F3,..., If there is a change in the image represented by the motion vectors x ₀ , x ₁ , x ₂ , x ₃ ,. Frame) every other frame, ie the second,
The fourth frame is designated as the interpolation frame and the interpolation frame processing is executed to form interpolation frame data F2X, F4X... As shown in FIG. F4X... Correspond to the remaining non-interpolated frames, that is, non-interpolated frame data F1X, F3X, F5X... Corresponding to the first, third, fifth. along with the data of x ₀ , x ₁ , x ₂ , x ₃ ...
The data is transmitted to the receiving device as transmission data DATA.

Here, the transmission data DATA is composed of image data subjected to high-efficiency encoding processing as shown in FIGS. This high-efficiency coding is, for example, as shown in FIG. 2A, at time t = t ₁ , t ₂ , t _3, ..., The moving picture images PC1, PC2, PC3,. In the transmission processing, the video signal has the feature that the autocorrelation is large, and it is designed to improve the transmission efficiency by compressing the digital data to be transmitted. And an inter-frame encoding process.

In the intra-frame encoding process, as shown in FIG. 2A, for example, images PC1, PC2, PC3,... Are interpolated between pixel data one-dimensionally or two-dimensionally along a horizontal scanning line direction. , And thus the transmission frame image data of the number of compressed bits is formed for each of the images PC1, PC2, PC3,....

The inter-frame encoding process is performed as shown in FIG.
As shown in FIG. 3, images PC1 and PC2, PC
Images PC12, PC2 representing the deviation between 2 and PC3 ...
.., And these are used as motion vector data x ₀ , x ₁ , x ₂ , x _3, ... Representing the motion of the image, and difference data between adjacent images in sequence, as the initial image P at time t = t ₁ .
It is transmitted to the receiving device together with the image data of C1 (consisting of intra-frame encoded data). Thus, the images PC1, PC
2. It is possible to transmit a video signal with high efficiency encoding to digital data having a smaller number of bits as compared with the case of transmitting all pixel data of PC3.

The moving picture signal transmission system according to this embodiment converts image data to be transmitted as shown in FIG.
A predetermined number of frames (for example, 10 frames) are divided into one block, and the block data... BL (N-
1), BLN, BL (N + 1)... Are sequentially subjected to high-efficiency encoding processing, and then transmitted from the transmitting apparatus to the transmission path in that order.

Block data BLN (N =... N−1,
..) Are the first frame data D1
Has intra-frame encoded processing data as
Interframe coding processing data is included as the tenth frame data D2 to D10.

Here, the intra-frame encoding process is shown in FIG.
As described above, the image data includes the difference data for all the pixels forming the image for one frame, and the receiving apparatus sequentially adds the difference data for the one frame to the frame image data representing one image. To reproduce. On the other hand, the second-
As described above with reference to FIG. 2B, the tenth frame data D <b> 2 to D <b> 10 are motion vector data and difference data representing the difference between the inter-frame images only for the changed pixels in the successive frame images. Is converted.

Thus, in practice, the first frame data D1 constitutes transmission data having a relatively low compression efficiency (thus having a large number of bits) by constituting data representing the differences of all pixels for one frame. Whereas the second
The tenth frame data D2 to D10 constitute transmission data having a relatively high compression efficiency (accordingly, having a small number of bits) representing only movement between image data.

The transmission device of the moving picture signal transmission system is shown in FIG.
When the transmission data DATA is formed by performing a high-efficiency encoding process on the moving image video signal VD shown in FIG. 7A, the first transmission non-interpolated frame data F1X is formed as intra-frame encoded data, and , The fifth transmitted non-interpolated frame data F3X, F5X... Are converted into inter-frame coded data (motion vector data (x ₁ +
_{x 2), (x 3 +} x 4) ...... and the image and a deviation data) and by connexion formed, transmit interpolated frame data F2X between further such a configuration transmission non-interpolated frame data,
F4X... Are adaptively added as necessary, and block data... BLN (N =... N-1, N, N + 1) using 10 frames of transmission frame data as a whole.
……).

Here, the transmitting apparatus converts the sequentially interpolated frame data F0, F1, F2, F3,... Constituting the moving picture video signal VD as transmission interpolation frame data F2X, F4X.
.., Motion vector data x ₀ representing a change in image information between
The transmission interpolation frame data F2X, F4X... are decoded by an interpolation operation based on x ₁ , x ₂ , x ₃ ... by a plurality of predetermined arithmetic methods (the image reproduced thereby is called a predicted frame image). ), The one with the smallest error among the predicted frame images is selected as the optimal transmission interpolation calculation data and transmitted to the receiving device side.

That is, first, the K-th (K = 2, 4,...)
..)), The transmission interpolation frame data FK corresponding to the interpolation frame data FK of the frame (this is called the current frame).
When obtaining X, the transmitting apparatus uses the first interpolation calculation processing method S
The following equation as P1

(Equation 1) , The intra-frame encoded data FK _INTRA for the K-th interpolated frame data FK is calculated as the transmission interpolated frame data FKX.

Second, the transmitting apparatus uses the following equation as a second interpolation calculation processing method SP2.

(Equation 2) For the K-th current frame data FK,
The inter-frame coded data F (K + 1) _INTER with the next frame data F (K + 1) is obtained as transmission interpolation frame data FKX. At this time, the inter-frame coded data F (K + 1) _INTER is composed of motion vector data [−−] between the current frame data FK and the next frame data F (K + 1).
x _K ] and the frame image moved so that the next frame data F (K + 1) is returned by the amount of the motion vector data [−x _K ] and the image data representing the deviation between the current frame image to be transmitted. Data F (K + 1) [-x
_K ].

Third, the transmitting apparatus uses the following equation as a third interpolation calculation processing method SP3.

(Equation 3) , The inter-frame coded data F (K-1) between the previous frame data F (K-1) and the current frame data FK.
_The transmission interpolation frame data FKX composed of _INTER is obtained.
The inter-frame coded data F (K-1) _INTER includes the previous frame data F (K-1) and the current frame data FK.
Between the motion vector data [x _K-1 ] and the frame image obtained by moving the previous frame data F (K-1) by the motion vector data [x _K-1 ] and the current frame image to be transmitted. F (K
-1) [x _K-1 ].

Fourth, the transmitting apparatus uses the following equation as a fourth interpolation calculation processing method SP4.

(Equation 4) , The inter-frame coded data F (K + 1) _INTER between the current frame and the next frame and the inter-frame coded data F (K-1) between the current frame and the previous frame
The average value with _INTER is obtained as transmission interpolation frame data FKX. The transmission interpolation frame data FKX includes the deviation data of the current frame reproduced based on the next frame data F (K + 1) and the motion vector [−x
_K ] and the deviation data of the current frame reproduced based on the previous frame data F (K-1) and its motion vector [xK _-1 ].

The transmitting apparatus determines, among the transmission interpolation frame data F2X obtained by the interpolation calculation processing methods SP1 to SP4 of the equations (1) to (4), the transmission interpolation frame data F2X between the current frame data to be transmitted. The data having the smallest deviation is selected and transmitted to the receiver as transmission interpolation frame data FKX constituting transmission data DATA.

If the video signal is encoded with high efficiency by such a method, when obtaining the transmission interpolation frame data FKX, the calculation obtained by the interpolation calculation processing steps of the equations (1) to (4) is used. By selecting the interpolated frame data having the smallest error among the results as the transmission data DATA, even if the error becomes abnormally large due to inappropriate calculation of the motion vectors x ₀ , x ₁ , x _2, ... In this case, the optimum transmission interpolation frame data can be selected and transmitted, so that when the frame image of the interpolation frame is reproduced in the receiving device, the image quality of the frame image can be further improved.

(2) Configuration of Embodiment (2-1) Image Signal Transmission System, Transmitting Unit FIG. 4 shows an image signal transmitting system to which the above-described image signal encoding method is applied. Transmission data D obtained by encoding the moving picture video input signal VD _IN with high efficiency
The data is converted into ATA, transmitted from the transmission circuit unit 12 to the reception circuit unit 14 via the transmission line 13, and the reception data DATAX obtained in the reception circuit unit 14 is converted into a moving image output video signal VD _OUT in the reception unit 15.

The transmitting section 11 receives the moving picture video input signal VD _IN at the image data input section 21, and as shown in FIG. 5, the luminance signal Y and the chrominance signal C constituting the moving picture video input signal VD _IN.
The _R and C _B to form an image data PIC constituting a moving by giving a single field drop circuit 25 and single field line thinning circuit 26 through an analog / digital converter circuit 22 as well as 23 and 24, which time base conversion Image data input unit 21 (FIG. 4) via circuit 27
Is sent to the image data encoding circuit 31 as an output.

As shown in FIG. 5, the image data encoding circuit section 31 fetches the image data PIC into the pre-filter 32 having a frame memory structure, and then, converts the image data PIC into the unit block data.
In step 3, the transmission unit block data composed of pixel data of 8 pixels (in the horizontal direction) × 8 lines (in the vertical direction) is read out, and the transmission unit block data is used as frame input data S1, and the data selection circuit 34 and the frame memory configuration are used. This is given to the interpolation frame input memory 35 and the motion vector detection circuit 36.

In this embodiment, a frame mode designation signal S2 separately sent from the system controller is applied to the selection input terminal SEL of the data selection circuit 34 and the write enable input terminal WRT1 of the interpolation frame input memory 35. Thus, as shown in FIG. 6, a frame input data S1 (FIG. 6 (A)) is the time t _10, t _11,
At t ₁₂ , t ₁₃ , t ₁₄ , t ₁₅ , t _16, ..., the 0th, 1st, 2nd, 3rd, 4th, 5th, 6th,. , F3, F4, F5, F
6, the frame mode designating signal S2 (FIG. 6B) is sequentially output from the interpolation frame encoding mode data INTPL, the intra-frame encoding mode data INTRA, and the interpolation frame encoding mode data I
NTPL, inter-frame coding mode data INTER,
The interpolated frame encoding mode data INTPL, the inter-frame encoding mode data INTER, and the interpolated frame encoding mode data INTPL are supplied.

The interpolated frame input memory 35, during the interpolation frame coding mode interval T _PL-frame mode designating signal S2 is decreased to the interpolation frame coding mode specifying data INTPL, write enable input WRT1 a write operation state (i.e. 6C, the interpolated frame encoding mode section T _PL , that is, from time t ₁₀ to the time t ₁₀ , as shown in FIG. 6D.
_{t 11, t 12 ~t 13,} t 14 ~t 15, t 16 ~t 17 arrived at the timing of ...... has frame input data S1 (FIG. 6
(A)), that is, 0th, 2nd, 4th, 6th ...
Th frame data F0, F2, F4, F6 captures ...... therein, the subsequent interval, i.e. the time t ₁₁ ~t
_{12, t 13 ~t 14, t} 15 ~t 16 frame data F0 captured the between ......, F2, F4, F6 have been made to retain ....

Thus, the interpolation frame input memory 35
The 0th, 2nd, 4th, 6th... Frame data captured in the interpolation frame encoding mode section _TPL is
Interpolated frame data S repeated for the next two frame periods
3 (FIG. 6D) so that the data can be transmitted to the second input terminal A2 of the data selection circuit 34.

The data selection circuit 34 determines that the frame mode designating signal S2 is the interpolation frame encoding mode data INTP.
Timing t ₁₀ ~t ₁₁ are decreased to L, t ₁₂ ~t _13, t
₁₄ ~t ₁₅ (shown Te code (A2) Niyotsu in FIG 6 (E)) by selecting, t ₁₆ ~t ₁₇ the second input terminal A2 with ...... given to the second input terminal A2 (F-1) of the interpolated frame data S3 (FIG. 6D),
(F0), (F2), (F4)... Are output as current frame input data S4 as shown in FIG. 6 (E) (FIG. 6 (E)). This together with the data selection circuit 34 is the interpolation frame coding mode interval T intraframe coding mode interval following the _PL T _RA and inter-frame coding mode interval T timing of _{ER t 11 ~t 12, t 13} ~t 14, t _{In the} period from _{15 to} t ₁₆ , the intra-frame encoding mode designating data INTR
A, and inter-frame encoding mode designation data INTER
Is given (FIG. 6B), the data (F1), (F1), (F2) of the frame input data S1 given to the first input terminal A1 (indicated by the symbol (A1) in FIG. 6E).
3), (F5),... Are transmitted as the current frame input data S4 as shown in FIG.

As a result, the current frame input data S4 becomes the frame data of the non-interpolated frame at the timing when the frame data F1, F3, F5... Of the non-interpolated frame of the frame input data S1 (FIG. 6A) arrive. F
.. As the current frame input data S4, and the frame data F0, F2, F4... Of the previous and previous frames at the timing when the frame data F2, F4, F6.
.. Are transmitted as the current frame input data S4.

As a result, the data selection circuit 34 sets three sets of frame data (F1, F2, F3), (F3, F4, F5), (F
5, F6, F7)... For the non-interpolated frame data (F1, F3), (F3, F5),
(F5, F7)... Are sequentially fetched as current frame input data F4, and then frame data F2, F4, F6. , (F3,
F5, F4), (F5, F7, F6)... Among non-interpolated frame data (F1, F3), (F3, F5),
(F5, F7)... Based on the frame data F2, F4, F6,... Of the interpolated frame, and the result of the predicted calculation is interpolated frame data F2 to be actually transmitted as the current frame input data S4. , F4, F6
.. And the adaptive prediction data forming circuit 41 (FIG. 4).

In FIG. 4, the current frame input data S4
Is supplied from the image data encoding circuit unit 31 to the adaptive prediction data forming circuit unit 41 as data constituting the prediction input data S5 together with the motion vector data, and is the predicted image data subjected to the prediction operation in the adaptive prediction data forming circuit unit 41. By feeding back the adaptive prediction operation result data S6 to the image data encoding circuit 31, the image data encoding circuit 31 converts the image data representing the deviation between the image data of the current frame input data S4 and the predicted image data into the transmission data DATA. It can be transmitted to the receiving unit 15 as a part.

To realize such a function, the image data encoding circuit 31 is provided with a data selection circuit 34 as shown in FIG.
Subtracts the current frame input data S4 output from the
2 as an addition input, and the prediction frame data S8 obtained as a part of the adaptive prediction operation result data S6 of the adaptive prediction data forming circuit 41 is supplied to the subtraction circuit 42 as a subtraction input. Thus, the subtraction circuit 42 adds the predicted frame data S8 to the current frame input data S4.
Is input to the discrete cosine conversion circuit 43 as transmission frame data S9.

The discrete cosine conversion circuit 43
For each frame data of the current frame input data S4 provided to the addition input terminal of the subtraction circuit 42, when the data of the predicted frame data S8 provided to the subtraction input terminal is 0, the discrete cosine conversion circuit 43 The conversion output data S10 obtained by performing high-efficiency encoding processing on the transmission frame data S9 in the intra-frame encoding mode is sent to the quantization circuit 44.

On the other hand, when frame data corresponding to the frame data of the current frame input data S4 is given as the predicted frame data S8, the discrete cosine transform circuit 43 operates in the inter-frame coding mode or the interpolation frame coding mode. The converted output data S10 obtained by performing the high-efficiency encoding process is supplied to the quantization circuit 44.

As a result, the quantization circuit 44 quantizes the converted output data S10 by a quantization step of a value corresponding to the quantization control data S11 given from the data generation amount calculation circuit 45, and runs the quantized data S12. After encoding into transmission encoded data S13 having encoded data suitable for transmission in the length Huffman encoding circuit 45, the transmission buffer circuit 51
The data is written into the buffer memory 52 (FIG. 5) of FIG.

The transmission buffer circuit 51 stores the transmission data D
As the ATA, data of a data amount corresponding to an allowable transmission amount that can be transmitted through the transmission circuit unit 12, the transmission path 13, and the reception circuit unit 14 is read out from the buffer memory 52 at a predetermined transmission speed and transmitted, so that the transmission unit 11 The data formed in step (1) can be reliably transmitted to the transmission unit 15.

On the other hand, when the amount of data generated in the transmission unit 11 becomes excessively large or small compared with the amount of transmission of the transmission data DATA, the data generation amount calculation circuit 45 detects this and detects the amount of data. By controlling the quantization step value used at 44, the buffer memory 5
2 is controlled so as not to overflow, so that all data formed by the transmission unit 11 to be transmitted can be taken into the buffer memory 52 without excess or deficiency.

The adaptive prediction data forming circuit 41 corrects the motion vector of the interpolated frame data F2, F4,... In the moving picture video signal VD (FIG. 1) to be transmitted in accordance with the accuracy of the motion vector. If the correct motion vector is transmitted, the transmission data is compressed by this amount, whereas if the motion vector is incorrect,
The adaptive prediction operation result data S6 (FIG. 4) capable of adaptively transmitting the deviation data corresponding to the incorrect motion vector error is formed by the following configuration.

(2-2) Adaptive Prediction Data Forming Circuit The adaptive prediction data forming circuit 41 inversely quantizes the quantized data S12 sent from the quantization circuit 44 in the inverse quantization circuit 55, as shown in FIG. After that, the quantized data S1 quantized by the quantization circuit 44 by being inversely transformed by the discrete cosine inverse transformation circuit 56.
2 is once decoded into image data before being sent to the run-length Huffman encoding circuit 45, and is supplied to the adaptive prediction circuit 58 as the current frame decoded data S15 via the addition circuit 57.

The adaptive prediction circuit 58 converts the current frame decoded image represented by the current frame decoded data S15 according to the motion vector data S16 supplied from the motion vector detection circuit 36 based on the current frame decoded data S15. The predicted frame data S8 representing the frame image moved by the vector data S16 or the restored frame image is generated. As shown in FIG. 7, the adaptive prediction circuit 58 writes the current frame decoded data S15 into the previous frame memory 61, at which time the image data held in the previous frame memory 61 is stored in the previous frame memory 61. The data is transferred to the previous / previous frame memory 62 as frame decoded data S21. At this time, the previous / previous frame memory 62 sends out the image data held so far as the preceding / previous frame decoded data S22.

As described above, the previous frame memory 61 and the previous previous frame memory 62 are quantized at a timing one frame before the current frame when the image data of the current frame is quantized. The previous frame decoded data S21 obtained by decoding the previous frame data and the previous previous frame decoded data S22 obtained by decoding the previous previous frame data quantized at a timing two frames earlier can be obtained. The decoded data of the previous frame S21 and the decoded data of the previous frame S22 are converted into motion correction data forming circuits 63 and 6, respectively.
4

The motion correction data forming circuits 63 and 64 are constituted by read variable memories 63A and 64A controlled by corrected motion vector data S23 and S24 obtained from the motion vector calculation circuit 65, and linear interpolation circuits 63B and 64B, respectively. Then, of the pixel data of the previous frame decoded data S21 and the previous previous frame decoded data S22, the frame image data obtained by moving by the amount of the corrected motion vector represented by the corrected motion vector data S23 and S24 is used as the previous / next frame data. Next frame motion correction data S
25 and the previous frame motion correction data S26 are supplied to the predicted frame data forming circuit 66.

Here, the current frame decoded data S15 is
As shown in FIGS. 5 and 6, the image data encoding circuit 3
1 as the current frame input data S4 (FIG. 6 (E)) transmitted from the data selection circuit 34 of FIG.
Each time F1, F0, F3, F2, F5, F4,... Are transmitted, as shown in FIG. 6 (F), the frame decoded data corresponding to the frame data F1U, F0
U, F3U, F2U, F5U, F4U... Are obtained through the addition circuit 57 (FIG. 5) and supplied to the previous frame memory 61 (FIG. 7).

Here, the frame mode designating signal S2 is applied to the write enable input terminal WRT2 of the previous frame memory 61.
(FIG. 6 (B)), and as shown in FIG. 6 (G),
The writing operation (that is, the ON operation) is performed at the timing when the frame mode designating signal S2 (FIG. 6B) becomes the intra-frame coding mode data INTRA or the inter-frame coding mode data INTER, whereby the previous frame is set. As shown in FIG. 6H, the previous frame decoded data S21 of the memory 61
Outputs the written previous frame encoded data in the frame section next to the frame section in which the writing operation state (ON operation state) has been entered, and the previous frame memory 61 writes this output state in the next frame section. The device is maintained by entering the device enable state (that is, the OFF operation state).

As a result, the preceding frame decoded data S21
Represents frame data F2 as frame input data S1.
, F3, F4 to F5, and F6 to F7 at the timing between two frames (FIG. 6A) sequentially.
First, third, fifth... Frame data F1, F3, F
5 Frame decoded data corresponding to F1U, F
3U, F5U (FIG. 6 (H)) are sequentially transmitted.

Similarly, the previous / previous frame memory 62 sends the frame mode designating signal S2 to the write enable terminal WRT.
3 and the frame mode designation signal S2 becomes the intra-frame coding mode data INTRA or the inter-frame coding mode data INTER (FIG. 6 (B)).
As shown in (I), it is controlled to the write enable state (that is, the ON operation state), and the previous frame decoded data S21 transmitted from the previous frame memory 61 at this time is
(FIG. 6H) is taken in.

Thus, as shown in FIG. 6J, the previous / previous frame memory 62 stores the frame input data S1 (FIG.
(A)) As the fourth, fifth, sixth,... Frame data F4, F5, F6... At the timing of arrival, the frame decoded data F1U, F1U, F
Pre-decoded data S22 consisting of 3U... (FIG. 6 (J))
Is transmitted.

As a result of the above operation, the interpolation frame data... As the current frame input data S4 (FIG. 6E) is obtained.
.., F2, F4, F6,... At the timing when the preceding frame decoded data S21 (FIG. 6 (H)) and the preceding frame decoded data S22 (FIG. 6 (J)) are obtained. , F4, F6... As a reference, frame decoded data of the next frame and the previous frame... F3U and F1U, F5U and F3U, F7
U and F5U... Are generated, and the decoded frame data of the next frame and the previous frame and the associated corrected motion vector data S23 (FIG. 6 (L)) and S2
4 (FIG. 6 (M)), the previous / next frame motion correction data S25 and the previous frame motion correction data S26 are generated.

The corrected motion vector data S23 (FIG. 6)
(L)) and S24 (FIG. 6 (M)) correspond to the motion vector detection circuit 36 (FIG. 5) in the motion vector calculation circuit 65.
Is calculated based on the motion vector data S16 given by Motion vector calculating circuit 65 as shown in FIG. 7, the motion vector data _{S16 (...... x 0, x 1} ,
x _2, x ₃ motion vector data written in the sequential write delay circuit 72 and 73 to the delay circuit 72 and 73 made of a one-frame delay memory through the input circuit 71 to a composed) ..., frame mode designating signal S2 (FIG. 6 (B))
There each interpolation frame coding mode interval T _PL becomes interpolation frame coding mode data INTPL, transfers the movement of the first frame memory configuration vector detecting circuit 74 and 75.

Thus, at the timing when the K-th (K =..., 1, 2,...) Frame data FK arrives as the frame input data S1 (FIG. 6A), the K-th frame is used as a reference. Vector data x representing the motion of the image data from the current frame and the previous frame.
_K and xK _-1 are sent from the delay circuits 72 and 73 to the motion vector detection circuits 74 and 75 (FIG. 6 (K1)).
And (K2)), the K transferred to the motion vector detection circuits 74 and 75 in the interpolation frame encoding mode section _TPL .
Th and K-1 th motion vector data x _K, and x _K-1 and continues sending from 2 between frame sections motion vector detection circuit 74 and 75 subsequent (Fig. 6 (K3) and (K
4)).

As a result, the frame input data S1 (FIG. 6)
(A)) as when the K-th frame data FK in the interpolation frame coding mode interval T _PL arrives, the motion from the circuit 72, 73, 74, 75 vector detection data x
_K , xK _-1 , xK _-2 , and xK _-3 are sent.
The circuit when the K-th frame data FK in the coding mode interval T _ER inter-frame has arrived 72,73,7
Motion vector detection data x _K from 4,75, x _K-1, x
_K-1 and xK _-2 are sent.

For example, the frame input data S1 (FIG. 6)
(A)) represents a first set of frame data,
K = 2nd frame data F as the current frame in the section where the third frame data F1 to F3 have arrived.
(In other words in the interpolation frame coding mode interval T _PL 2 it is coming, the current frame input data S4 (FIG. 6
(E)) in the 0th frame section frame data F0 has arrived), the motion vector detection circuit 74 transmits the motion vector data x ₀ as the motion vector detection data x _K-2, the motion vector detection circuit 75 Becomes a state in which the motion vector data x- ₁ is transmitted as the motion vector detection data xK _-3 .

Thus, the first set of frame data F
At the timing of the interpolation frame coding mode interval T _PL of the section 1~F3 is coming, the current frame input data S4 (FIG. 6 (E)) is the 0th frame data F0
Is supplied to the adaptive prediction circuit 58, and the motion vector calculation circuit 56 supplies the adaptive prediction circuit 58 with the interpolation frame supplied as the current frame input data S4, that is, the frames before and after the 0th frame data F0. Data F (-1) (FIG. 6 (J)) and F1 (FIG. 6 (H))
Motion vector data x ₀ and x ₋₁ (FIG. 6 (K
3) and (K4))).

On the other hand, when the K = 1st frame data F1 arrives as the current frame in the inter-frame coding mode section T _RA one frame before (FIG. 6A),
The motion vector detection circuit 74 calculates the motion vector data x _K-1 as the motion vector detection data x _K-1 corresponding to the previous frame.
_{At the same time} as transmitting ₀ (FIG. 6 (K3)), the motion vector detecting circuit 75 enters a state of transmitting the motion vector data x- ₁ as the vector detection data xK _-2 corresponding to the previous previous frame (FIG. 6 (K4)). )).

[0061] The 1 K = 3 th frame data F3 to interframe coding mode interval T _ER as entry connexion current frame after the frame arrives (FIG 6 (A)), the motion vector detection circuit 74 the previous frame The motion vector data x ₂ is sent out as the motion vector detection data x _K-1 corresponding to (FIG. 6 (K3)), and the motion vector detection circuit 7
5 is the vector detection data x _K-2 corresponding to the previous previous frame.
A state for sending motion vector data x ₁ (FIG. 6 (K4)).

In the same manner, a set of frame data successively successively, that is, third to fifth frame data F3 to F5,
Also in the timing of the interpolation frame coding mode interval T _PL of the fifth to seventh frame data F5~F7 ......, the frame data F2 for the current frame input data S4, F4 ......
And the next and next frame data F1 and F3, F3 and F5
, Motion vector data (x ₂ , x ₁ ),
(X ₄ , x ₃ ) are the motion vector detection circuits 74 and 75
Generated in

In addition to this, a motion vector detecting circuit 74,
75 is the current frame data F3, F5,... When in the inter-frame coding mode section _TER of the previous frame.
... and the previous frame (F2, F1), (F4, F
3) Motion vector data (x ₂ ,
x _1), (with x _4, x ₃₎ for generating a ......, when in the state of the interframe coding mode interval T _ER of the next frame, the current frame data F5, F7 ... before and before the previous frame ( F4, F3), (F6, F5) motion vector data for _{...... (x 4, x 3)} , (x 6, x 5) ......
Occurs.

When the frame mode designating signal S2 is in the state of the interpolation mode designating data INTPL (FIG. 6B), the motion vector calculating circuit 65 sets the frame data three frames before the K-th interpolation frame FK as a reference. F
(K-3) of the motion vector data x _K-3 (ie, the output data of the motion vector detection circuit 75) is input to the input terminal A11 of the data selection circuit 76 and the corrected motion vector data S
24 (FIG. 6 (M)) as the previous and previous frame decoded data S
22 is supplied to a given motion correction data forming circuit 64.

At the same time, the motion vector calculation circuit 65 outputs the frame mode designating signal S2 to the interpolation mode designating data I.
In the state of NTPL (FIG. 6B), the motion vector data x _K−2 of the frame data F (K−2) two frames before the K-th interpolated frame FK (ie, motion vector detection) After the output data of the circuit 74 is inverted by the polarity inversion circuit 77, the corrected motion vector data S23 is input through the input terminal A21 of the data selection circuit 78.
As (FIG. 6 (L)), the motion compensation data forming circuit 63 to which the preceding frame decoded data S21 is supplied is supplied as motion vector data -xK _-2 representing a motion that restores the motion of the frame image.

Thus, when the frame mode designating signal S2 comes to the timing of the interpolated frame encoding mode section _TPL , the preceding frame decoded data S22 composed of the K-3rd frame data F (K-3) U (FIG. 6 (J)), the frame image data moved by the amount of the motion vector data x _K−3 , that is, the (K−2) th frame image data is the previous frame motion correction data S26 representing the predicted interpolation frame image data. From the motion correction data forming circuit 64.

At the same time, the previous frame decoded data S21 (FIG. 6
(H) The frame image data returned by the motion vector data −x _K−2 from the frame image data of (H), that is, the K−2 th frame image data is formed as the motion correction data as the previous / next frame motion correction data S25. Circuit 63
Sent from.

For example, the frame input data S1 (FIG. 6)
(A)) as the interpolation coding mode section T _{PL at the} timing when the K = 4th frame data F4 has arrived (therefore, the current frame input data S4 (K−2 as FIG. 6 (E)).
= Timing at which the second frame data F2 has arrived), the motion vector calculation circuit 65 determines that K-2 =
K-1 = third non-interpolated frame data F which is the next frame centered on the second interpolated frame data F2
3 sends out a motion vector data -x ₂ as returned to the frame data F2 as compensation motion vector data S23 (FIG. 6 (L)), pre-made frame data K
The motion vector x _1, such as moving -3 = from the first non-interpolated frame data F1 to the frame data F2 is sent as a correction motion vector data S24 (FIG. 6 (L)).

Similarly, at the timing when the K = 6th frame data F6 arrives as the frame input data S1 (FIG. 6 (A)), the motion vector calculation circuit 65 outputs the motion vectors as the corrected motion vector data S23 and S24. sends a vector data -x ₄ and x _3, the same operation applies to the following other interpolation mode section.

In addition to this, the motion vector calculation circuit 65
When the frame mode designating signal S2 is the inter-frame coding mode data INTER, the motion vector data xK _-2 and xK _-1 are added by the adding circuit 79 and the corrected motion data is input through the input terminal A22 of the data selecting circuit 78. It is supplied to the motion correction data forming circuit 63 as vector data S23.

Here, when the frame mode designating signal S2 is K
At the timing of the inter-frame coding mode data INTER, the current frame input data S4 (FIG. 6)
(E) shows that the K-th non-interpolated frame data FK following the interpolated frame has arrived, whereas the previous frame decoded data S21 (FIG. 6 (H)) shows the frame of the interpolated frame. Data F (K-2) is maintained as it is.

Then, the motion correction data forming circuit 63 converts the frame image into the image data of the next non-interpolated frame (that is, the frame two frames later) based on the non-interpolated frame data F (K-2) two frames before. The motion vector data x _K−2 + x _K−1 corresponding to the movement up to is supplied to the motion correction data forming circuit 63, so that the next non-interpolated frame data F (K−2) is eventually obtained based on the previous non-interpolated frame data F (K−2). The non-interpolated frame data FK is predicted and calculated.

For example, when the K = third frame data F3 arrives as the frame input data S1, the previous frame decoded data S21 (FIG. 6)
(H)) is K-2 = 1st frame decoded data F1
U, and the K = third inter-frame coded image data F3 is predicted based on the motion vector data x ₁ + x ₂ (FIG. 6 (L)) based on the frame decoded data F1U.

Similarly, at the timing of K = 5th frame data F5, K = 5th interframe coding based on K-2 = 3rd previous frame decoded data (FIG. 6 (H)). Image data x ₃ motion vector data
+ X ₄ (FIG. 6 (L)). In the same manner, non-interpolated frame data in the inter-frame coding mode section is similarly predicted.

The prediction frame data forming circuit 66 forms the image data having the smallest deviation from the current frame input data S4 as the prediction frame data S8 using the previous / next frame motion correction data S25 and the previous frame motion correction data S26. First, the frame mode designation signal S
2 is the intra-frame encoded data INTTRA, the second is the interpolated frame encoding mode data INTPL, and the third is the inter-frame encoded mode data INTPL.
For each of the three cases of ER, predicted frame data S8 is formed.

The predicted frame data forming circuit 66 has a data selecting circuit 81 for performing a data selecting operation according to the frame mode specifying signal S2.
Is the intra-frame coding mode data INTRA (FIG. 6).
When (B)) is reached, “0” is passed through the selection input terminal A31.
The image data S31 is sent from the adaptive prediction circuit 58 to the subtraction circuit 42 (FIG. 5) as prediction frame data S8.

The "0" image data S31 is data representing a blank image containing no image information.
(FIG. 5) transmits the current frame input data S4 coming from the data selection circuit 34 as it is to the discrete cosine conversion circuit 43 as transmission frame data S9, thereby transmitting intra-frame encoded data as the converted output data S10. State.

The frame mode designating signal S2 is the inter-frame coding mode designating data INTER (FIG. 6B).
, The data selection circuit 81 sends the inter-frame coded image data S32 given from the inter-frame coding data selection circuit 82 to the subtraction circuit 42 (FIG. 5) as the predicted frame data S8 through the selection input terminal A32. .

The inter-frame coding data selection circuit 82 is constituted by a two-input selection circuit, and its first selection input terminal A4
1 is given “0” image data S31, and the second
Of the motion correction data forming circuit 63, the first minimum correction data priority circuit 83 is provided with the intra-frame deviation detection circuit 84 and the first motion deviation detection circuit 85. The selection control signal S35 for selecting the detection output data having the minimum value in response to the detection output data S33 and S34 is supplied to the inter-frame encoding data selection circuit 82.

The intra-frame deviation detection circuit 84 receives the current frame input data S4 and receives the transmission unit block average value circuit 8
6, the average value of the pixel data of the transmission block is calculated based on the pixel data of the current frame input data S4, and the average value is compared with the current frame input data S4 in the comparison circuit 87 as the reference data S36. The difference between the value of the pixel data of the input data S4 and the average value of the image data around the pixel data (the difference is determined by the image data to be transmitted and the image data to be transmitted by the intra-frame encoded data). Represents the error of
The detection output data S33 is provided to the minimum correction priority circuit 83.

As a result, the intra-frame deviation detecting circuit 84 converts the detected output data S33 representing the error between the frame data to be currently transmitted and the predicted frame data obtained by decoding the intra-frame encoded data into the first minimum value. It can be provided to the correction data priority circuit 83.

The first motion deviation detection circuit 85 compares the previous / next frame motion correction data S25 of the motion correction data forming circuit 63 with the pixel data of the current frame input data S4 by giving it to the comparison circuit 91 as reference data. ,
The deviation data S41 representing the deviation is supplied to an absolute value sum circuit 92.
And integrates the integration result into the detection output data S34.
Is transmitted.

Here, since the frame mode in which the input terminal A32 is selected in the data selection circuit 81 is the section of the inter-frame encoding mode INTER (FIG. 6B), the inter-frame encoding of the current frame input data S4 is performed. The contents in the mode are frame data F3 of the non-interpolated frame,
F5 (FIG. 6E). On the other hand, the motion correction data forming circuit 63 generates the frame decoded data F1U and F3U of the non-interpolated frame given by the previous frame decoded data S21 of the previous frame memory 61.
Based on (FIG. 6 (H)), frame data F3, F5,... Obtained by moving the corrected motion vector data (x ₁ + x ₂ ), (x ₃ + x ₄ ) are transmitted. You are in a state to do.

Then, the comparing circuit 91 of the first motion deviation detecting circuit 85 converts the frame data F3, F5... Of the current frame input data S4 and the frame decoded data F1U, F3U. (Based on the image data to be transmitted and the motion vector data (x ₁ + x ₂ ), (x ₃ + x ₄ )). (Which represents an error with respect to the image data to be described) is input to the absolute value summing circuit 92 as the deviation data S41.

As described above, the inter-frame coding mode section T
_At the timing of _ER , the first minimum correction data priority circuit 83 outputs the detection output data having the smallest error among the detection output data S33 of the intra-frame deviation detection circuit 84 and the detection output data S34 of the first motion deviation detection circuit 85. Is selected, and the detection output data S3 of the first motion deviation detection circuit 85 is selected.
4 is smaller than the previous / next frame motion correction data S2
5 is supplied to the data selection circuit 81 as inter-frame coded image data S32 through the inter-frame coding data selection circuit 82, whereby the frame data F1 predicted based on the motion vector for the non-interpolated frame as the predicted frame data S8. F3...

As a result, the non-interpolated frame F1, the output frame of the subtraction circuit 42 (FIG. 5)
.. Are sent to the discrete cosine conversion circuit 43, whereby the image data of the non-interpolated frames, ie, the third, fifth, seventh,..., Th frame data F3, F5, F7,. The data is supplied to the transmission buffer circuit section 51 as inter-coding data.

Next, the predicted frame data forming circuit 66
When the frame mode designation signal S2 becomes the interpolated frame encoding mode data INTPL (FIG. 6B), the interpolated frame encoded image obtained in the interpolated frame encoded data selection circuit 95 from the input terminal A31 of the data selection circuit 81 The data S45 is sent to the subtraction circuit 42 (FIG. 5) as the predicted frame data S8.

The first, second and fourth input terminals A51, A52 and A54 of the interpolation frame coded data selection circuit 95
Are provided with “0” image data S31, previous / next frame motion correction data S25, and previous frame motion correction data S26, respectively. At the same time, the previous / next frame motion correction data S25 and the previous frame motion correction data S26 are added to the third input terminal A53 of the interpolation frame coded data selection circuit 95 by the addition circuit 97 of the average value calculation circuit 96, and then 1 The average value motion correction data S46 obtained by the division in the / 2 division circuit 98 is provided.

The interpolation frame coded data selection circuit 95 selects one of the data at the input terminals A51 to A54 in accordance with the selection control signal S47 formed in the second minimum correction data selection circuit 99, and thus the interpolation frame code. at the timing of mode section T _PL, it has been made the image data as to reproduce the most high quality interpolated image at the receiving side so as to be selected to Adaputeibu.

First, the second minimum correction data selection circuit 9
9 receives the detection output data S33 of the intra-frame deviation detection circuit 84 as the first selection condition input, and when the detection output data S33 reaches the minimum value, this state is regarded as interpolation frame data as the intra-frame decoding method. When data other than the data of the frame type is transmitted, it is determined that only the data having an error larger than the data of the intra-frame decoding method can be generated on the transmission side, and the selection of the interpolation frame encoded data selection by the selection control signal S47. The circuit 95 selects the "0" image data S31 given to the input terminal A51, and uses this as prediction frame data S8.
To be sent. At this time, the discrete cosine transform circuit 43 sends out the converted output data S10 composed of the data of the intra-frame decoding method, and transmits the converted output data S10 to the receiving side as image data corresponding to the frame data F2, F4,... Of the interpolated frame.

Second, the second minimum correction data selection circuit 9
9 is the detection output data S of the first motion deviation detection circuit 85
When the detection output data S34 becomes minimum as compared with other input data by receiving the input data, this state is defined as frame data F2, F4 of the interpolated frame and a motion vector between the next frame image data. It is determined that the data is data capable of reproducing the highest quality image on the receiving side (that is, the data having the highest accuracy), and the selection control signal S4
7, the previous / next frame motion correction data S25 provided to the second input terminal A52 by the prediction frame data S25.
The number 8 is sent to the subtraction circuit 42 (FIG. 5).

In the interpolated frame coding mode section _TPL , as shown in FIG.
At the timing of the K-th frame data FK (FIG. 6A), the data selection circuit 78 of the motion vector calculation circuit 65 is transmitting the motion vector data −x _K−2 (FIG. 6 (L)). And the current frame input data S4 (FIG. 6 (E)) is the (K-2) th frame data F
At the same time as (K-2), the previous frame decoded data S21 (FIG. 6 (H)) has become the (K-1) th frame data F (K-1). As a result, the motion correction data forming circuit 63 outputs the previous / next frame motion correction data S25.
As, similar to the frame data F (K-2) (slave connexion current frame input data S4 representing an image returned by vector data -x _K-2 motion this based on the next frame data F (K-1) K -2nd frame) is decoded and output.

Therefore, if the value of the detection output data S34 obtained from the first motion deviation detecting circuit 85 is "0", this means that the motion compensation data is formed based on the motion vector data to be transmitted to the receiving side. This means that the predicted image data formed by the circuit 63 matches the image data of the current frame input data S4 replaced with the motion vector. At this time, the predicted frame data forming circuit 66 converts the predicted image data into the predicted frame data S
Since the transmission frame data S9 can be supplied to the subtraction circuit 42 (FIG. 5) as 0, the content of the transmission frame data S9 becomes 0.
As a result, the transmission data synthesizing circuit 46 has to send only the motion vector data to the buffer memory 52 as the image data of the interpolation frame. The amount of transmission data can be reduced.

Then, the transmitted motion vector data -x _K-2 transmits that the image data predicted based on the motion vector data -x _K-2 coincides with the current frame input data S4 to be transmitted. Since the data is confirmed by the unit 11, high-quality image data can be reliably obtained as image data reproduced by the interpolation operation on the receiving side.

On the other hand, if there is a deviation between the previous / next frame motion correction data S25 and the current frame input data S4, this means that the motion vector actually contains an error. At this time, the predicted frame data forming circuit 66 sends the image data including the error to the subtraction circuit 42 (FIG. 5) as the predicted frame data S8, so that the image data corresponding to the error is transmitted frame data. The data is supplied to the discrete cosine conversion circuit 43 as S9, and the transmission data synthesis circuit 46 transmits the image data corresponding to the error to the receiving side together with the motion vector data. As a result, when there is an error in the motion vector, image data for correcting the error can be transmitted to the receiving unit 15, so that, in this case as well, high quality image data can be reproduced on the receiving unit 15 side.

Third, the second minimum correction selecting circuit 99
Upon receiving the detection output data S50 obtained from the second motion deviation detection circuit 101, when the detection output data S50 is the minimum, the detection control signal S47 moves to the fourth input terminal A54 of the interpolation frame coded data selection circuit 95. The subtraction circuit 42 uses the previous frame motion correction data S26 obtained from the correction data forming circuit 64 as the predicted frame data S8.
(FIG. 5).

The second motion deviation detection circuit 101 compares the previous frame motion correction data S26 with the current frame input data S4 in the comparison circuit 102 to obtain deviation data S51, and outputs the deviation data S51 in the absolute value summation circuit 103. The arithmetic processing is performed to obtain the detection output data S50.

Thus, the second minimum correction data selection circuit 9
9 is when the detection output data S50 of the second motion deviation detection circuit 101 has become minimum compared with other input data,
This state corresponds to the frame data F2, F of the current interpolation frame.
4... In place of the image data of 4..., The motion vector data between the previous frame to be transmitted is determined to be the highest-accuracy data capable of reproducing the highest quality image on the receiving unit 15 side, and selection control is performed. The subtraction circuit 42 uses the previous frame motion correction data S26 provided to the fourth input terminal A54 by the signal S47 as the predicted frame data S8.
(FIG. 5).

As described above, in the interpolated frame encoding mode section _TPL , as shown in FIG. 6, the frame input data S1 (FIG. 6A) is changed to the K-th frame data F.
At the timing of K, the data selection circuit 76 of the motion vector calculation circuit 65 is in the state of transmitting the motion vector data xK _-3 (FIG. 6 (M)), and the current frame input data S4 (FIG. 6 (E)). )) Is the (K-2) th frame data F (K-2), and at the same time, the previous / previous frame decoded data S22 (FIG. 6 (J)) is the (K-3) th frame data F (K-3). is there. As a result, the motion correction data forming circuit 64 moves the frame data F (K-3) based on the previous frame data F (K-3) by the motion vector data xK _-3 as the previous frame motion correction data S26.
(K-2) Accordingly, similarly to the current frame input data S4, the (K-2) th frame is decoded and output.

Therefore, if the detection output data S50 obtained from the second motion deviation detection circuit 101 is "0", this is based on the motion vector data xK _-3 to be transmitted to the receiver 15. This means that the prediction data formed by the motion correction data forming circuit 64 matches the image data of the current frame input data S4 replaced with the motion vector, and at this time, the prediction data is changed to the prediction frame data S8. Can be supplied to the subtraction circuit 42 (FIG. 5), the content of the transmission frame data S9 becomes 0, so that the discrete cosine conversion circuit 43 does not have error data that actually needs to be transmitted as the converted output data S10. Thus, the transmission data synthesizing circuit 46 outputs the motion vector data as the image data of the interpolation frame. Only need to transmit only the _K-3 in buffer memory 52, becomes the minute amount of transmitted data to be Genchijimi.

The transmitting unit 11 notifies the transmitted motion vector data x _K-3 that the image data predicted based on the motion vector data x _K-3 matches the current frame input data S4 to be transmitted. Since the data is confirmed on the side, high-quality image data can be obtained as the image data reproduced by the interpolation operation on the receiving unit 15 side based on the confirmation.

On the other hand, the previous frame motion correction data S
26 and the current frame input data S4, this means that the motion vector data xK _-3 actually contains an error. When the image data is sent to the subtraction circuit 42 as the predicted frame data S8, only the image data corresponding to the error is supplied to the discrete cosine conversion circuit 43 as the transmission frame data S9. Will transmit the image data corresponding to the error to the receiving unit 15 together with the motion vector data. As a result, when there is an error in the motion vector xK _-3 , image data for correcting the error can be transmitted to the receiving side, so that high quality image data can be reproduced on the receiving side.

Fourth, the second minimum correction data selection circuit 9
Reference numeral 9 denotes a third input terminal of the interpolated frame coded data selection circuit 95 according to the selection control signal S47 when the detection output data S55 obtained from the third motion deviation detection circuit 105 is received and the detection output data S55 is minimum. A53, the average value motion correction data S46 given from the average value calculation circuit 96.
As the predicted frame data S8, the subtraction circuit 42 (FIG. 5).
To send to. The third motion deviation detection circuit 105 outputs the detection output data S by subjecting the deviation data S56 obtained by comparing the average value motion correction data S46 in the comparison circuit 106 to the absolute value summation operation in the absolute value summation circuit 107.
55.

Thus, the second minimum correction data selection circuit 9
9 indicates that when the detection output data S55 of the third motion deviation detection circuit 105 becomes minimum as compared with other input data,
This state corresponds to the frame data F2, F of the current interpolation frame.
4 motion vector data between the next frame and the previous frame to be transmitted in place of the image data _{...... (-x 2, x 1)} , (- x 4, x 3) ...... it is highest at the reception side It is determined that the data is the highest-precision data capable of reproducing the image of the image quality, and the average value motion correction data S46 provided to the third input terminal A53 by the selection control signal S47.
As the predicted frame data S8, the subtraction circuit 42 (FIG. 5).
To be sent.

In the interpolated frame encoding mode section _TPL , as described above with reference to FIG. 6, the frame input data S1 (FIG. 6A) is changed to the K-th frame data FK.
, The motion vector data calculation circuit 6
The data selection circuits 78 and 76 of No. 5 are transmitting the motion vector data -xK _-2 and xK _-3 (FIGS. 6 (L) and 6 (M)), and the current frame input data S4.
(FIG. 6E) shows the (K-2) th frame data F (K-
2) and at the same time, the previous frame decoded data S21
(FIG. 6 (H)) and the previous / previous frame decoded data S22
(FIG. 6 (J)) shows the (K-1) th frame data F (K-1) U and the (K-3) th frame data F (K
-3) U. As a result, the motion correction data forming circuit 63
Is K-2 as the previous / next frame motion correction data S25.
The frame decoded data F (K-2) obtained by returning the next frame decoded data F (K-1) U by the motion vector data -x _K-2 based on the frame data F (K-2). ) U (thus, the (K-2) th reference frame in the same manner as the current frame input data S4), and outputs it.
Is the frame decoded data F obtained by moving the previous frame decoded data F (K-3) U by the motion vector data xK _-3 as the previous frame motion correction data S26.
(K-2) U (thus, the (K-2) th reference frame like the current frame input data S4) is decoded and output.

Therefore, the contents of the average value motion correction data S46 obtained from the average value calculation circuit 96 are the same as the predicted frame data predicted based on the next frame decoded data F (K-1) U and the previous frame data F (K -3) An interpolated image having an average value of frame data predicted based on U is formed as predicted image data.

If the value of the detection output data S55 obtained from the third motion deviation detection circuit 105 is "0",
This means that the prediction interpolation consisting of the average value of the prediction image data formed by the motion correction data forming circuits 63 and 64 based on the motion vector data -xK _-2 and xK _-3 to be transmitted to the receiving side. This means that the image data matches the image data of the current frame input data S4 represented by the motion vector, and at this time, the predicted image data is used as the predicted frame data S8 by the subtraction circuit 42 (FIG. 5). The transmission frame data S9 can be supplied.
Is 0, the discrete cosine transform circuit 43 does not have error data that actually needs to be transmitted as the converted output data S10, and thus the transmission data synthesizing circuit 46 outputs the image of the interpolation frame. Only the motion vectors -xK _-2 and xK _-3 need to be sent to the buffer memory 52 as data, and the amount of transmission data can be reduced accordingly.

The transmitted motion vector data −x _K−2 and x _K−3 are the current frame to be transmitted by the image data predicted based on the motion vectors −x _K−2 and x _K−3. By confirming that the input data S4 coincides with the input data S4, high-quality image data can be obtained as the image data reproduced by the interpolation calculation on the receiving side based on the data.

On the other hand, the average value motion correction data S46
If there is a deviation between the current frame input data S4 and the current frame input data S4, this means that the motion vector actually contains an error. By transmitting the data as the data S8 to the subtraction circuit 42, the image data corresponding to the error is supplied to the discrete cosine conversion circuit 43 as the transmission frame data S9.
Transmits image data corresponding to the error to the receiving side together with the motion vector data.

As a result, when there is an error in the motion vector, image data for correcting the error can be transmitted to the receiving side, so that high-quality image data can be reproduced on the receiving side. Thus, the predicted frame data forming circuit 66
In the interpolation frame coding mode section _TPL , the K-th frame data FK is the frame input data S1 (FIG. 6).
(A)), the predicted frame data S
As shown in FIG. 6 (N), 8 is decoded by motion vector data (-xK _-2 ) based on "0" image data or the (K-1) th frame data F (K-1). Frame data F (K-2) [F (K-1)] or K-
Frame data F (K-2) [F (K-3)] decoded by motion vector data-xK _-3 based on the _third frame data F (K-3), or the K-1st frame data And K-3rd frame data F (K-1) and F (K-
3) frame data F (K-2) [F (K-1), F (K-
3)] is adaptively selected and transmitted.

The predicted frame data forming circuit 66
In the inter-frame encoding mode section _TER , the K-th frame data FK is the frame input data S1 (FIG. 6).
(A)), the predicted frame data S
As shown in FIG. 6 (N), “0” image data,
Or frame data FK [F (K-2)] decoded by motion vector data (xK _-2 + xK _-1 ) based on the (K-2) th frame data F (K-2). One is selected as adaptive and transmitted.

[0112] Further predictive frame data generating circuit is, K th frame data FK frame input data S1 in intraframe coding mode interval T _RA (Fig. 6 (A))
As shown in FIG. 6 (N), the image data “0” is transmitted as the predicted frame data S8.

(2-3) Receiving Unit The receiving unit 15 fetches the reception data DATAX obtained at the output terminal of the receiving circuit unit 14 into the buffer memory 121 of the image data encoding circuit unit 110 as shown in FIG.
The header separation unit 122 separates the header data D _HD , and sequentially performs inverse transform on the run-length Huffman inverse encoding circuit 123, the inverse quantization circuit 124, and the discrete cosine inverse transformation circuit 125, thereby adding the received frame data S 51 to the addition circuit. By supplying the selection prediction data 126 to the selection prediction data S52 obtained by the selection prediction circuit 127.

Receiving the addition output S53 of the addition circuit 126, the selection prediction circuit 127 determines the prediction conversion information specified by the receiving side based on the header data D _HD separated by the header separation circuit 122, and determines the prediction conversion information. The highly efficient encoded transmission data is decoded by forming the selection prediction data S52 based on. In the case of this embodiment, the header data D _HD includes frame mode data, prediction mode data, motion vector data, and quantization width data as prediction conversion information.

The frame mode data represents the encoding system used when encoding the transmission data of each frame in the transmission unit 11, and includes the intra-frame encoding mode,
Indicates the types of the inter-frame encoding mode and the interpolated frame encoding mode. The prediction mode data is the adaptive prediction data of the transmission unit 11 for frames other than the frame encoded in the intra-frame encoding mode, that is, frames encoded in the inter-frame encoding mode and the interpolation frame encoding mode. The type of image data predicted to be optimal for adaptation by the formation circuit unit 41 is shown.

The interpolation frame encoding mode section T
_In the case of the section of _PL (FIG. 6), the interpolation mode coded data selection circuit 95 outputs the "0" image data S3
1 is selected or the previous / next frame motion correction data S25
Is selected, the previous frame motion correction data S26 is selected, or the average value motion correction data S46 is selected.

When transmitting the frame data in the inter-frame coding mode section T _ER (FIG. 6), the prediction mode data is sent to the inter-frame coding data selection circuit 82.
It is configured by data indicating whether “0” image data S31 has been selected or previous / next frame motion correction data S25 has been selected. The motion vector data is a motion vector when image data moves between frames when transmitting frame data in the interpolated frame coding mode section _TPL and the interframe coding mode section _TER . The quantization width data is data representing the coding width used when quantizing in the quantization circuit 44 of the transmission unit 11.

Thus, the selection prediction circuit 127 determines, for the image data obtained at the output end of the adder circuit 126 from the discrete cosine inverse conversion circuit 125, various conditions used when the image data was encoded in the transmission section 11. , header data D _HD Niyotsu and haunch on the basis of the predictive conversion information represented Te, and decoded before the image data to be encoded using the prediction conversion information, select this predicted data S
By supplying 25 to the addition circuit 126, the image data consisting of the deviation data that arrives one after another is decoded and obtained as an addition output S 53 at the output terminal of the addition circuit 126.

The addition output S53 is subjected to inverse unit block processing in an inverse unit block circuit 128, and thereby the same signal format as the image data obtained at the input terminal of the unit block circuit 33 (FIG. 5) of the transmission unit. The time axis conversion circuit 13 of the reception data output unit 111 reproduces the image data of
Send to 1.

The time axis conversion circuit 131 reproduces the luminance signal S _Y1 and the chroma signal S _CL by performing time axis conversion of the image data supplied from the inverse unit block circuit 128 based on a predetermined clock signal CL. One field interpolation circuit 132 and one field line interpolation circuit 1
After performing the interpolation calculation in the 33, digital / by conversion to an analog signal in an analog converting circuit 134, 135, 136, the luminance signal Y _X and chroma signal C _RX and C _BX video output signal VD _OUT of the reception data composed of It is transmitted as an output signal of the output unit 111.

(3) Operation and Effect of Embodiment In the above configuration, the moving image input signal VD _IN input to the transmission unit 11 is provided from the image data input unit 21 to the image data encoding circuit unit 31 as image data PIC. Then, the image data encoding circuit unit 31 converts the data into the 0th, 1st,
The second, third, fourth,..., Th frame data F0, F
.., F2, F3, F4...
In accordance with the specification of the frame mode specifying signal S2 (FIG. 6 (B)), the first frame data F1 is subjected to the high-efficiency coding in the intra-frame coding mode. The fourth, sixth,... Frame data F2, F4, F6,... Are encoded in the interpolation frame encoding mode, and every other frame data between the interpolation frame encoding modes, ie, the third, Fifth, seventh,... Frame data F3, F5, F7
.. Perform a high-efficiency encoding process in the inter-frame encoding mode.

That is, the frame input data F1 (FIG. 6)
(A)) shows every other frame data F by the data selection circuit 34 and the interpolation frame input memory 35.
Frame data F0, F2, other than 1, F3, F5...
By shifting F4 and F6 by two frames, the first, zeroth, third, second, fifth,...
0, F3, F2, and F5 are arranged and the current frame input data S4 is obtained and supplied to the subtraction circuit 42, and also to the adaptive prediction circuit 58 of the adaptive prediction data formation circuit 41 to transmit the frame data to be transmitted to the receiving unit. Is decoded in the adaptive prediction data forming circuit 41, and prediction image data considered to be capable of reproducing the highest quality image on the reception side is selected for each frame, and transmission data corresponding to the prediction image data is transmitted to the reception side. I do.

That is, the image data encoding circuit 31 converts the current frame input data S4 from the subtraction circuit 42 to the discrete cosine transform circuit 43 and the quantized data S12 transformed by the quantizing circuit 44 to the inverse quantizing circuit 55 and the discrete The data is supplied to the adaptive prediction circuit 58 as the current frame decoded data S15 (FIG. 6F) through the cosine inverse conversion circuit 56 and the addition circuit 57.

The adaptive prediction circuit 58 stores the decoded current frame data S15 in the previous frame memory 61 and the previous previous frame memory 62 by storing them in two-frame intervals, and stores the decoded previous frame data S21 (FIG. 6 (H)). )
And the previous / previous frame decoded data S22 (FIG. 6 (J)) is generated, whereby various kinds of predicted image data are formed for each frame based on the relationship with the motion vector data obtained every moment in the motion vector calculation circuit 65. Then, by supplying the optimum image data of the predicted image data to the subtraction circuit 42 as the predicted frame data S8, the deviation between the current frame input data S4 and the predicted image data decoded by the motion vector data is obtained. Is combined with the motion vector data so as to be transmitted to the receiving side.

First, the frame input data S1 (FIG. 6)
(A)) at the timing of the intra-frame encoding mode section T _RA at which the first frame data F1 arrives,
The data selecting circuit 81 of the predicted frame data forming circuit 66 (FIG. 7) outputs the "0" image data S3 of the first input terminal A31.
1 is selected and the predicted frame data is selected (FIG. 6 (N)).
, The subtraction circuit 42 sends the first frame data F1 arriving as the current frame input data S4 to the transmission frame data S9 as it is. At this time, the transmission unit uses the intra-frame coded data. Is transmitted to the receiving unit side.

[0126] Subsequently, when the frame input data S1 is interpolated frame coding mode interval T _PL becomes the second frame data F2, predictive frame data generating circuit 66
The data selection circuit 81 shown in FIG. 7 enters a selection state in which the interpolated frame encoded image data S45 of the interpolated frame encoding selection circuit 95 is output as predicted frame data S8 from the input terminal A31.

At this timing, the 0th frame data F0 is obtained as the current frame input data S4 (FIG. 6 (E)), and the previous frame decoded data S21 of the previous frame memory 61 (FIG. 6 (H)) In a state where the frame data F1U is being output. At the same time, the motion vector calculation circuit 65 outputs the corrected motion vector data S2
3 in a state that is transmitted through the input terminal A21 of the data selector circuit 78 to the motion vector data -x ₀ (FIG. 6 (L)).

At this time, -1 is stored in the previous / previous frame memory 62.
Since the frame data F (−1) U is included,
Eventually, the interpolation frame decoded data selection circuit 95 of the prediction frame data formation circuit 66 has “0” image data S31
When the motion vector data by the -x ₀ connexion decoded frame data F0 before made by the F1 / next frame motion compensation data S25 as motion frame data F0 had it occurred in decoding the vector data x _-1 [F (-1)] and the average value motion correction data S46, which is the average value of these frame data F0 [F1] and F0 [F (-1)], are input. Is selected by the minimum correction data selection circuit 99, and the predicted frame data forming circuit 66 sends out the selected frame data as predicted frame data S8.

As a result, the transmission unit 11 obtains, in the subtraction circuit 42, transmission frame data S9 representing an error between the frame data selected as the predicted frame data S8 and the current frame input data S4. In step S8, together with the motion vector data used in generating the selected frame data and the header data D _HD including the prediction mode data indicating the type of the selected frame data as prediction conversion information, the data is transmitted to the receiving side as transmission data DATA. Send out.

Subsequently, when the frame input data S1 becomes the inter-frame coding mode period _TER for transmitting the third frame data F3 (FIG. 6A), the third frame data F3 is changed to the current frame input data. Data S4 (FIG. 6)
At the same time as being captured as (E)), the previous frame decoded data S21 in the previous frame memory 61 (FIG. 6 (H))
Maintain the state of transmitting the first frame data F1U. Motion vector calculating circuit 65 whereas the corrected motion vector data S23 (FIG. 6 (K)) the sum of the motion vector data x ₁ and x ₂ as, namely (x ₁ + x ₂₎
From the input terminal A22 of the data selection circuit 78.

Accordingly, the motion correction data forming circuit 63 outputs the first frame data F of the previous frame decoded data S21.
Based on 1U, the image is converted to motion vector data (x ₁
+ X ₂ ), the predicted image data F3 [F1] representing the third frame data is output as the previous / next frame motion correction data S25. However, at this timing, the data selection circuit 81 outputs the "0" image data S31 or the predicted image data F3 supplied from the input terminal A32 to the inter-frame encoding data selection circuit 82.
[F1] is sent as predicted frame data S8 (FIG. 6 (N)).

In this state, the motion vector data (x
_If the error of ₁ + x ₂ ) is not large, the predicted image data F
3 [F1] and the deviation between the third frame data F3 of the current frame input data S4 and the third frame data F3 are sufficiently small or 0, so that this is selected by the first minimum correction data priority circuit 83 and becomes the predicted frame data S8. Subtraction circuit 4
2 given. The error data thus obtained from the subtraction circuit 42 is the motion vector data (x ₁ + x ₂ ) and the predicted image data F3 [F
[1] together with the header data D _HD including the prediction mode data indicating that [1] has been selected as prediction conversion information.
1 to the receiving side.

Here, the motion vector data (x ₁ + x ₂ )
If the error of
Is detected by the minimum correction data priority circuit 83, the "0" image data is used as the predicted frame data S8, so that the image data transmitted from the transmitting unit 11 to the receiving side is switched to intra-frame encoded data.
As a result, the receiving unit 15 can reproduce a high-quality image.

Subsequently, the frame input data S1 (FIG. 6)
When (A)) becomes the fourth frame data F4, the current frame input data S4 (FIG. 6 (E)) becomes 2 at this time.
At the same time that the second frame data F2 is fetched, the previous frame decoded data S21 in the previous frame memory 61
(FIG. 6 (H)) transmits the third frame data F3U, and the previous frame decoded data S22 (FIG. 6 (J)) of the previous frame memory 62 transmits the first frame data F1U. become.

At the same time, the motion vector calculation circuit 65
Sends the motion vector data −x ₂ as the corrected motion vector data S23 (FIG. 6 (L)) from the input terminal A21 of the data selection circuit 78, and outputs the motion vector data −24 as the corrected motion vector data S24 (FIG. 6 (M)). Data x ₁
Is sent from the data selection circuit 76. As a result the motion correction data forming circuit 63 the previous frame decoding the third on the basis of the frame data F3U 2 th obtained by returning to the original only vector data -x ₂ motion which the frame data of the data S21 F2 [F3 ] Before
By transmitting it as the next frame motion correction data S25, it becomes possible to transmit this as predicted frame data S8 (FIG. 6 (N)).

At the same time, the motion correction data forming circuit 64
Is based on the first frame data F1U of the previous frame decoded data S22, and
By moving by _{one, the} second frame data F2
By transmitting [F1] as the previous frame motion correction data S26, the predicted frame data S8 (FIG. 6 (N))
Can be supplied.

At the same time, the average value calculating circuit 96
Image data F2 [F of the next frame motion correction data S25
3] and predicted image data F2 which is an average value of the predicted image data F2 [F1] of the previous frame motion correction data S26.
It becomes possible to supply the average value motion correction data S46 of [F3, F1] as the predicted frame data S8 (FIG. 6 (N)).

[0138] Thus, as shown in (FIG. 6 (N)), the interpolation frame coding data selector circuit 95, a "0" image data S31, the third frame data F3 and motion vector data -x ₂ connexion between the predicted image data F2 [F3] to expected, first by the image data F1 and the motion vector data x ₁ in connexion predicted prediction image data F
2 [F1] and the predicted image data F2 [F3, F1] obtained by calculating the average value from these predicted image data F2 [F3] and F2 [F1], the error of which is the largest The smaller one is selected by the second minimum correction data selection circuit 99.

At this time, the second minimum correction data selection circuit 9
9 determines that the difference between the predicted image data F2 [F3] predicted from the third frame F3 and the second frame data F2 of the current frame input data S4 is the smallest, the predicted frame data forming circuit 66 performs the prediction. The image data F2 [F3] is transmitted as predicted frame data S8. Thereby, the subtraction circuit 42 can obtain the correction data corresponding to the error as the transmission frame data S9, so that the transmission unit 11 eventually converts the image data representing the error into the motion correction vector data -x ₂ and 3
Th predictive mode data indicating that the those obtained on the basis of the frame data can be a transmission with header data D _HD comprising a predictive conversion information to the receiving section 15 side. Thus, the receiving unit 15 converts the second frame image into the motion vector data -x between the third frame image and the third frame image.
₂ can be reproduced.

On the other hand, when the second minimum correction data selecting circuit 99 determines that the deviation between the predicted image data F2 [F1] and the second frame data F2 of the current frame input data S2 is the minimum, The frame data forming circuit 66 supplies the predicted image data F2 [F1] to the subtraction circuit 42 as predicted frame data S8, so that the transmission unit 11 performs the prediction image data F2 [F1] and the current frame input data S4. The frame data F
2 together with the header data D _{HD including} , as prediction conversion information, prediction mode data indicating that the image data representing the error from the motion vector 2 is obtained from the motion vector data x ₁ and the first frame data F 1. To the side. Thus, the receiving unit 15 can reproduce the second frame data by moving the _first frame data by the motion vector data x1 based on the _first frame data.

Further, the second minimum correction data selection circuit 99
When it is determined that the average motion correction data S46 obtained from the average calculation circuit 96 is the minimum value, the predicted frame data forming circuit 66 outputs the predicted image data F2 [F3, F3
1] as the predicted frame data S8, the transmitting unit 11 transmits the first and second frame data F
1 and F3, the average value of the predicted image data for the second frame data obtained by moving the motion vector data −x ₂ and x _1, and ₂ of the current frame input data S 4
Th image data representing the error between the frame data F2, the predictive mode data indicating that the average value data and the motion vector data -x ₂ and x ₁ to the receiving unit 15 side together with the header data D _HD including the predictive conversion information Transmit. In this case the receiving unit 15, by interpolating calculates an average value of the second frame data obtained by moving from the third and first frame data F3 and F1 only motion vector data -x ₂ and x _1, 2 The second frame data can be obtained.

In this way, the receiving unit 15 can reproduce the image data based on the transmission data which is adaptively selected and transmitted. However, the transmitting unit 11 determines that the error is the smallest. Is small, and based on transmitting the image data, the receiving unit 15 can reproduce a high-quality frame image.

In the same manner, the frame input data S1
(FIG. 6A) is the fifth, seventh,... Frame data F5, F7 _,.
, The same operation as described above for the third frame data F3 is performed, and when the interpolation frame coding mode section _TPL becomes the sixth, eighth,... Frame data F6, F8,. The same data transmission processing as described above is executed for the frame data F4.

According to the above configuration, in the interpolation frame coding mode section _TPL and the inter-frame coding mode _TER , the transmission section 11 forms a plurality of prediction image data based on the transmission data transmitted to the reception side. Since the transmission data DATA is formed based on the predicted image data having the smallest error between the predicted image data and the current frame input data, the receiving unit 15 can always reproduce a high-quality frame image. it can.

[0145]

As described above, according to the present invention, after a first frame that has been intra-coded, inter-frame prediction is performed using a motion vector and difference data from the first frame using predictive coding. Forming a plurality of encoded second frames, and representing between the first and second frames and between the plurality of second frames, represented by motion vectors and difference data, and pre- or post-prediction. Or forming the third and fourth frames predicted in both directions,
The first frame, the second frame, and the third frame are transmitted in the order of the first frame and the next second frame, and the second frame is transmitted between the plurality of second frames. By transmitting in the order of the fourth frame, encoded video data having a large improvement in image quality in practical use can be easily restored on the decoding side.
Can be transmitted.

[Brief description of the drawings]

FIG. 1 is a schematic diagram illustrating the principle of a video signal encoding method.

FIG. 2 is a schematic diagram for explaining a high-efficiency encoding method and image data thereof;

FIG. 3 is a schematic diagram for explaining a high-efficiency encoding method and image data thereof;

FIG. 4 is a block diagram showing the overall configuration of the image signal transmission system.

FIG. 5 is a block diagram showing a detailed configuration of a transmission unit 11 of FIG.

6 is a time chart for explaining a decoding method of predicted image data in an adaptive prediction data forming circuit 41 in FIG. 5;

FIG. 7 is a block diagram showing a detailed configuration of an adaptive prediction circuit 58 in FIG. 5;

FIG. 8 is a block diagram showing a detailed configuration of a receiving unit 15 of FIG.

[Explanation of symbols]

11 ... transmitting unit, 15 ... receiving unit, 21 ... image data input unit, 31 ... image data encoding circuit unit, 33 ... unit block circuit, 34 ... data selection circuit, 36 ...
A motion vector detecting circuit, 41, an adaptive prediction data forming circuit, 42, a subtracting circuit, 51, a transmission buffer circuit,
52 buffer memory, 58 adaptive prediction circuit, 61
…… Previous frame memory, 62 …… Previous previous frame memory,
63, 64: a motion correction data forming circuit; 65: a motion vector calculating circuit; 66: a predicted frame data forming circuit.

Continuation of the front page (56) References JP-A-59-123383 (JP, A) JP-A-62-213494 (JP, A) JP-A-62-193383 (JP, A) Speech at the 1989 National Conference of the Institute of Television Engineers of Japan Proceedings, 19–20 “Adaptive Predictive Coding Method for Previous and Next Frames Suitable for Storage Media”. 485-486 (July 19, 1989)

Claims (2)

(57) [Claims] 1. A digital video signal is encoded and transmitted.
In the video signal transmission method, the first first frame(F1X)Is for intra-frame encoding
Composed of more encoded framesAnd  Subsequent second frames(F3X, F5X)Is
At least motion from the intra-coded frame
Frame represented by a vector and difference data
It is composed of frames encoded by inter prediction coding.
And  The first frame encoded in the intra-frame(F1
X)And the second frame encoded by the inter-frame prediction coding.
Laem(F3X)Between, at leastBeforeIs encoded in the frameThe firstH
Prediction code based on the difference data from the frame and the motion vector.
Pixels to be decoded, at leastNextOf the above-mentioned second frame,
From the difference data and motion vector from the frame
The pixel to be measured and encoded, at least the previous intra-frame encodedThe first
ofThe difference data and the motion vector from the frame
Next inter-frame predictive codedThe secondFrame
Predictive coded by the difference data and the motion vector.
Pixel, The above or other pixels Frame consisting of pixels selected from
Third frame predictively coded as a frame(F2X)
And  The second frame encoded by the inter-frame predictive coding.
(F3X, F5X)Between each other, at least the previous interframe predictive codedSecondH
Prediction code based on the difference data from the frame and the motion vector.
Pixels to be encoded, at least the next inter-frame predictive encodingSecondH
Prediction code based on the difference data from the frame and the motion vector.
The pixel to be encoded, at least the previous inter-frame predictive codedSecond
ofThe difference data and motion vector from the frame
Is interframe predictive codedSecondDifference from frame
Image to be predicted and coded by the
Elementary,The above or other pixels Frame consisting of pixels selected from
Frame encoded predictively as a frame(F4X)
Is provided,The first frame (F1X), the plurality of second frames
(F3X, F5X), the third frame (F2X
X) and the fourth frame (F4X)
Frame (F1X), the second frame (F3X),
The third frame (F2X), the second frame
(F5X), the order of the fourth frame (F4X)
Transmit A video signal transmission method characterized by the above-mentioned. 2. A digital video signal is encoded and transmitted.
In the video signal transmission device, the first first frame(F1X)For intra-frame encoding
Means comprising a more encoded frame, and a plurality of subsequent second frames(F3X, F5X)To
At least motion from the intra-coded frame
Frame represented by a vector and difference data
It is composed of frames encoded by inter prediction coding.
Means for performing, the intra-coded first frame(F1
X)And the second frame encoded by the inter-frame prediction coding.
Laem(F3X)Between, at leastBeforeIs encoded in the frameThe firstH
Prediction code based on the difference data from the frame and the motion vector.
Pixels to be decoded, at leastNextIs interframe predictive codedThe second
ofPredicted by the difference data from the frame and the motion vector
The pixel to be measured and encoded, at least the previous intra-frame encodedThe first
ofThe difference data and the motion vector from the frame
Next inter-frame predictive codedThe secondFrame
Predictive coded by the difference data and the motion vector.
Pixel,The above or other pixels Frame consisting of pixels selected from
Third frame predictively coded as a frame(F2X)
Means for forming the second frame
(F3X, F5X)Between each other, at leastBeforeIs interframe predictive codedSecondH
Prediction code based on the difference data from the frame and the motion vector.
Pixels to be decoded, at leastNextIs interframe predictive codedSecondH
Prediction code based on the difference data from the frame and the motion vector.
The pixel to be encoded, at least the previous inter-frame predictive codedSecond
ofThe difference data and motion vector from the frame
Is interframe predictive codedSecondDifference from frame
Image to be predicted and coded by the
Elementary, The above or other pixels Frame consisting of pixels selected from
Frame encoded predictively as a frame(F4X)
Means for formingThe first frame (F1X), the plurality of second frames
(F3X, F5X), the third frame (F2X
X) and the fourth frame (F4X)
Frame (F1X), the second frame (F3X),
The third frame (F2X), the second frame
(F5X) and transmitted in the order of the fourth frame (F4X).
Means to send Video signal transmission device comprising: