JP3426668B2 - Video coding method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a coding method in the case where an image signal is composed of a large number of channels such as a multi-lens type three-dimensional moving image.

[0002]

2. Description of the Related Art A conventional moving picture coding method is described, for example, in "International Standard for Multimedia Coding", edited by Hiroshi Yasuda, published by Maruzen P1.
As disclosed in 26-156, since the moving image is transmitted by one channel, there is no disclosure of a method or a problem in the case of encoding a plurality of channels.

[0003]

[Problems to be Solved by the Invention]
If the moving image coding method of the channels is to be extended to the multi-lens type three-dimensional as it is, each frame corresponding to the multi-lens requires a frame for performing the intra-frame coding in order to perform the coding individually, It can be predicted that the total amount of data increases as the number of channels increases, and as a result, the transmission efficiency of image signals deteriorates. Therefore, a coding method with a high compression rate is required.

The problem to be solved by the present invention is to realize a more advanced moving image compression method with a simple structure in view of the problems in such a simple extension of the prior art.

[0005]

The first invention of the present invention is 3
A method for encoding a multi-channel image signal forming a three-dimensional moving image, wherein intra-frame encoding is performed on a block-by-block basis only in a frame of one of the channels when intra-coding each channel. Is a moving picture coding method for obtaining the I picture by predicting coding the difference between the other channels and the frame of the intra-coded channel and obtaining the I picture. Is changed after a predetermined frame.

A second aspect of the present invention is a method for encoding a multi-channel image signal which constitutes a three-dimensional moving image, and when each channel is intra-frame encoded, a frame of any one channel is used. Only when intra-frame coding is performed in block units to obtain an I picture, and for other channels, motion compensation in the time axis direction is performed between the frame of the channel and the frame of the intra-coded channel. A moving picture coding method in which a motion vector is calculated and predictive coding is performed using this motion vector to obtain an I picture, in which a channel for intraframe coding is changed after a predetermined frame.

[0007]

In both the structure of the first invention and the structure of the second invention, only the data of one channel among all the channels is intra-coded, and the other channels are subjected to the predictive coding from the intra-coded frame. Since this is done, the amount of data transmitted can be reduced.

[0008]

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

FIG. 1 is a block diagram showing an embodiment of a multi-view 3D inter-picture prediction method. In the figure, I, I '
Is an I-picture (Intra-coded Picture), and B is a B-picture (Bidirectionally Predic
tive-coded Picture: bidirectional predictive coded image), P is P
Each picture (Predictive-coded Picture) is shown. Then, change the number of channels from ch1 to chN.
The number of N-channel and GOP (Group Of Picture) frames up to the above is defined as M frames.

In the method of FIG. 1, in image coding, intra-frame coding is performed only with an arbitrary I picture frame (frame number X) of an arbitrary channel (channel ch1 in the figure), and I picture frames of channels ch2 to chN. Predictive coding is performed for the frame corresponding to (1) picture frame in consideration of the positional relationship with the I picture frame of the channel ch1.

A second I picture frame (frame number X +
In M), this time, intraframe coding is performed only on the frame of channel ch2, and other channels ch1, ch3 to chN
Then, predictive coding is performed using the second I picture frame of the channel ch2 (this is referred to as an I'picture frame). Similarly, intraframe coding is performed only for one channel for every M frames, and predictive coding is performed for other channels using this frame.

For frames 2 to M for each channel ch1 to chN, forward prediction or bidirectional prediction coding is performed using I and I'picture frames that are temporally forward or backward, and P picture frames are generated. And B picture frame.

In this way, the intra-frame coding is performed only for one channel among the frames corresponding to the I-picture frame in sequence, and the predictive coding is performed using the I-picture frame for the remaining frames. Coding of the image signal is realized.

Next, a moving picture coding apparatus used in each of the channels ch1 to chN will be described with reference to FIG. It is assumed that each of the channels ch1 to chN is equipped with the same device, and that the input data is further divided into block units composed of 16 × 16 pixels.

First, the channel with the frame number X described above is used.
Consider the case of encoding ch1. A mode determination circuit 14 compares the original image data input from the first input terminal 16 with the prediction error image data, and determines either of the original image data and the prediction error image data according to the determination result. When the switch 2 is switched, it is sent to the encoding unit at the next stage. Since the channel ch1 is assumed to be encoded now, the original image data is sent.

The encoder is a DCT (Discrete Cosine Tran).
sform: Discrete cosine transform circuit 3, quantization circuit 4, V
An LC (Variable Length Coder) 5 and a buffer 6 are provided, and the encoded moving image signal is output from the buffer 6. The quantized moving image data output from the quantization circuit 4 is the inverse quantization circuit 7 and the inverse DCT.
It is input to the adder 9 via the circuit 8 and becomes the locally decoded image data of the previous frame and is stored in the motion compensation frame memory 10.

When the original image data is input to the motion vector detection circuit 11 and the frame is a P or B picture frame, the image data of the frame preceding and succeeding the frame and the image data of the frame are A motion vector is calculated for each macroblock. This motion vector is sent to the VLC 5 and the motion compensation frame memory 10.

The output from the frame memory 10 is sent to the adder 9 via the second switch 12. As described above, the data which is switched by the first switch 2 and sent is DCT-converted by the DCT circuit 3 and then obtained D
The CT coefficient is quantized by the quantization circuit 4, then variable-length coded by the variable-length coding circuit 5, and finally transmitted or recorded through the buffer 6. At this time, the mode determination circuit 14 operates the first switch 2 and the second switch 12 in an interlocking manner. Therefore, when performing the intra-frame encoding, the original image data directly passes through the first switch 2 from the first input terminal 16 and the motion compensation frame 10
The locally decoded data from is fed back to the adder 9 via the second switch 12.

The third switch 13 is switched according to the result of the mode decision circuit 14, and when the frame is an intra-frame encoded I frame, this data is sent from the third switch 13 via the output terminal 18 to another. It is sent to the encoder of the channel (same as the configuration of FIG. 2).

Next, referring to FIG. 1, encoding in the case of channels ch2 to chN in frame number X will be described. One method for obtaining the I'picture frame is to output the block-based original image data input from the first input terminal 16 from the motion compensation frame memory 10 by the subtracter 1. The difference between the decoded data of the frame and the prediction data of the current frame is calculated, and the difference data is output to the mode determination circuit 14 as a prediction error image and also output to one terminal of the first switch 2. There is a way to be done.

That is, in the case of calculating the prediction error image data, the current data obtained by using the decoded data of the previous frame sent from the motion compensation frame memory 10 by the subtracter 1 with respect to the data passed through the first input terminal 16 is used. The difference from the predicted data of the frame is taken, and this difference data is passed through the first switch 2 to D
Send to CT circuit 3.

The locally decoded data that has passed through the motion compensation frame memory 10 is fed back to the subtracter 1 and is also sent to the second switch 12 and the third switch 13 as described in the case of intraframe coding. To be

When predicting the I'picture frame of the channel in this way, the data of the I picture frame of the other channel is fetched from the second input terminal 17 to the predictor 15, and the prediction is performed here. You will need a picture frame.

An example of the predicting method of the predictor 15 will be described below with reference to FIG. In FIG. 3, the left shows the I picture frame which is the X frame of the channel ch1, and the right shows the I'picture frame of the channels ch2 to chN.

T of each frame (t is an arbitrary natural number)
The pixel value of the pixel a and the predicted pixel value of the pixel a ′ corresponding to the th block coordinate (i, j) (i and j are arbitrary natural numbers) are A (a) and A ′ (a), respectively.

Further, it is assumed that the pixels of the frames of the channels ch2 to chN are predicted in order from the upper left corner to the lower right corner of the frame as indicated by the arrow, and the pixel (a'-1) to the left of the pixel a'obtained immediately before is predicted. The pixel value of is set to A (a'-1).

In this way, the respective values are set and the prediction calculation as shown in the equation 1 is performed.

[0028]

[Equation 1]

FIG. 4 shows a circuit block when the prediction algorithm according to the equation 1 is developed in hardware. In the same figure, the data of the I picture frame of the other channel input from the second input terminal 17 in FIG. 2 is stored in the memory 101. Then, one pixel of data is sent out from the memory 101.

The switch 104 is switched depending on the coordinates of the data for one pixel in the block. If the coordinates (i, j) = (i, 1) of the pixel, the data sent from the memory 101 is directly switched. Take 104.

On the other hand, if (i, j) ≠ (i, 1), the data sent from the memory 101 is added to 1 by the adder 102.
Coordinates (i, j-1) sent from the pixel delay device 106
The predicted data of the pixels of
Is divided by.

The data passing through the divider 103 is sent to the buffer 105, and one block of data is sent out every time the buffer 105 stores one block of data, and the motion compensation frame memory 10 of FIG. Sent to.

Other prediction methods of the predictor 15 are possible. For example, as shown in the circuit block diagram of FIG. 5, the I picture frame data of another channel is input from the first input terminal 201 and the I picture frame data of the channel to be predicted is input from the second input terminal 202 in block units. , To the vector calculator 203.

The vector calculator 203 uses it for well-known time-axis direction prediction based on the data of both block units (for example, see Hiroshi Yasuda's "International Standard for Multimedia Coding" published by Maruzen, P126-156). A vector corresponding to the motion vector is calculated and sent to the reconstructor 204.

The decompressor 204 uses the I picture frame data of another channel and the vector calculator 203.
Data of the I picture frame of the channel to be predicted is restored by the vector calculated by.

The data of the I picture frame thus restored is input to the subtractor 1 of FIG. 2 through the output terminal 205, and the difference from the data of the I picture frame of the relevant channel is calculated to obtain the I'picture. Data as a frame is obtained.

[0037]

As described above, the present invention enables reduction of encoded data in encoding data having a plurality of channels such as a multi-lens type three-dimensional image.
This can be expected to improve the transmission efficiency.

[Brief description of drawings]

FIG. 1 is a multi-lens system 3 to which a moving image encoding device of the present invention is applied.
It is a functional block diagram of the prediction method between dimension images.

FIG. 2 is a circuit block diagram showing a configuration of a moving image encoding device for one channel.

FIG. 3 is a diagram illustrating a prediction method of the predictor in FIG.

FIG. 4 is a circuit block diagram showing a hardware configuration of the predictor of FIG.

5 is a circuit block diagram showing another hardware configuration of the predictor of FIG.

[Explanation of symbols]

1 subtractor 2, 12, 13, 104 switch 3 DCT circuit 4 Quantization circuit 5 VLC 6 buffers 7 Inverse quantization circuit 8 Inverse DCT 10 Motion compensation frame memory 11 Motion vector detection circuit 14 mode judgment circuit 16, 17, 18 Input (output) force terminal

Continuation of front page (56) Reference JP-A-4-361499 (JP, A) JP-A-4-196998 (JP, A) JP-A2-131697 (JP, A) JP-A-2-130094 (JP , A) JP-A 2-50689 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) H04N ^7/ 24-7/68 H04N 13/00-15/00

Claims (2)

(57) [Claims] 1. A method of encoding a multi-channel image signal that constitutes a three-dimensional moving image, wherein when each channel is intra-frame encoded, only one of the frames of one channel is encoded in a block unit in a frame unit. Encodes I
With obtaining the picture, and other channels to obtain the I-picture performs predictive coding takes the difference between the frame of the channel encoded within the frame kinematic
This is an image coding method, in which a channel for intra-frame coding is set after a predetermined frame.
A moving picture coding method characterized by changing to . 2. A method for encoding a multi-channel image signal forming a three-dimensional moving image, wherein when each channel is intra-frame encoded, an intra-frame is performed in block units only in a frame of any one channel. Encodes I
In addition to obtaining a picture, for other channels, a motion vector when performing motion compensation in the time axis direction between the frame of the channel and the frame of the intra-coded channel is calculated, and this motion vector is used. Video for Predictive Coding to Obtain I-Picture
A coding method, in which a channel for intra-frame coding is set after a predetermined frame.
A moving picture coding method characterized by changing to .