JPS62102685A - High-efficiency encoder - Google Patents

Abstract

PURPOSE:To obtain a high-efficiency encoder capable of improving the picture quality of a picture restored at the reception side by identifying a parameter at each block obtained by the division of one field. CONSTITUTION:Forecast data Ik(x, y) for an existing picture element is expressed as a prescribed linear coupling. The forecast equation is obtained by extracting a forecast value for a picture element of an existing field while an adjacent picture element having the strongest correlation is as a representative value and applying the correction in the time space direction. A parameter identifying section 2 uses reference data from a block circuit 1 to identify the parameter at each block by the least square method. That is, a true value Ik of a picture element of the existing field is a forecast value Ik' of the picture element corresponding to that obtained by a prescribed equation superimposed with a forecast error (e). Parameters W1-W35 are identified by using picture elements at representative points of a prescribed number obtained by thinning out for each plural samples.

Description

DETAILED DESCRIPTION OF THE INVENTION C. Industrial Application Field This invention relates to a high-efficiency encoding device applied to the transmission of image data.

[Summary of the invention]

The present invention provides a high-efficiency encoding device that reduces the number of bits per pixel of image data such as a digital television broadcast signal. ■It is divided into several blocks, and for each block, the pixel data of the current field is predictively encoded from the pixel data of several past fields, and by identifying the parameter that minimizes the sum of squares of the prediction error. , to obtain good restored image quality with very little distortion.

[Conventional technology]

An interframe encoding method that performs three-dimensional, ie, spatiotemporal, processing is known as a high-efficiency encoding method that reduces the number of bits per pixel. Interframe encoding methods include those based on motion detection and those based on motion compensation. In the former method, motion is detected based on the presence or absence of a frame difference, and only areas where there is no frame difference (that is, areas where there is no movement) are replaced with the data of the previous frame.

The latter method uses a block matching method etc. to find the position information (movement correction amount) between the current and previous frames, and then manipulates the previous frame image based on this movement correction amount to create correspondence between frames. It is something. In the block matching method, a screen is divided into a plurality of blocks, the amount and direction of movement are determined for each block, and the amount and direction of movement are transmitted.

[Problem that the invention seeks to solve]

The interframe encoding method using motion detection has a problem in that -r moving images have many moving parts and a low compression ratio.

Furthermore, the interframe coding method using motion compensation has the disadvantage that the compression rate cannot be said to be sufficiently low because distortion occurs due to block division and the amount of motion is transmitted for each block.

Furthermore, both methods have the disadvantage that when a moving object moves, the pixel data of the original area is lost (so-called uncovered background problem occurs).

Therefore, it is an object of the present invention to provide a highly efficient encoding device that can achieve a much higher compression rate than conventional devices.

Another object of the present invention is to provide a high rate encoding device that can cope with various movements caused by a plurality of moving objects by performing various corrections in the time direction.

Still another object of the present invention is to provide a highly efficient encoding device that eliminates problems such as edge blurring and uncovered background by performing various corrections in the spatial direction.

Also, application 1! IIi people first set the compression ratio to extremely high (
High-efficiency encoding device for television signals (patent application 1973)
9-174412). The object of the present invention is to improve this high-efficiency encoding device.

That is, what is shown in the above-mentioned application is to obtain the pre/llI value for a pixel in the current field by extracting a neighboring pixel with the strongest correlation as a representative value, and applying correction in the spatio-temporal direction to this representative value. The parameters for correction are 1
Identification is performed to minimize the sum of squares of prediction errors for all pixel data in the field. By determining parameters for each field and transmitting these parameters, the compression ratio can be extremely high. On the other hand, there is a drawback that the prediction error becomes large and the image quality is not sufficiently good when the image is restored on the receiving side.

Therefore, an object of the present invention is to provide a highly efficient encoding device that can improve the image quality of an image restored on the receiving side by identifying parameters for each block formed by dividing one field. It is in.

[Means for solving problems]

This invention comprises a memory for storing pixel data of several fields in the past, pixel data in a block obtained by dividing the current field, and pixel data of several past fields stored in the memory, which correspond to each other. A parameter identification unit that identifies parameters that define a spatiotemporal relational expression defined by linear one-shot precision from pixel data in a block, and pixel data of several past fields based on the parameters identified for each block. Prediction of the pixel data of the current field? The present invention is a high-efficiency encoding device characterized by having a prediction unit that performs JIII and transmitting identified parameters.

[Effect]

This invention predicts current motion from pixel data of several past fields. In this invention, since the motion information of each of a plurality of moving objects is included in the above pixel data, in other words, motion vectors having various directions and speeds also have strong temporal correlations, the amount of motion is There is no need to transmit motion correction at the level of each pixel. Since it is regarded as a temporal change in motion vector, motion models such as constant continuous motion (represented by the data of the past 2 fields) or constant acceleration motion (represented by the data of the past 3 fields) do not depend on the direction or speed of the motion vector. Since it can be handled uniformly as a motion model, it is only necessary to correct the deviation from the motion model.Therefore, according to this invention, it is possible to increase the compression ratio. Since the correction is performed on the original signal, there are no problems with professional distortion or anchored background.

Furthermore, in the present invention, since the calculation for parameter identification is performed for each block, the image of the received signal 1u) 1 can be restored satisfactorily.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings. A description of this one embodiment is provided in the following sequence.

a. Encoding device C. Decoding device C. Parameter identification a. Encoding device FIG. 1 shows an embodiment of the present invention, that is, the configuration of an encoding device provided on the transmitting side.

In FIG. 1, 1 indicates a blocking circuit, and 2:
The parameter identification part is shown. In this blocking circuit 1,
A digital television signal digitized with a predetermined sampling frequency of 14 waves, that is, image data in the current field, is input, and reference data -1, -2, and -3 in the past three fields is input to the prediction unit 3. It is supplied from field memories 4 and 5, respectively. The output signal of the blocking circuit 1 is supplied to the parameter identification section 2.

In the blocking circuit l, the order of the television signal is converted into the order of blocks. FIG. 2S shows an example of the blocking circuit 1. In FIG. In FIG. 2, 11.12.13.
Each of 14 represents a scanning line conversion circuit. Scanning line conversion circuit 1
1 is supplied with image data of the current field from an input terminal 7. The image data of the previous field is supplied to the scanning line conversion circuit 12 from the input terminal 8 . The scanning line conversion circuit 13 is supplied with image data of the previous field from the input terminal 9. The scanning line conversion circuit 14 has an input terminal 10
The image data of the field before the previous one is supplied from the field. At the output terminals 15, 16, 17, 18 of these scanning line conversion circuits 11, 12, 13, 14, data of each field is generated in block order.

As an example, in the case of a block formed by dividing one field into four, image data is outputted to the output terminal 15 block by block in the order of the numbers given in FIG. 3 regarding the image fiFk of the current field. Pond output terminal 16.17
．． 18 shows the previous field image Fk-1. Image Fk-2 of the previous field. Mae Mae F Mae Mae Shifield k-
Each image data of 3 is the 3rd image data as well as the field image Fk.
Each block is output in the order of the numbers given in the figure. The l1lff order of data output within l blocks is similar to television scanning. Scanning line conversion circuit 11
, 12, 13, and 14 each have a memory for blocking.

If one field is (210 lines x 684) pixels,
The total number of pixels is 143.640 pixels. As shown in Figure 3, when one field is divided into four, one professional,
The number of pixels included in the image is (105X342=35.91
0) Becomes a pixel. The smaller the 1-pro, 1'-size is, the lower the compression ratio will be, but on the other hand, the restoration error can be reduced.

The data stored in each of the field memories i and 5.6 is one prediction data, and using this prediction data and the image data of the current field, the parameter identification unit 2 calculates the sum of the squares of the prediction error by the least abandonment method. For example, 35 parameters W1 to w35 each having 8 focal points, which are the minimum, are identified for each block. The parameter identification unit 2 includes a line delay circuit and a sample delay circuit for adjusting the spatial positional relationship. Four parameters W1 to w35 for each field identified by the parameter identification unit 2
(in the case of 4 divisions) is the transmission data.

These parameters W1 to W35 are set to 1 for input data.
-1 after the field that is behind the field.

Reference numeral 3 indicates a prediction unit, and field memories 4.5.6 are written with prediction data from the prediction unit 3, and each of the field memories 4 and 5.6 stores information on past three fields.
Image data (prediction data) of 1, -2, and -3 are stored. The prediction unit 3 is located near the pixel to be predicted and calculates a predicted value for the current pixel using 35 pieces of prediction data included in the past three fields and parameters W1 to W35. . For this reason, the prediction unit 3
Also includes a plurality of line delay circuits and a plurality of sample delay circuits for adjusting spatial relationships. Furthermore, the prediction unit 3 uses a parameter wl-w35 for each block.
and generate prediction data in an order similar to television scanning.

The reason why the parameters are identified on the transmitting side using predicted data rather than real image data is to ensure the sameness with the image data restoration on the receiving side.

The predicted value for the pixel data in the current field (Fig. 5A) is the 10 pixel data in the vicinity of -1 in the previous field (Fig. 5B), and 1 in the vicinity of -2 in the field before that.
It is obtained as a linear one-shot precision of a total of 35 pixel data, including 5 pixel data (FIG. 5C) and 10 pixel data in the vicinity of -3 in the two fields before and after (FIG. 5D).

FIG. 5A - In the fifth diagram, the solid horizontal lines represent the lines scanned in field k and field-2, and the dashed horizontal lines represent the lines scanned in field-1 and field-3. represent. Let y be the line at the position where the pixel data in the current field is included,
The line located above it is y+1, and the line y
The line located in the upper law of ↑1 is y+2. The lines below line y are y-1 and y-2, respectively.
It is said that

In Figures 5A to 5, the vertical solid lines indicate sampling in each field (91i, which indicates the sampling position one sample before the sampling position X of pixel data in the current field and the sampling position two samples before this). They are X-1 and x-2, respectively. Also,
The sampling position after the sampling position X and the later sampling position are respectively defined as X % l and xT2.

Prediction data L(x, y) for the current pixel is expressed by the following linear equation.

Ty (x, y) = wl X↑−+
(x-2,y+1)÷W2XTll-4(x-1,y+
1) + w3 x↑−+ (x, y=1)+
W4×↑−+ (x=1.II) + w5 X↑
ni-+ (x+2.y+1)-! −w5X↑−+ (
x-2, y-1) +w7 X Tk-+ (x-L
y-1) + w3x - + (x, y-1) +w
9 x Tk-+ (xd, y-1)↓wlOx
T k-+ (x+2.y-1)" wll X r
k-z (x-2, y+2) Book W12X T k-
Z (x-Ly+2)-4-w13X ↑ k-z
(x, y+2) -I-W 14 X ↑ k
-z (X21.y=2)+w15X rk-2(
x;2. y+2) -1-w16x "i' to -2(
x-2,y)-kw17x↑w-z (X-Ly)
−j−w13x↑−2(x,y)=w19X r
k-2(x+1.y) -1-w20x↑h-z (
x+2. y)+W21X↑h-z (x-2, y-2)
4 W22X↑w-z (x-1, y-2)” w
23x T k-z (x, y-2) ÷w24×↑
h-z (x+1.y-2)+W25X T t+-z
(x:2.y-2) 10W26X↑w-x (x-2,
y+1) +W27X↑-3(x-Lv+1)-4
-w28X↑h-:+ (x, y+1) -1-w
29X Li (x; 1.yd)'-w3QX f
w-3(x;2.y+1) + W31 - to X↑
:+ (x-2, y-1)-! -W32X↑w-y
(x-Ly-1) +w33x↑ -1 (x, y-
1) Ki W34 x ↑ -3 (xTl, y-1) + w
35 This means that it is determined by applying the following corrections.

The parameter identification unit 2 uses the reference data from the blocking circuit 1 to identify parameters for each block by the least squares method. The true value of the pixel in the O1 current field is the predicted value of the pixel corresponding to this obtained by the above formula (
Since the prediction error e is superimposed on 16↑5, (
e−↑, −Ik), and the parameters w1 to w are such that the square of this prediction error is minimized for a predetermined number of pixels.
35 is calculated as described below.

In this case, the highest accuracy can be obtained by calculating the parameters W1 to w35 using the least squares method using all predicted pixels included in the block. It is practical to identify the parameters w1 to w35 using a predetermined number of representative point pixels.

In addition, in areas where there is no data at the periphery of the screen, the 6th
As shown in the figure, the same data as data a % h on the screen may be substituted as data outside the screen. Alternatively, as shown by the broken line in the sixth M, identification may be performed in a region that is one line inside and two samples inside.

pixel data of two past fields may be used for the current field, and in that case, a constant speed fake motion model will be expressed as a three-dimensional motion model.

b. Decoding device The decoding device that receives the encoded transmission data described above includes field memories 24 and 2 as shown in FIG.
5.26 and the parameters w1~ of each received block
w35 is supplied and the field memory 24.2
5.26 and a prediction unit 23 to which data of the past three fields is supplied. The prediction unit 23 forms restored data, that is, a digital television signal. On the receiving side, in order to restore one digital television word, an initial value for three fields is transmitted prior to the transmission of parameters W1 to W35, and this initial value is stored in the field memory 24, 25, 26. The prediction unit 23 records and clears the parameters of each received block, and the prediction unit 23 outputs prediction data in the same order as television scanning.

C. Identification of Parameters An example of a method using the method of least squares for identifying the parameters W1 to w35 performed in the parameter leg 2 on the transmitting side described above will be described below.

The above-mentioned linear equation for calculating the prediction/jIII data (x, y) can be expressed by the following determinant when prediction is made for one block of the current field.

Expressing the above equation together using a matrix and a vector, we get 〒−↑ ・W However, ↑ is the next (mXn) vector, and ↑ is (m x
The matrix W of n, 35) is a 35th order vector. (mXn) represents all pixels or representative points in the ■block.

On the other hand, the data (true value) of the pixel or representative point of the current field
The vector T formed by arranging is the (mXn)-order vector, and if e is the (mXn)-order hector, then the following equation is obtained. The above formula becomes e=1-cho・V^r. A parameter W that minimizes the sum of the prediction error vector e by a minimum value S is determined. The above formula is transformed as follows. 1 and T indicates the transposed matrix.

e” e=(T-↑W)T(D-↑W)-:r”
'I-4' ↑W -W"T" T = w ding ↑ "7tw Upper 2, e" The parameter W that minimizes e satisfies the following formula.The output of this formula is
(Ri1(e)yo Literature "System Identification Department (Publisher: Incorporated Association of Instrumentation and Automation M'<10 Society 9 Publication date: February 1, 1982)
0 (first edition)), Chapter 4, Section 11, Paragraph 2.

W, w = (↑7 ↑) - 1 block TT As it stands, in the case of (mxn) pixels in one block, a large matrix of (mxn, 35) will be handled, which is not practical. Therefore, the above equations are processed by converting them into matrices and vectors of small order.

That is, the (35, 35) matrix of (P-↑1 ・↑),
A 35th-order vector of (Q=T”・engineering) is used.

xl Tk-+ (X, -2, Vj A1) End-+
(Xi -1+yi ;1)1□ ...↑-+ (x, ;2.yJ-1)! two
2×Tw(xl, V, i) The above-mentioned P and Q are formed from the prediction data of the past three fields supplied to the parameter identification section 2. and,
The parameter W is calculated by (P −' Q).

[Effects of the Invention] This invention predicts the current motion from pixel data of the past several fields. Therefore, since the motion information of each of the plurality of moving objects is included in the above pixel data, There is no need to transmit the amount of motion, and only the parameters (coefficients for prediction) for each block need to be transmitted (,■
The average number of bits per pixel can be extremely reduced.

In addition, in this invention, since motion correction is regarded as a temporal change in the level of each pixel, constant velocity motion (expressed by the data of the past two fields) or constant acceleration motion (expressed by the data of the past two fields) is independent of the direction and speed of the motion vector. Since it can be handled uniformly as a motion model (expressed as three-field data), it is sufficient to simply correct the deviation of the motion model.

Therefore, according to the present invention, the compression ratio can be increased.

First, since the correction is performed three-dimensionally, problems such as block distortion and anchor hard bank ground do not occur.

Historically, in this invention, when identifying the parameters as the predicted i series using the least squares method, the compression ratio decreases somewhat because the identification is performed for each block, but the image restored on the receiving side is not distorted. is small and has good image quality.

[Brief explanation of drawings]

Fig. 1 is a block diagram of an embodiment of the present invention, Fig. 2 is a block diagram of an example of a block conversion circuit, Fig. 3 is a schematic diagram used to explain an example of block formation, and Fig. 4 is a block diagram showing a configuration for receiving encoded transmission data according to an embodiment of the present invention, and FIGS. 5M and 6 are schematic diagrams used to explain the embodiment of the present invention. Explanation of main symbols in Drawing 52 1: Niblocking circuit, 2: Parameter identification section, 3: Preliminary)
Lateral sounds 3, 4. 5.6: Field memory. Agent Patent Attorney Tadashi Sugiura Tomo x-2x-1x X+i x÷2 5th 5th Figure AC Figure 5B Figure 5D

Claims (1)

[Claims] A memory for storing pixel data of the past several fields, pixel data in a block obtained by dividing the current field, and pixel data of the past several fields stored in the memory, which correspond to the above block. means for identifying parameters that define a spatiotemporal relational expression defined by linear combination from pixel data in a block, and based on the parameters identified for each block, from pixel data of the past several fields. and means for predicting pixel data of the current field, and transmitting the identified parameters.