JPH01103387A - Compression device for picture data by prediction coding - Google Patents

Abstract

PURPOSE:To reduce the time till the start of display and to improve the compression rate by providing a means compressing a picture data of each frame by an in-frame prediction coding and a means compressing the data by inter-frame prediction coding. CONSTITUTION:A picture data of each frame inputted from an input terminal IN is stored in a compression data memory 4 as a compression data by the in-frame prediction coding via an in-frame coding circuit 1 and a quantization circuit 2 or the like. Moreover, the input data is stored in a compression data memory 13 as a compression data by the inter-frame prediction coding via an arithmetic circuit 10 and a quantization circuit 11 or the like. Then a comparison control circuit 17 compares the data quantity in the memories 4, 13 via the counts of the address counters 15, 16 and selects the less data quantity as the final compression data and the picture data by one frame in the frame memory 7 is transferred to a reference frame memory 9 when the data of the memory 4 is selected. Thus, the compression pate is improved and the time till the start of display is reduced.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an image data compression device used in image data recording/reproducing systems, transmission systems, and the like.

(Prior Art) In image data recording/reproducing systems, transmission systems, and the like, data compression is performed using predictive encoding in order to reduce the recording capacity and transmission capacity of frame-structured image data.

This predictive coding uses an appropriate method that utilizes the correlation between image data to predict the data that is likely to appear next, calculates the difference between this predicted value and the data that actually appears as a prediction error signal, and calculates the prediction error. This is done by variable length encoding the signal. In other words, the distribution of the error signal shifts to the small-level region as the probability of prediction accuracy increases, so if variable-length encoding is performed that assigns a short code to small-level data for such an error signal, the data can be reduced. Compression becomes possible. Furthermore, in order to further increase the compression rate, a compression method in which the prediction error signal is quantized and then variable-length encoded is also commonly used.

The simplest of the above prediction methods is to use correlation with immediately preceding image data in the same frame or correlation with image data at the same position in the immediately preceding frame and use these as predicted values. There is also a two-dimensional predictive coding method that calculates a predicted value by combining image data on multiple lines within the same frame or the immediately preceding frame. Therefore, predictive coding methods are broadly divided into intraframe predictive coding methods that use correlation within a frame and interframe predictive coding methods that use correlation with the preceding frame.

In the intra-frame predictive coding method, as shown in FIG.
2, D3... are created. Furthermore, in a typical interframe predictive coding method, as illustrated in FIG. 5, compressed data Di is created by intraframe predictive coding for the first frame 1, and Regarding v2 to vlO, compressed data D2 to DIO are created by interframe predictive coding using the image data of the immediately previous frame as a reference. Furthermore, for the 11th frame Vll, compressed data Dll is created again by intra-frame predictive coding, and for the nine subsequent frames, inter-frame predictive coding is again performed using the image data of the immediately previous frame as a reference. be exposed. In this way, compression is performed by intra-frame predictive coding at predetermined frame intervals, and data compression is performed for each frame in between by inter-frame predictive coding with the immediately preceding frame.

In a data compression method using interframe predictive coding, in principle, image data of frames at predetermined intervals may be recorded or transmitted as they are without being compressed, but in order to further increase the compression rate, an example is shown in FIG. Intraframe predictive coding is often performed at predetermined frame intervals. In addition, with data compression methods using interframe predictive coding, subsequent compressed data may become impossible to restore due to recording or transmission failures, or the image quality of subsequent frames may deteriorate due to accumulation of errors during compression and restoration. In order to prevent this, uncompressed data or compressed data based on intraframe predictive coding is recorded or transmitted at predetermined frame intervals.

(Problems to be Solved by the Invention) Among the above-mentioned conventional image data compression methods, those using intra-frame predictive coding have problems when the intra-frame correlation is smaller than the inter-frame correlation, for example, if there is small movement within the screen. j
When a lH picture appears, this method has the disadvantage that the data compression rate is lower than that of a method using interframe predictive coding.

On the other hand, in the interframe predictive coding method, when the interframe correlation is smaller than the intraframe correlation, for example, when a simple figure with large movement appears on the screen,
The disadvantage is that the data compression rate is lower than that of intraframe predictive coding. In addition, with this interframe predictive coding method, playback and
When attempting to start displaying, the compressed data must be restored by going back to the frame of uncompressed data that precedes this particular frame, which also has the drawback that it takes time to start displaying.

(Means for Solving the Problems) An apparatus for compressing image data by predictive coding according to the present invention includes a first data compression means for compressing image data of each frame by intraframe predictive coding; a second data compression means for compressing the image data by interframe predictive coding with the image data of the reference frame;
．． The data amount of the compressed data of each frame by each of the second data compression means is compared, and the compressed data with the smaller data amount is selected as the final compressed data of this frame, and the compressed data by the first data compression means is selected as the final compressed data of this frame. means for updating the image data of the reference frame in the second data compression means each time the image data is selected as the final compressed data, and the image data of the reference frame in the second data compression means is updated according to the motion of the screen or the picture pattern. It is configured to increase the data compression rate by dynamically selecting the one with the smaller amount of data.

Hereinafter, the operation of the present invention will be explained in detail together with examples.

(Embodiment) FIG. 1 is a block diagram showing the configuration of an image data compression apparatus using predictive encoding according to an embodiment of the present invention.

In FIG. 1, IN is an input terminal for image data to be compressed, l is an intra-frame encoding circuit, 2 is a quantization circuit, 3 is a variable length encoding circuit, 4 is a compressed data memory, and 5 is an inverse quantization circuit. , 6 is an intra-frame decoding circuit, 7 is a frame memory, 8 is a switch, 9 is a reference frame memory, and 10 is a subtraction circuit.

Furthermore, 11 is a quantization circuit, 12 is a variable length encoding circuit,
13 is a compressed data memory, 15 and 16 are address counters, 17 is a comparison control circuit, 18 is a switch, and OUT is an output terminal for compressed data.

The image data of each frame appearing on the input terminal IN is
In the intraframe encoding circuit l, the signal is converted into an intraframe prediction error signal sequence based on data at a predetermined position within the frame. This intra-frame prediction error signal sequence is quantized by a quantization circuit 2, then converted into a variable-length code sequence by a variable-length encoding circuit 3, and is stored as compressed data by intra-frame prediction encoding in a compressed data memory 4. Stored. On the other hand, the quantized intra-frame prediction error signal sequence output from the quantization circuit 2 is returned to the original intra-frame prediction error signal sequence in the inverse quantization circuit 5, and the intra-frame prediction error signal sequence outputted from the quantization circuit 2 is returned to the original intra-frame prediction error signal sequence in the The image data is added to the data at a predetermined position, thereby being restored to the original image data and stored in the frame memory 7. The data stored in this frame memory 7 is
The signal is transferred to the reference frame memory 9 via the switch 8 which is selectively closed depending on the comparison result of the comparison control circuit 17.

Further, the image data of each frame appearing on the input terminal IN is subtracted by the image data of the reference frame held in the reference frame memory 9 in the subtraction circuit 10, and becomes an interframe prediction error signal sequence. This inter-frame prediction error signal sequence is quantized by a quantization circuit 11, and then converted into a variable-length code sequence by a variable-length encoding circuit 12, and becomes compressed data by inter-frame prediction encoding, and is stored in a compressed data memory 13. is stored in

The comparator circuit 15 compares the count values of the address counters 15 and 16 for each frame, and compares the count values of the address counters 15 and 16.
The amount of compressed data by intra-frame predictive coding stored in 4 and the amount of compressed data by inter-frame predictive coding stored in the compressed data memory 13 are compared, and the one with the smaller amount of data is output as the final compressed data. The switch 18 is controlled so that the signal is output to the terminal OUT. When the comparison control circuit 17 selects the compressed data stored in the compressed data memory 4 by intra-frame predictive encoding as the final compressed data, the comparison control circuit 17 closes the switch 8 to close the frame memory 7.
The image data for one frame currently stored in is transferred to the reference frame memory 9 as reference frame image data for subsequent interframe predictive coding.

FIG. 2 is a conceptual diagram for explaining the operation of the video data compression apparatus shown in FIG. 1. vl, v2, v3...
... is the image data of each frame to be compressed, and Dl
, D2, D3, . . . are the final compressed data of the corresponding frames. Compressed data D1 shown with diagonal lines at the beginning is compressed data by intra-frame predictive coding, and the following compressed data D2, D3...D6 are compressed by inter-frame predictive coding using frame vi as the reference frame. It is data.

Compressed data D2, D3, D by interframe predictive coding
4...The data amount of D6 is based on the reference frame as frame v.
Since it is fixed at 1, the amount of data gradually increases to reflect the accumulation of movements within the screen. In the example in FIG. 2, in frame v7, the amount of compressed data by interframe predictive coding exceeds the amount of compressed data by intraframe predictive coding, and as a result, compression by intraframe predictive coding is indicated by diagonal lines. Data D7 is selected as the final compressed data. This frame v7 is stored in the reference frame memory 9 and becomes a new reference frame for subsequent interframe predictive coding, and subsequent frames V8, V9, etc. are interframe predictive codes based on frame V7. The compressed data D8, D9, . . . are selected as the final compressed data.

Furthermore, at the beginning of each of the final compressed data DI, D2, D3, etc., there is an identifier for identifying whether the data is compressed by intra-frame predictive encoding or by inter-frame predictive encoding. is attached.

FIG. 3 is a block diagram showing an example of the configuration of a restoration device for restoring image data compressed by the compression device shown in FIG. 2 is an inverse quantization circuit, 23 is a switch, 24 is a prediction mode switching circuit, 25 is an intra-frame decoding circuit, 2
6 is a reference frame memory, 27 is an adder circuit, and 0 is an output terminal.

Compressed data D1 of each frame appearing on input terminal I,
D2, D3... are converted into a quantized prediction error signal sequence in the variable length code inverse conversion circuit 21, and
2, it is restored to a prediction error signal sequence before quantization. On the other hand, the prediction mode switching circuit 24 determines whether the compressed data is based on intraframe predictive coding or interframe predictive coding based on the identifier added to the beginning of the compressed data of each frame appearing on input terminal I. In the former case, the switch 23 is switched to the lower side in FIG. 3, and in the latter case, the switch 23 is switched to the upper side in FIG. 3.

In the example of FIG. 2, since the first compressed data D1 is obtained by intraframe predictive encoding, the prediction error signal sequence output from the dequantization circuit 22 is supplied to the intraframe decoding circuit 25 via the switch 23. is restored to image data by adding the reference point in the frame and the prediction error signal sequence,
It is stored in the reference frame memory 26. The video data stored in this reference frame memory 26 is stored in an adding circuit 27.
The data is then supplied as is to output terminal 0, and becomes restored data v1'' in FIG.

When the next compressed data D2 appears, the switch 23 is switched to the upper side in FIG. The sum is added, and the restored moving image data V2' shown in FIG. 2 is supplied to the output terminal O.

Regarding the subsequent compressed data D3, D4, D5, and D6,
As in the case of the compressed data D3, it is added to the restored data Vl' of the reference frame in the adding circuit 27, and the second
Restored video data shown in the figure ■3', V4'', V5'
, V6' and are supplied to output terminal 0.

The next compressed data D7 is generated by intra-frame predictive coding, so it passes through the switch 23, becomes the restored data V7' of the reference frame in the intra-frame decoding circuit 25, passes through the reference frame memory 26 and the adder circuit 27, and is sent to the output terminal O. is supplied to Subsequent compressed data D8, D9...
, restoration is performed based on the restoration data v7° of the reference frame held in the reference frame memory 26.

When the restoration device shown in FIG. 3 is installed in a playback device for image data recorded while being compressed by the compression device shown in FIG. 2, playback and display start from a specific frame, for example, frame ■6°. is assumed to have been ordered. The playback control unit (not shown) first successively outputs the compressed data DI of the nearest reference frame preceding the frame ■6° designated as the playback start frame, restores the reference frame v1°, and stores it in the reference frame memory 26. The frames V6', V7', V3', . . . are restored and displayed while reading out the compressed data D6, D7, D8, . In this way, according to the compression method of the present invention, the intermediate (7
)7tz-mes V2', V3', V4', and ■5'' do not need to be restored, and the time required to display the restored screen is shortened.

(Effects of the Invention) As explained in detail above, the image data compression device using predictive coding according to the present invention includes a first data compression means using intraframe predictive coding and a second data compression means using interframe predictive coding. and a means for selecting the one with the smaller amount of data from each compressed data as the final compressed data. Since the configuration dynamically selects the image data, the compression rate of image data can be effectively increased.

Further, according to the image data compression device according to the present invention,
Specify and play a specific frame of compressed image data,
When starting restoration and display, unlike compressed data using conventional interframe predictive coding, there is no need to reproduce or restore intermediate frames. This also has the effect of shortening the time until display starts.

[Brief explanation of the drawing]

FIG. 1 shows the pressure w1 of image data according to an embodiment of the present invention.
A block diagram showing the configuration of the device, FIG. 2 is a conceptual diagram for explaining the operation of the image data compression device shown in FIG. 1, and FIG. 3 shows how the image data compressed by the compression device shown in FIG. 1 is restored. A block diagram showing an example of the configuration of a restoration device for
FIG. 4 is a conceptual diagram for explaining a data compression method using intra-frame predictive coding, and FIG. 5 is a conceptual diagram for explaining a data compression method using inter-frame predictive coding. IN...Input terminal for image data to be compressed, l...
Intraframe encoding circuit, 2.11... quantization circuit, 3
．． 12... Variable length encoding circuit, 4.13... Compressed data memory, 5... Inverse quantization circuit, 6... Intraframe decoding circuit, 9... Reference frame memory, 15.1
6... Address counter, 17... Comparison control circuit,
OUT... Output terminal for compressed image data, ■...
- Input terminal for compressed data to be restored, 21... variable length code inverse conversion circuit, 22... inverse quantization circuit, 24...
Prediction mode switching circuit, 25... Intraframe decoding circuit, 26... Reference frame memory, 0... Output terminal for restored data, ViV2, v3... Image data to be compressed, Di, D2, D3...Compressed image data, vlo, v2°, v3°...Restored image data. 2nd paper note 3 Figure 41 or 5

Claims (1)

[Scope of Claims] A first data compression means that compresses the image data of each frame by intraframe predictive encoding; and a first data compression means that compresses the image data of each frame by interframe predictive encoding with the image data of a reference frame. Compare the amount of compressed data of each frame by the second data compression means and the first and second data compression means, and select the compressed data with the smaller data amount as the final compressed data of this frame. and means for updating the image data of the reference frame in the second data compression means each time compressed data by the first data compression means is selected as the final compression data. Image data compression device using