JPH0132711B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an apparatus for predictively encoding a multilevel image signal, and more particularly to a predictive encoding apparatus for adaptively switching a prediction method according to the characteristics of a screen.

In the prediction of multivalued image signals such as television signals, (1) in addition to the fixed use of one prediction method, (2) the fixed combination of multiple prediction methods, ( 3) Various methods have been considered, such as adaptively switching between multiple predictions depending on the characteristics of the screen.

In this third adaptive prediction method, which performs prediction by adaptively switching the prediction method, blocks are formed in units of multiple pixels, the optimal prediction method is detected for each block, and the optimal prediction method used is determined. One method is to transmit the code indicating the current value and the prediction error at the same time, and the other is to investigate the local properties of the image signal using only the information corresponding to the already encoded pixels and estimate the optimal prediction method for the next pixel. There is a method. In the latter case, similar estimation is possible on the receiving side, so there is no need to transmit a code indicating the prediction method used, and the circuit configuration is that much simpler. This invention belongs to this latter method.

If only the determination result at the current pixel time is used to estimate the optimal prediction method at the next pixel time, a prediction error that is different from that of the surrounding pixels, for example 1 even though the surrounding prediction error is small, is assumed to be used. The estimation of the optimal prediction method may be incorrect due to so-called arcuate points where a large prediction error appears only at points. This phenomenon tends to occur with signals on which random noise is superimposed. Even in such cases, if the occurrence state of the optimal prediction method used is understood two-dimensionally, accurate estimation can be made without being strongly influenced by the existence of isolated points.

This point will be explained with reference to FIG. Here, it is assumed that two types of prediction methods () and () are used. If the time of the currently input pixel signal on the scanning line (line) on the screen is i, then the pixel time on the currently considered line is i-2,
For i-1 and i, the prediction errors by prediction method () are 5, 4, and 3, respectively, as shown in Figure 1A.
Assume that the prediction errors by the prediction method () are 3, 5, and 4, respectively, as shown in FIG. 1B. Therefore, pixel times i-2, i-1,
The optimal prediction method for i is both prediction method (),
Since the prediction error caused by () is smaller, as shown in the current line in FIG. 2, which shows the optimal prediction method for the encoded pixel, respectively.

Since the optimal prediction method at the current pixel time i is (), if only the determination result at the current pixel time is used, the prediction method at the next time (i+1) is (). If the occurrence state of the optimal prediction method at a pixel time one or more pixel times before, one line before, and two lines before is shown in FIG. 2, then (), ( in the vertical direction of the screen at pixel time i+1) ), you can immediately see that () is more likely.

Such a phenomenon in which the optimal prediction method suddenly changes usually occurs at the contour portion. Since the contour part is, so to speak, a part where the brightness is discontinuous, it is necessary to estimate the optimal prediction method in relation to the surroundings. Furthermore, since images typically contain a large number of contours, it is especially important to utilize surrounding relationships.

On the other hand, a figure may suddenly appear to the right from time i+1 on the current line in which a prediction method that cannot be accurately estimated based on the circumstances surrounding the occurrence of the surrounding optimal prediction methods is optimal. In such a case, no prediction will be accurate, so taking into account the optimal prediction method one line before is not disadvantageous compared to a method that uses only the determination result of the current pixel time.

Therefore, in this invention, the estimation accuracy is improved by estimating the optimal prediction method at the next pixel time by considering the spatial distribution state of the optimal prediction method corresponding to the encoded pixel, and in turn, the predictive coding efficiency is improved. In other words, the purpose is to reduce the amount of information transmitted or recorded.

According to this invention, a prediction error signal is created by calculating the difference between an input multilevel image signal and a prediction signal from a selector, and the prediction error signal is encoded and output, and the prediction error signal is encoded and output. The decoded signal is input to a predictor, and the predictor generates a predicted signal using a plurality of prediction methods. The plurality of predicted signals and locally decoded signals are compared in an optimal prediction method determining circuit to rank the optimal prediction methods. Part or all of the rankings are stored in a memory circuit over multiple pixel times, and the prediction method to be used for the next image time is determined according to the spatial distribution of the rankings of multiple lines of the memory contents of the storage circuit. The selection circuit is determined by the circuit, and the selector is controlled by the determination result to select and output the corresponding prediction signal in the predictor. In this way, predictive encoding is performed with high efficiency on multivalued image signals.

Embodiments of the predictive encoding device according to the present invention will be described in detail below with reference to the drawings. Third
The figure shows an example in which two prediction methods are used as the simplest case. A multivalued image signal inputted from an input terminal 11 is supplied to a subtracter 12 . The subtracter 12 receives the predicted signal selected by the selector 14 from among the two types of predicted signals generated by the predictor 13, and calculates the difference between it and the input multilevel image signal supplied from the terminal 11. . This difference, ie, the prediction error, is supplied to a code converter 16 and an adder 17 after being quantized by a quantizer 15, which normally limits the level of use in order to reduce the amount of information generated. The adder 17 generates a locally decoded signal by adding the prediction signal supplied from the selector 14 and the quantized prediction error signal supplied from the quantizer 15. This locally decoded signal is supplied to the predictor 13, and the predictor 13 generates a plurality of predicted signals using this locally decoded signal.

The two types of prediction methods performed by the predictor 13 are a prior value prediction (DPCM) method that uses a local decoded signal from one pixel before, and an interframe prediction method that uses a local decoded signal from one frame before. If used, this predictor 13 is realized by a circuit in which a 1-pixel delay element and a 1-frame delay element are arranged in parallel.

Next, a method of generating a control signal for the selector 14 will be explained. 2 produced by predictor 13
The prediction signal based on the seed prediction method is supplied to the selector 14, and at the same time, is also supplied to the optimum prediction method determining circuit 18. This decision circuit 18 compares and determines which of the two predicted signals is closer to the locally decoded value of the multivalued image signal supplied from the adder 17, that is, which has a smaller difference, that is, at that pixel time. A ranking is made to determine which prediction method is suitable, and the result (in this example, a 1-bit signal) is transferred to the storage circuit 19. The memory circuit 19 is composed of, for example, a memory element of about one line (one horizontal scanning line in the case of a television signal), and its output at pixel time i is the previous line and the previous line in the example shown in FIG. From each time i-1 to i+2 in each line, information indicating a total of 10 optimal prediction methods for times i-1 and i in the current line is used to determine a control signal for the selector 14. Therefore, in this example, a 10-bit output signal is supplied from the memory circuit 19 to the determination circuit 21.

Various determination algorithms can be considered for the determination circuit 21, but here we will use a method that uses the spatial distribution state of information indicating the optimal prediction method of the same type, that is, the distribution state on the screen. In the example shown in FIG. 2, the methods () and () are clearly separated and distributed between pixel times i and i+1.
Therefore, in this case, method () is easily selected as the prediction method at the next pixel time i+1. If the methods () and () are mixed and it is difficult to make a decision, a decision by majority vote may be considered.

This determination in the determination circuit 21 can normally be performed using a read-only memory, so-called ROM. That is, in the above example, a 10-bit signal is supplied to the determination circuit 21, but this is supplied to the input address of the ROM, and various images are pre-programmed at the position corresponding to the bit pattern made up of 10 bits. Write down the results of statistical judgment. This will be explained using the example shown in FIG. The upper 4 bits of the input 10-bit signal are from time i-1 to i+2 of the previous line.
, the following middle 4 bits are a signal indicating the optimal prediction method from time i-1 to i+2 of the previous line, and the last lower 2 bits are a signal indicating each optimal prediction method for times i-1 and i of the current line, and method () If () is represented by the code 0 and the method () by the code 1, then at the current time i, a 10-bit bit pattern of 0011001100 from the upper bit to the lower bit is obtained from FIG. Judgment circuit 21
Now, use this as the input address and enter the corresponding address.
In this case, it is sufficient to write the code 1 representing the method (). This kind of thing is determined and written in advance for all addresses. Therefore, in actual judgment, the most convenient one can be selected and used without making any particular distinction between majority voting, judgment based on spatial distribution pattern, etc. Judgment circuit 21
The selection of the prediction method by the selector 14 is controlled by the output signal of . Note that the determination result in the determination circuit 21 is 1.
Output with pixel delay.

The prediction error signal supplied to the code converter 16 uses an unequal length code to further increase the information reduction effect, and when transmitting or recording on a storage medium, necessary control signals are added to adjust the transmission speed or recording speed. output after speed matching.

Next, a decoding device for decoding information encoded in this manner will be described in detail with reference to FIG. The encoded multilevel image signal transmitted or read from the recording medium is first supplied to the code inverse converter 23 through the terminal 22 . The code inverse converter 23 extracts a prediction error signal from the encoded multivalued image signal, performs inverse conversion from an unequal length code to an equal length code, and prepares for the following decoding. The prediction error signal inversely converted into an equal length code by the code inverse converter 23 is supplied to the adder 24. The adder 24 receives the prediction signal selected by the selector 26 from among the two types of prediction signals generated by the predictor 25, and combines this prediction error signal with the code inverse converter 2.
A decoded multivalued image signal is obtained by addition with the output of No. 3. This decoded multivalued image signal is simultaneously supplied to an external image signal output terminal 27, a predictor 25, and an optimal prediction method determining circuit 28, respectively.

Using this decoded image signal, the predictor 25 generates a plurality of predicted signals. The optimal prediction method determination circuit 28 uses this decoded image signal and a plurality of prediction signals (two types in this example) from the predictor 25 to perform the determination circuit 18 of the encoding device described with reference to FIG. The optimum prediction method is determined according to the same rules, and the result is transferred to the storage circuit 29.
Similar to the storage circuit 19 in the encoding device shown in FIG.
supply to The determination circuit 31 controls the selection of the selector 26 using the same rules as the determination circuit 21 of the encoding device. Note that the predictor 25, the selector 26, the optimal prediction method determining circuit 28, the memory circuit 29, and the determining circuit 3
1 is the predictor 1 in the encoding device.
3, selector 14, optimum prediction method determining circuit 18,
It is the same as the memory circuit 19 and the determination circuit 21, and the connection relationship between them is also the same.

Next, a case will be described in which three or more types of prediction methods are used. If we consider n types of prediction methods (n≧3), selectors (14 in the encoding device and 26 in the decoding device, hereinafter referred to as 14 and 26) and optimal prediction method determining circuits 18 and 28 n types of prediction signals are simultaneously supplied to the input terminal. This decision circuit 18, 2
As the output of 8, the prediction methods are ranked in descending order of the prediction error, and information indicating the prediction methods of only the top several types is transferred to the storage circuits 19 and 29 and stored therein. The memory circuits 19 and 29 use memory elements of 1 bit/pixel in the embodiment described above with reference to FIG.
If we use a memory element with a bit length long enough to express They are lined up.

If the output signals of the memory circuits 19 and 29 at this time use the same information (10 types) of the optimal prediction method for pixel time as in the previous embodiment, the total number is 10 x (bit length that can express n) x m bits, and the determination circuits 21 and 31 determine the prediction method at the next pixel time by majority vote with an appropriate weight applied depending on the ranking, for example.

As yet another embodiment, a case will be described in which the input multilevel image signal is a color television signal having a color signal subcarrier. For example, in an NTSC color television signal, the phase of the color signal subcarrier is inverted line by line, and since one frame consists of an odd number of lines (525 lines), the phase is also inverted frame by frame. Therefore, prediction signals whose color signal subcarriers do not match in phase have very large prediction errors and are not suitable as prediction signals.
Therefore, 2 lines before the phase matches, 1 field before (equal to 262 lines before), 1 frame ±1
If the first embodiment is executed using a prediction signal from a line before, two frames before, etc., the present invention can be applied to an NTSC color television signal, and its effects are significant.

Furthermore, as another modification, a case may be considered in which the optimal prediction method at pixel time i of the current line is excluded from consideration when determining the prediction method at the next pixel time i+1. In this case, since the occurrence state of the surrounding optimal prediction method is used, even if the optimal prediction method at pixel time i is not considered, the high coding efficiency is hardly reduced, and the allowable calculation time is significantly increased. Therefore, it can be constructed using ordinary semiconductor integrated circuits such as transistor-transistor logic (TTL).

[Brief explanation of drawings]

FIG. 1 is a diagram showing an example of prediction errors by two types of prediction methods, FIG. 2 is a diagram showing an example of spatial distribution of the optimal prediction method, and FIG. 3 is a diagram showing an example of the predictive coding device according to the present invention. FIG. 4 is a block diagram showing an embodiment of a decoding device for decoding information encoded by the encoding device according to the present invention. 11, 22: input terminal, 12: subtractor, 13,
25: Predictor, 14, 26: Selector, 16: Code converter, 17, 24: Adder, 18, 28: Optimal prediction method determining circuit, 19, 29: Storage circuit, 2
1, 31: Judgment circuit, 23: Sign inverse converter, 2
7: Output terminal.

Claims (1)

[Claims] 1. In predictive coding of a multivalued image signal, means for generating a plurality of prediction signals for the same signal using different prediction methods, a selection means for selecting one of the plurality of prediction signals, and an output of the selection means an encoding means for encoding a prediction error between a prediction signal and the multi-valued image signal; a decoding means for locally decoding the multi-valued image signal from the output of the encoding means and the output of the selection means; a determining means for comparing the decoded multi-level image signal and the plurality of prediction signals and ranking the suitability of the prediction methods corresponding to the plurality of prediction signals; and a part or all of the rankings. a storage means for storing the above over a plurality of pixel times, and a prediction method to be used at the next pixel time using the spatial distribution state of the ranks of the plurality of lines stored in the storage means, and the selection means An adaptive predictive coding device for an image signal, comprising a control means for controlling the image signal.