JPS5915553B2 - Predictive coding device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a predictive encoding device used when encoding binary image signals such as facsimile signals.

Conventional predictive coding methods of this type include one-dimensional prediction or two-dimensional prediction (IEICE Journal Vol. 56-D).
, /f63, PP, 170-177, March 1973) is used alone throughout the paper.

However, when looking at a page in units of a certain number of pixels, for example, in units of scanning lines, even within the same page, there are areas where the correlation in the sub-scanning direction is strong, such as the center of characters or areas with vertical lines. ,
Two-dimensional prediction has fewer prediction errors, but in areas where the correlation in the sub-scanning direction is weak, such as the transition from a blank space to the beginning of a character q or the transition area from the end of a character to a blank space, , the degree of correlation in the sub-scanning direction differs for each scanning line, so when performing one-dimensional or two-dimensional prediction alone, the prediction error is smaller when one-dimensional prediction is performed. Since it does not appear sufficiently, it has the disadvantage that the amount of information after encoding cannot be reduced. In addition, when two-dimensional prediction is used alone, the amount of information after encoding is generally reduced compared to when one-dimensional prediction is used, but if there is a code error in the transmission path, However, the disadvantage is that code errors are propagated. An object of the present invention is to provide a predictive encoding device that eliminates the drawbacks of the conventional predictive encoding methods described above.

The predictive encoding device of the present invention is an encoding device that predictively encodes a binary image signal in units of a certain number of pixels, and includes a plurality of sub-measurement means that perform a plurality of predictions in units of the certain number of pixels. , means for comparing prediction effects of a plurality of predictions made by the plurality of secondary measurement means, means for selecting any one prediction method from the comparison results, and encoding a prediction error of the selected prediction method. It consists of the means to

Next, the present invention will be explained in detail with reference to the drawings.

FIG. 1 is a diagram showing the first embodiment of the present invention, in which two types of predictions are performed, the prediction error of each prediction is encoded, the number of coded bits after encoding is compared, and the number of coded bits is The configuration of a predictive encoding device that selects the smaller one is shown. In FIG. 1, the binary image signal 1 (for example, 2 scanning lines) once stored in the tsufua memory 2 is as follows.
The control signal of the switching signal generator 10 reads out the buffer memory 2 and controls the first predictor 3 (e.g. the second predictor 3).
It realizes the prediction function shown in the figure, and can be easily realized by a combination of an exclusive OR gate and a shift register), and is converted into a prediction error signal by the multiplexer 5.
The encoder 6 (for example, a conventionally used encoder such as a run-length encoder) encodes the encoder 6 via the encoder 6 (for example, a conventionally used encoder such as a run-length encoder or the like).
The bits encoded in this manner are counted by a counter 7 and input to a comparator 9. Next, in a similar operation, the image signal 1 read out from the buffer memory 2 by the control signal of the switching signal generator 10 is passed through the second predictor 4, the encoder 6 and the counter 8, and the encoded bits are counted. and input to the comparator 9. The comparator 9 selects the output of the counter 7 and the counter 8 which has the smaller number of encoded bits, and supplies it to the switching signal generator 10. The switching signal generator 10 receives the output signal from the comparator 9 and reads out the image signal 1 from the buffer memory 2 again. This read image signal 1 is encoded in the encoder 6 via either the predictor 3 or 4 selected by the control signal of the switching signal generator 10, and is passed through which predictor. After the mode code is added by the encoder 6, it is passed through the gate 11 to the output terminal 12.
is output from. In this example, even if the number of predictors is three or more, a similar configuration can be adopted, and the amount of information after encoding is of course reduced. The same thing applies to ℃. As mentioned above, by using two predictors 3 and 4, the prediction is made not for the entire page but for a certain number of pixels (two scanning lines). Since a prediction method with a small amount of information after encoding is selected, the predictive encoding device has a small amount of information after encoding compared to the conventional one-dimensional or two-dimensional prediction of only one type. FIG. 2 is a diagram showing an example of a prediction function used in predictors 3 and 4 in FIGS. 1, 3, and 4.
In Fig. 2, the reference letter A is the prediction function F1-d at ℃
In the case of one-dimensional prediction, the signal e is predicted from the previous signal d in the same scanning line, and if it matches d, the sign of 0 is
Indicates that if there is no match, a code of 1 is given. Reference alphabet B is for two-dimensional prediction of prediction function F2-b+d,
This shows that the signal e is predicted from the signal b of the previous scanning line and the signal d of the same scanning line. The reference alphabet C indicates that the signal e is predicted from the signals A, b, and c of the previous scanning line and the signal d of the same scanning line in the case of two-dimensional prediction of the prediction function F3-d(1+b+c)+Ibc. . FIG. 3 is a diagram showing a second embodiment of the present invention, in which two types of prediction methods are used, the number of bits of misprediction of each prediction error signal is compared, and the one with the smaller number of bits is selected. This figure shows the configuration of a predictive encoding device in this case.

In the figure, the binary image signal 1 (corresponding to two scanning lines) once stored in the buffer memory 2 is stored in the buffer memory 2.
The first predictor 3 and second predictor 4 convert the signal into a prediction error signal, and the counter 7 and counter 8 count the number of mispredicted bits, respectively. The numbers of mispredicted bits counted by the counters 7 and 8 are compared by a comparator 9, and the one with the smaller number of bits is selected and input to the switching signal generator 10. The switching signal generator 10 receives the output signal from the comparator 9 and reads out the image signal 1 from the buffer memory 2 again. The image signal 1 read out in this way is converted into a prediction error signal by either the selected predictor 3 or 4, and then sent to the encoder 6 under a control signal from the switching signal generator 10. It is then encoded. After this, encoder 6
The coded bits encoded in
is output from. In the prediction device of the embodiment shown in FIG. 3, two predictors 3 and 4 are used, so that the prediction is performed not for the entire page but for a fixed number of pixels (two scanning lines). Since the number of mispredicted bits of the prediction error signal is compared, the control system is simpler than the device of the first embodiment, and it has almost the same effect. FIG. 4 is a diagram showing a third embodiment of the present invention, in which two types of prediction methods are used:
Using dimensional prediction and two-dimensional prediction, encode the prediction error of each prediction method, compare the difference in the number of coded bits after encoding, and determine if the difference is less than a certain value or if the two-dimensional prediction is n-line scanning line. This figure shows the configuration of a predictive encoding device in the case where one-dimensional prediction is selected in consideration of the influence of code errors on the transmission path. The device shown in FIG. 4 has the same configuration as that shown in FIG. 1, except that a counter 13 and a line number monitoring circuit 14 are added. Fourth
The illustrated apparatus operates differently from the apparatus of FIG. 1, as described below. That is, for every two scanning lines, the comparator 9 outputs a signal for selecting the one-dimensional prediction predictor 3 if the difference in the number of encoded bits between the counter 7 and the counter 8 is within a certain value. The counter 13 counts the number of times the predictor 4 for two-dimensional prediction is selected consecutively. The output of this counter 13 is monitored by a line number monitoring circuit 14, and when the number of lines exceeds n lines, the one-dimensional prediction predictor 3
A signal for selecting is output to the switching signal generator 10. It goes without saying that the technical idea described in this embodiment can be applied to the configuration of the second embodiment shown in FIG. 3 by adding a counter and a line number monitoring circuit.

In the predictive encoding device shown in the embodiment of FIG. 4, the difference in the number of coded bits after encoding is compared by using two predictors, that is, a one-dimensional predictor and a two-dimensional predictor. , when the difference between them is less than a certain value or when two-dimensional prediction continues for n lines, one-dimensional prediction is selected, which has the effect of being less susceptible to code errors on the transmission path. The predictors used in the three examples above have no prediction, that is,
Of course, there are cases where the original image signal is input as is.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing a first embodiment of the present invention, and FIG.
The figure is a diagram for explaining the prediction operation of the predictor used in the present invention, FIG. 3 is a block diagram showing a second embodiment of the present invention, and FIG. 4 is a block diagram showing a third embodiment of the present invention. It is a block diagram. In Figures 1, 3 and 4, reference numeral 2 is a buffer memory, reference numerals 3 and 4 are predictors, reference numeral 5 is
is a multiplexer, reference numeral 6 is an encoder, reference numerals 7 and 8 are counters, reference numeral 9 is a comparator, reference numeral 10
11 is a switching signal generator, reference numeral 11 is a gate, reference numeral 13 is a counter and reference numeral 14 is a line number monitoring circuit.

Claims (1)

[Claims] 1. A coding device that predictively encodes a binary image signal in units of a certain number of pixels, including a plurality of prediction means that perform a plurality of predictions in units of the certain number of pixels, and a plurality of prediction units performed by the plurality of prediction means. The method is characterized by comprising: means for comparing the prediction effects of the predictions; means for selecting any one prediction method from the comparison results; and means for encoding the prediction error of the selected prediction method. predictive coding device.