JPH09289643A - Image encoding device and method therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a device capable of reducing a generation code amount by providing a state value generation means, the number of times of state generation, a prediction correction value generation means, a correction prediction value generation means, a prediction error generation means and an entropy encoding means. SOLUTION: In a counter circuit 108, the numger N(S) of times of generation and a cumulative error E(S) are taken out corresponding to a state number outputted from a context generation circuit 103 and they are delivered to a prediction correction value generation circuit 105 and a counter updating circuit 109. In the prediction correction value generation circuit 105, a prediction correction value C is obtained by the arithmetic operation of E(S)/N (S). Also, a subtraction circuit 107 obtains a prediction error (e). A selector 112 compares the three values of the prediction error (e) and fixed values Th and-<=h, ranks them by the sizes, selects the second largest value and outputs it to a signal line 113. The counter updating circuit 109 adds an input value from the signal line 113 to the E (S).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image coding apparatus and method for coding an image using a prediction error based on peripheral pixel values.

[0002]

2. Description of the Related Art Conventionally, there is a predictive coding method as one of the image coding methods used in an image coding apparatus. In this predictive coding method, the pixel value of the pixel of interest is predicted from the surrounding pixels, and the prediction error is entropy coded. In this case, some prediction methods are prepared, and these are adaptively selected and used, or when entropy-encoding, divided into several states and entropy-encoded for each state. Various improved schemes based on the scheme have been proposed. It should be noted that even in JPEG, which is the international standard encoding method for natural images, the predictive encoding method is adopted as the lossless encoding method.

As one of the methods for improving the coding efficiency in this predictive coding method, a method of dividing the state and correcting the predicted value by using the average value of the prediction error in each state (hereinafter,
This is called error feedback). In this method, for example, when the pixel value of the pixel of interest is I, the state around the pixel of interest is Si, and the predicted value is p, the state Si has been changed so far.
The number N (Si) of occurrences of the error and the cumulative addition E (Si) of the prediction error in the state Si are held, and the average prediction error Em
= E (Si) / N (Si) is obtained, the corrected prediction value p'is obtained from p '= p + Em, and the prediction error e = Ip' is encoded to improve the prediction accuracy.

[0004]

However, in the image coding apparatus using the above-described conventional error-feedback predictive coding method, for example, when a large prediction error suddenly occurs, the value fed back thereafter becomes large. Error feedback may not work effectively because it is affected. Therefore, the amount of generated code becomes large, there is a problem that it takes a long time to transmit an image, and a large capacity is required to store an image.

The present invention has been made in view of the above-mentioned problems, and an object of the present invention is to make an error feedback effectively act even when a large prediction error suddenly occurs and to reduce the generated code amount. An object of the present invention is to provide an image encoding device and method that can reduce the number of images.

[0006]

In order to achieve the above object, the present invention provides an image coding apparatus for coding an image based on a predicted value obtained from pixel values of pixels around a pixel of interest. From means for generating a state value for the surrounding pixels, the number of occurrences of the state for each state of the generated state value, and a value obtained by accumulating a prediction error generated in the state with a predetermined limit. Means for generating a predicted correction value,
Means for generating a modified predicted value obtained by modifying the predicted value using the predicted modified value; means for generating a predicted error from the modified predicted value and the pixel value of the pixel of interest; and entropy coding of the predicted error. And means for doing so.

[0007]

Embodiments of the present invention will be described below in detail with reference to the accompanying drawings. <First Embodiment> FIG. 1 is a block diagram showing the arrangement of an image coding apparatus according to the first embodiment of the present invention. In the figure, 101 is a signal line for signal input, 1
Reference numeral 02 is a buffer for storing image data (image data series) for two lines, 103 is a context generation circuit for generating context from peripheral pixels of the pixel of interest, 104
Is a prediction circuit that generates a prediction value from surrounding pixels, 105
Is a prediction correction value generation circuit that generates a prediction correction value.
Further, 106 is an addition circuit, 107 is a subtraction circuit, and 108.
Is a counter circuit for storing the number of state occurrences and cumulative error for each context, 109 is a counter updating circuit, 110
Is a Huffman coding circuit, 111 is a signal line for outputting a coded output signal, 112 is a selector, and 11
Reference numeral 3 is a signal line for outputting an output signal from the selector.

In the present embodiment, a case where a 16-value (4-bit) monochrome image is encoded will be described as an example.

The Huffman coding circuit 110 shown in FIG. 1 is preliminarily constructed with a Huffman table generated based on statistics of prediction errors in general images. Further, all the data held by the counter circuit 108 is initialized to "0".

First, from the outside of the device to the signal line 101,
Pixels to be encoded are input in raster scan order. The buffer 102 stores a signal input from the signal line 101 for two lines. In addition, the context generation circuit 103 receives from the buffer 102 the peripheral pixels a,
Take out b and c.

Incidentally, FIG. 2 shows the positions of the peripheral pixels a, b and c with respect to the pixel of interest x.

Next, in the context generation circuit 103,
Using these pixels, the state number S is obtained by S = a × 256 + b × 16 + c (1) and output. The counter circuit 108 extracts the number of occurrences N (S) and the cumulative error E (S) according to the state number S output from the context generation circuit 103, and outputs them to the prediction correction value generation circuit 105 and the counter update circuit 109. hand over. The prediction correction value generation circuit 105 obtains the prediction correction value C by calculating E (S) / N (S). However, when N (S) = 0, C = 0.

The prediction circuit 104 also includes a buffer 102.
The peripheral pixels a, b, and c are taken out from and the predicted value p is generated from the following prediction formula.

P = a + b + c (2) The addition circuit 106 adds the C output from the prediction correction value generation circuit 105 to the prediction value p output from the prediction circuit 104 to generate a correction prediction value p ′. To do. Further, the subtraction circuit 106 obtains the prediction error e by e = x-p '... (3). The selector 112 compares the prediction error e with the three values of the constant values Th and −Th, ranks them according to their magnitude, selects the second largest value, and outputs it to the signal line 113. .

The counter updating circuit 109 has a signal line 113.
The input value from is added to E (S). Further, the counter updating circuit 109 increments the value of N (S) and compares it with a predetermined value R. If N (S) = R, E (S) and N (S) are compared. Cut it in half. Then, the N (S) and E (S) thus updated are returned to the counter circuit 108.

The Huffman encoding circuit 110 refers to the Huffman table corresponding to the above-mentioned state number S, Huffman-encodes the prediction error e, and obtains the obtained code on the signal line 111.
Output to In the device according to the present embodiment, the above-described processing is repeated up to the last pixel input from the signal line 101 to perform predetermined encoding.

As described above, according to the present embodiment, the obtained prediction error is compared with a total of three values of two constant values,
By ranking and adding the second largest value to the cumulative error and performing feedback based on the updated cumulative error and the number of state occurrences, effective error feedback can be realized and the prediction accuracy is improved. It is possible to perform coding with a small amount of generated code. <Second Embodiment> FIG. 3 is a block diagram showing the arrangement of an image coding apparatus according to the second embodiment of the present invention. In the figure, 301 is a signal line for signal input, 3
02 is a buffer for storing image data of 2 lines,
Reference numeral 303 is a context generation circuit that generates a context from the peripheral pixel value of the pixel of interest, 304 is a prediction circuit that predicts the value of the pixel of interest from the peripheral pixel value, 305 is a prediction correction value generation circuit, 306 is an addition circuit, and 307 is Subtraction circuit, 3
08 is a counter circuit, 309 is a counter updating circuit, 31
0 is a Huffman coding circuit, 311 is a signal line for an output signal, 312 and 313 are multiplication circuits, and 314 and 31.
Reference numeral 5 is a selector.

In this case, 8 bits (0 to 255)
Will be described as an example of encoding a monochrome image of (value).

The Huffman encoding circuit 310 of FIG. 3 has a Huffman table generated on the basis of statistics of a general image prediction error, as in the apparatus according to the first embodiment shown in FIG. Build. Further, all the data held by the counter circuit 308 is initialized to "0".

First, the pixel value x to be encoded is input to the signal line 301 from the outside of the apparatus in the raster scan order. The buffer 302 stores the signal input from the signal line 301 for two lines, and the context generation circuit 303
Extracts the peripheral pixels a, b, c, d, and e of the pixel of interest from the buffer 302 and obtains the average value m of a to e. Then, a to e are binarized by using the average value m as a threshold value, and the context S is generated by arranging in order from the MSB.

In this binarization, the value is the average value m.
If it is smaller than "0", it is "1" if the value is equal to or larger than m. The state number S thus obtained is an integer value of 0 to 31 represented by 5 bits.

FIG. 4 is a diagram showing the positions of the surrounding pixels a, b, c, d, and e with respect to the pixel of interest x in the present embodiment.

In the counter circuit 308 of FIG. 3, the number of occurrences N (S) and the cumulative error E (S) are taken out in accordance with the state number S output from the context generating circuit 303, and the predicted correction value is generated. It is passed to the circuit 305 and the counter updating circuit 309. In the predicted correction value generation circuit 305, E
The predicted correction value C is obtained by (S) / N (S). However, when N (S) = 0, C = 0.

The prediction circuit 304 also includes a buffer 302.
The surrounding pixels a and b are taken out from and the prediction value p is generated from the following prediction formula.

P = (a + b) / 2 (4) In the addition circuit 306, p output from the prediction circuit 304
To the corrected correction value generation circuit 105 to generate a corrected prediction value p ′. Also, the subtraction circuit 307
The prediction error e is obtained by e = x-p '(5). In addition, the selector 315 is
When the value of C output from the predicted value correction circuit 305 is "0", it is "1", and when it is not "0", it is c.
Is output.

The multiplication circuit 312 multiplies the output value of the selector 315 by n (n is an arbitrary integer). In addition, the multiplication circuit 31
3 inverts the sign of the multiplication result of the multiplication circuit 312. Then, the selector 314 includes the multiplication circuit 312 and the multiplication circuit 3.
The output from 13 and the three values of e output from the subtraction circuit 307 are ranked according to the magnitude of the value to calculate the second value.

On the other hand, the counter updating circuit 309 adds the output value of the selector 314 to E (S) and increments N (S). Next, N (S) is compared with a predetermined value R, and if N (S) = R, E (S) and N
(S) is halved. Then, the N (S) and E (S) thus updated are returned to the count circuit 308.

The Huffman coding circuit 310 is described above.
The prediction error e is Huffman-coded by referring to the Huffman table corresponding to the state number S, and the obtained code is converted into the signal line 3
11 is output. In this way, the above processing is performed by the signal line 3
The image coding is repeated until the last pixel input to 01.

As described above, also in this embodiment, the cumulative error and the number of state occurrences are updated based on the value obtained by multiplying the output from the predicted value correction circuit, and the feedback control is performed. Effective error feedback can be realized.

The present invention is not limited to the above-described embodiment, and for example, as the prediction method of the pixel value of interest, the previous value prediction may be simply used, or some prediction methods may be used. You may prepare and switch it at a suitable time.

In the above embodiment, Huffman coding is used as a means for entropy coding.
Other entropy coding schemes such as arithmetic coding may be used. Even when the present invention is applied to a system including a plurality of devices (for example, a host computer, an interface device, a reader, a printer, etc.), it can be applied to an apparatus including one device (for example, a copying machine, a facsimile device, etc.). You may apply.

Further, an object of the present invention is to supply a storage medium having a program code of software for realizing the functions of the above-described embodiment to a system or apparatus, and to supply a computer (or CP) of the system or apparatus.
U and MPU) read out and execute the program code stored in the storage medium.

In this case, the program code itself read from the storage medium realizes the function of the above-described embodiment, and the storage medium storing the program code constitutes the present invention. As a storage medium for supplying the program code, for example, a floppy disk, a hard disk, an optical disk, a magneto-optical disk,
A CD-ROM, CD-R, magnetic tape, non-volatile memory card, ROM or the like can be used.

Moreover, not only the functions of the above-described embodiments are realized by executing the program code read by the computer, but also the OS (operating system) running on the computer based on the instructions of the program code. It goes without saying that this also includes the case where the above) performs a part or all of the actual processing and the processing realizes the functions of the above-described embodiments.

Further, after the program code read from the storage medium is written in the memory provided in the function expansion board inserted into the computer or the function expansion unit connected to the computer, based on the instruction of the program code. , The CPU provided in the function expansion board or the function expansion unit performs some or all of the actual processing,
It goes without saying that the processing may realize the functions of the above-described embodiments.

[0036]

As described above, according to the present invention,
The prediction value is corrected by using the prediction correction value obtained from the number of occurrences of the state value generated for the surrounding pixels and the value obtained by cumulatively adding a predetermined limit to the prediction error that has occurred in the state. Even if a large prediction error occurs, effective error feedback can be realized and the image coding efficiency can be improved.

[0037]

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an image encoding device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing a positional relationship between a pixel of interest and peripheral pixels in the first embodiment.

FIG. 3 is a block diagram showing a configuration of an apparatus according to a second embodiment.

FIG. 4 is a diagram showing a positional relationship between a pixel of interest and peripheral pixels according to the second embodiment.

[Explanation of symbols]

 101, 111, 113, 301, 311 Signal line 102, 302 Buffer 103, 303 Context generation circuit 104, 304 Prediction circuit 105, 305 Prediction correction value generation circuit 106, 306 Addition circuit 107, 307 Subtraction circuit 108, 308 Counter circuit 109 , 309 Counter updating circuit 110, 310 Huffman coding circuit 112, 314, 315 Selector 312, 313 Multiplier

Claims (8)

[Claims] 1. An image coding apparatus for coding an image based on a predicted value obtained from pixel values of surrounding pixels of a pixel of interest, said means for generating a state value for said surrounding pixels, and said generating means. Means for generating a prediction correction value from the number of occurrences of the state for each state of the state value and a value obtained by cumulatively adding a predetermined limit to the prediction error generated in the state, and using the prediction correction value A means for generating a modified prediction value obtained by modifying the prediction value; a means for generating a prediction error from the modified prediction value and the pixel value of the pixel of interest; and means for entropy coding the prediction error. A characteristic image encoding device. 2. The image coding apparatus according to claim 1, wherein the predetermined limit is a limit of cumulative addition of the generated prediction error according to a preset upper limit value and a lower limit value. 3. The image coding apparatus according to claim 2, wherein the prediction correction value is used as the upper limit value and the lower limit value. 4. The image coding apparatus according to claim 1, wherein the entropy coding includes at least Huffman coding and arithmetic coding. 5. An image encoding method for encoding an image based on a prediction value obtained from pixel values of pixels surrounding a pixel of interest, the step of generating a state value for the surrounding pixels, Generating a predicted correction value from the number of occurrences of the state for each state of the state value and a value obtained by cumulatively adding a predetermined limit to the prediction error generated in the state, and using the predicted correction value A step of generating a modified predicted value in which the predicted value is modified, a step of generating a predicted error from the modified predicted value and a pixel value of the pixel of interest, and a step of entropy coding the predicted error. Characteristic image coding method. 6. A prediction error generating unit that generates a prediction error using a prediction value and a pixel value to be encoded, a control unit that controls the value of the prediction error, and a prediction error that is controlled by the control unit. And a correcting means for correcting the predicted value. 7. The image coding apparatus according to claim 6, wherein the correction by the correction unit is performed based on a plurality of prediction errors controlled by the control unit. 8. A step of generating a prediction error using the prediction value and a pixel value to be encoded, and a step of controlling the value of the prediction error and correcting the prediction value based on the controlled prediction error. An image coding method comprising: