JPH0654304A - Picture encoder - Google Patents

Abstract

PURPOSE:To obtain stable picture quality with high control precision in quantization by preventing the local variation, etc., of the quality in a compressed picture. CONSTITUTION:When an input picture Si is inputted, a total amount arithmetic means 21 obtains a total amount S21 and imparts it to a quantization control means 23. An encoding amount calculating means 22 calculates an encoding amount S22 from the output encoding data S13 of an encoding means 13 and imparts it to a control means 23. The control means 23 controls a quantizing means 12 so as to permit the encoding amount of the picture to be a prescribed value and also to permit the quantization quality of the applying picture not to change in the internal part of the picture. Also the control function is updated by the encoding amount S22. The quantizing means 12 quantizes picture data S11 from which the redundunce is removed by an information amount compressing means 11. The quantization output data S12 are encoded by the encoding means 13 so as to be outputted to an external part with a buffer 14.

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 for coding a moving image or the like, and more particularly to a control system of image quality and code amount.

[0002]

2. Description of the Related Art Conventionally, as a technique in such a field,
For example, some documents were described in the following documents. IEEE TRANSACTIONS ON COMMUNICATIONS
), COM-32 [3] (1984-3) (US) Wen-H
siung CHEN and William K. PRATT “Scene Adaptive Coder” P.P. 225-
In 232 image compression, particularly in real-time compression of moving images, it is necessary to control the code amount of images due to restrictions such as transmission channels. As the conventional code amount control method, the buffer control method described in the above document is most often used. The configuration example will be described below with reference to the drawings.

FIG. 2 is a block diagram showing the configuration of a buffer control type image coding apparatus described in the above document.
This image coding apparatus includes an information amount compression means 11 that removes redundancy of an input image S _i and outputs image data S11. The information amount compression means 11 is composed of, for example, a discrete cosine transform (DCT) means, a prediction (DPCM) means, etc., and the output side thereof has a quantization means 12 and an encoding means 1.
3 is connected. The quantization means 12 uses the quantization step size S15 to eliminate redundant image data S.
11 is quantized, and the quantized output data S12 is given to the encoding means 13, which is constituted by a Midread type linear quantizer or the like. The encoding means 13 encodes the quantized output data S12 and outputs the encoded data S13, and a variable length code (Huffman code or the like) is often used. A buffer 14 is connected to the output side of the encoding means 13, and the buffer 14 is feedback-connected to the quantizing means 12 via the buffer control means 15.

The buffer 14 temporarily stores the encoded code data S13 and outputs the output image S _o at a predetermined timing.
And the function of outputting the occupancy S14 (the code amount of the image accumulated in the buffer 14).
It is composed of IFO (First In First Out). Generally, the input of the buffer 14 is irregular and the speed (bit rate) is indefinite. Buffer control means 15
Measures the occupancy S14 of the buffer 14 at regular intervals, and calculates the next quantization step width S15 from the occupancy S14.
And has a function of giving it to the quantizing means 12.

Next, the operation of FIG. 2 will be described with reference to FIGS. 3 and 4. FIG. 3 is a diagram for explaining the Midread type linear quantizer, where x on the horizontal axis is input, y on the vertical axis is output, and h is the quantization step width. FIG. 4 shows buffer 1
It is a figure which shows the method of determining the quantization step width S15 from the occupation amount S14 of 4. B of FIG. 4 shows a buffer occupancy amount, B _max is an upper limit buffer occupancy amount, and B _min is a lower limit buffer occupancy amount. h represents a quantization step width, h _max represents a maximum step width, and h _min represents a minimum step width.

When the input image S _i is input to the information amount compression means 11, the information amount compression means 11 removes the redundancy of the input image S _i and sends the image data S 11 to the quantization means 12. The quantizing means 12 quantizes the image data S11 based on the quantizing step width S15 given from the buffer controlling means 15 and sends the quantized output data S12 to the coding means 13. If the quantizing means 12 is constituted by, for example, the Midread type linear quantizer shown in FIG. 3, the quantization can be expressed by the following equation (1).

[0007]

[Equation 1] The encoding means 13 is quantized quantized output data S1.
2 is encoded and the encoded data S13 is sent to the buffer 14. The buffer 14 stores the temporarily stored code data S13.
Are output to the transmission channel at a constant speed (bit rate) according to the transmission channel. In the conventional buffer control type image encoding device, the buffer 1 provided at the final stage
The occupancy S14 of 4 is measured by the buffer control means 15, and the buffer control means 15 controls the quantization means 12 by changing the characteristic of the quantization means 12, for example, the quantization step width S15, according to the occupancy quantity S14. , The occupancy S14 of the buffer 14 is controlled so as not to exceed a predetermined upper limit or lower limit. That is, the buffer control unit 15 measures the occupancy S14 of the buffer 14 at regular intervals, and the occupancy S14 is measured.
Then, the next quantization step width S15 is determined and given to the quantization means 12. Occupancy S1 of this buffer 14
4, the method of determining the quantization step size S15 is shown in FIG. As shown in FIG. 4, for the buffer occupancy S14 = B measured at a certain time, the next quantization step width S15 = h is determined by the following equation (2).

[0008]

[Equation 2]

[0009]

However, the apparatus having the above structure has the following problems. (A) In the conventional image coding apparatus, the quantization step width S15 of the quantization means 12 is controlled by the occupied amount S14 of the buffer 14 regardless of the image quality. Therefore, in the same image, a part with good image quality and a part with poor image quality are mixed, and the numerical evaluation value (for example, the signal-to-noise ratio (S
/ N ratio)) The deterioration of the image quality is felt. (B) In a moving image, there is a large variation in image quality between frames, and similarly, the image quality deteriorates more than the numerical evaluation value. (C) FIGS. 5 (a) and 5 (b) are explanatory views of conventional problems showing the above phenomenon.

FIG. 5A is a diagram showing the relationship between the passage of time and the buffer occupancy B. As shown in this figure, the code amount by the encoding means 13 changes due to the variation in the local characteristics of the image, and as a result, the buffer occupation amount B changes. FIG. 5B is a diagram showing the relationship between the passage of time and the image quality evaluation value (eg, S / N ratio). As shown in this figure, if the buffer occupancy B changes, the quantization step width S output from the buffer control means 15 accordingly.
Since 15 changes, the S / N ratio also changes. As a result, S /
The N ratio varies, and the image quality deteriorates. The present invention provides an image encoding device that solves the problem that the above-mentioned conventional technique has, such as local variations in quality of compressed images.

[0011]

In order to solve the above-mentioned problems, the first invention is to quantize the information amount compression means for removing the redundancy of the input image and the image data redundantly removed by the information amount compression means. An image coding apparatus comprising: a quantizing means for coding, a coding means for coding the quantized output data of the quantizing means, and a buffer for temporarily storing the image data coded by the coding means. In, a statistical quantity calculating means for calculating the statistical properties of the input image and outputting the statistical quantity of the image, a code quantity calculating means for calculating the code quantity of the image based on the output of the coding means, and a quantizer And control means. The quantizing control means controls the quantizing means so that the code amount of the image becomes a predetermined code amount based on the statistical amount of the image, and the control is performed from the calculated code amount of the image. It has an update function to update the function. According to a second aspect of the invention, in an image coding apparatus comprising the information amount compression means, the quantization means, the coding means, and the buffer of the first invention, the statistical properties of the input image are calculated and the statistics of the image are calculated. Quantity calculation means for outputting the quantity, code quantity calculation means for calculating the code quantity of the image based on the output of the coding means, and quantization error power of the image from the quantized output data of the quantization means. Quantization error calculation means and quantization control means are provided. The quantizing control means controls the quantizing means so that the quantizing quality or code amount of the image becomes a predetermined value based on the statistical amount of the image, and the calculated code amount of the image. And the update function for updating the control function from the quantization error power. According to a third aspect of the present invention, the control function of the second aspect of the present invention controls the quantizing means based on the code amount of the image when the code amount of the image is determined and determines the quantization quality of the image. Is configured to control the quantization means based on the quantization quality.

[0012]

According to the first aspect of the present invention, since the image coding apparatus is configured as described above, when the input image is input to the statistical amount calculating means and the information amount compressing means, the statistical amount calculating means:
Statistics such as correlation and variance of the input image are calculated and sent to the quantization control means. The information amount compression means sends the image data from which the redundancy between the pixels of the input image or the redundancy between the current image and the previous image is removed to the quantization means. In the quantizer,
The image data from which the redundancy is removed is quantized based on the control data such as the quantization step width given from the quantization control means, and the quantized output data is sent to the encoding means. The encoding means encodes the quantized output data and sends the encoded data to the buffer. The buffer temporarily stores the code data and outputs the code data to the outside at a predetermined timing. The code amount calculation means calculates the code amount for each image, for example, based on the output of the coding means, and sends the code amount to the quantization control means.
In the control function of the quantization control means, control data such as a quantization step width is calculated based on the statistics of the image so that the quality of the image becomes constant, and the quantization means is controlled by the control data. At the same time, the quantization control means also predicts, for example, the code amount necessary for encoding the image based on the output of the code amount calculating means, and updates the control function using the ratio of the code amount and the actually measured code amount. To do. According to the second and third inventions, the information amount compressing means, the quantizing means, the encoding means, the buffer, the statistic calculating means, and the code amount calculating means perform the same operations as in the first invention. The quantizing error calculating means calculates, for example, the quantizing error power for each pixel from the output of the quantizing means, and sends the quantizing error power to the quantizing control means. The control function of the quantization control means controls the quantization step width and the like so that the image quality such as the quantization error power or the code amount becomes a predetermined value based on the image statistics from the statistics calculating means. Data is calculated and the quantizing means is controlled by the control data. At the same time, the quantization control means, for example, by comparing the given predetermined code amount or the calculated code amount necessary for encoding the image with the code amount input from the code amount calculating means, Update the control function. Further, in the quantization control means, for example, by comparing the actual quantization error power measured by the quantization error calculation means with a predetermined quantization error power or the calculated quantization error power of the image, Update the control function. Therefore, the above problem can be solved.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS First Embodiment FIG. 1 is a block diagram showing the configuration of an image coding apparatus according to a first embodiment of the present invention, in which elements common to those shown in FIG. Is attached. This image encoding device removes redundancy of the input image S _i by using means such as orthogonal transform (for example, DCT) and outputs the image data S11 from which redundancy has been removed, as in the conventional case.
1, the quantizer 12, the encoder 13, and the buffer 14 are connected to the output side thereof. The quantizing means 12 quantizes the image data S11 using the given control data (for example, quantizing step width) S23 and outputs the quantized output data S12 to the encoding means 13. The encoding means 13 has a function of encoding the quantized output data S12 and outputting the encoded data S13 to the buffer 14. The buffer 14 has code data S
13 has a function of temporarily storing 13 and outputting the output image S ₀ encoded at a constant bit rate according to the FIFO rule to the outside. This image coding apparatus is different from the conventional one in that the statistical amount calculating means 21 for inputting the input image S _i , the code amount calculating means 22 connected to the output side of the coding means 13, and the code This is the point where the quantity calculation means 22 and the quantization control means 23 connected to the output side of the statistic calculation means 21 are provided.

The statistic calculation means 21 obtains the statistical properties (for example, the average, variance, correlation, etc.) of the input one input image S _i , and the quantized control means 23 calculates these statistics S21. It has a function to output to. The code amount calculation unit 22 outputs the image data S output from the encoding unit 13.
13, the code amount S22 of one image or several blocks is calculated, and the code amount S22 is calculated by the quantization control means 23.
It has a function to output to. The quantization control means 23
Based on the given image statistic S21, the average quantization error power of the image is calculated so that the code amount of the image becomes a predetermined code amount, and the quantization step is calculated from the calculated average quantization error power. It has a control function of determining the width and controlling the quantizing means 12 by the control data S23, and at the same time,
It has an update function for updating the quantization error power calculation function from the ratio of the given code amount S22 of a given image or several blocks to a predetermined code amount.

Next, the operation of FIG. 1 will be described. Input image S
_{When i} is input, the statistic calculator 21 uses the following equations (3) to (6) for the input M x N input image S _i = g (x, y), for example. , Its average value μ, variance σ ² , horizontal correlation coefficient ρ _H , and vertical correlation coefficient ρ _v .

[0016]

[Equation 3] Then, those statistics S21 = (σ ² , ρ _H , ρ _v )
Is output to the quantization control means 23. Information amount compression means 11
Then, for example, the image is divided into small blocks n × n, the discrete cosine transform (DCT) is performed on the image, and the correlation is removed. DC
At T, an input image S _i = g _b (x, y) of a block
On the other hand, the output image data S11 = y _b (u, v) is _expressed by the following equations (7) to (9).

[0017]

[Equation 4] The quantizing means 12 uses the quantizing step width Δ given from the quantizing controlling means 23 to input the input image data S1.
1 = y _b (u, v) is quantized by the equation (1) in FIG. 3 and the output level is output to the encoding means 13 as the quantization result.

[0018]

[Equation 5] The encoding means 13 is quantized quantized output data S1.
For example, the Huffman code or the zero-run composite Huffman code is assigned to 2 and coded, and the code data S1
3 is output to the buffer 14. The buffer 14 is composed of, for example, a FIFO memory, while the code data S1 is used.
Input 3 at any time, and temporarily store. On the other hand, the accumulated data is output at a constant speed (bit rate) on a first-come-first-served basis. The code amount calculation means 22 inputs the code data S13, cumulatively adds the code amount in image units or several blocks, and the added code amount S22 is quantized by the quantization control means 2.
Output to 3. The quantization control means 23 first calculates the energy of the compressed output data by the information amount compression means 11 of the normalized image input data using the given correlation coefficients ρ _H and ρ _v . For example, when a block DCT is used as the information amount compression unit 11, the quantization control unit 23 has the energy (dispersion) σ _N ² (u, u of each component of the normalized image input data (σ ² = 1) after DCT. v) is calculated by the following equation (11). Here, the quantization error power σ of each component after DCT
_e ² (u, v) is expressed as the following equation (12), and (1
From the equation (3), the average code amount b (bits / pix) in pixel units
el) is calculated.

[0019]

[Equation 6] Then, the average code amount b is a predetermined code amount b ₀ (bits /
pixel) and so as to determine the sigma _em ^2, further sigma _em ²
Is used to obtain the quantization step width Δ (= S23) as in the equation (14), and the quantization step width Δ is given to the quantization means 12.

[0020]

[Equation 7] At the same time, the quantization control means 23 compares the code amount b ₁ (bits / pixel) calculated by the code amount calculating means 22 with a predetermined code amount b _0, and sets the coefficient k as shown in equation (15). Calculate and update sequentially.

[0021]

[Equation 8] As described above, in the image coding apparatus according to the present embodiment, based on the statistical property of the input image S _i , the code amount of the image becomes a predetermined value, and the quantization quality of the image is the inside of the image. Since the quantization control means 23 controls so that there is no change, stable image quality can be obtained and the quantization can be easily controlled. Further, since the actual code amount of the image is measured and the quantization control means 23 is updated so that the actually measured code amount becomes a predetermined code amount, it is possible to always realize control with a small error adapted to the image.

Second Embodiment FIG. 6 is a block diagram showing the arrangement of an image coding apparatus according to the second embodiment of the present invention. Elements common to those in FIG. 1 showing the first embodiment are shown in FIG. Are assigned common reference numerals. In this image coding apparatus, instead of the quantizing control means 23 of FIG. 1, a quantizing control means 23A having a different configuration is provided. Further, the quantizing error calculating means 24 is connected to the output side of the quantizing means 12, and the output side of the quantizing error calculating means 24 is connected to the quantizing control means 23A. Other configurations are the same as those in FIG. The quantizing error calculating means 24 receives the quantizing output data S12 from the quantizing means 12, calculates the quantizing error power S24 of one image or several blocks, and calculates the quantizing error power S24.
It has the function of outputting to 3A. Quantization control means 23
A determines the quantization step width so that the quality (for example, quantization error power) or code amount of the image becomes a predetermined value based on the image statistical amount S21 output from the statistical amount calculation means 21. , Has a control function of controlling the quantizing means 12 by the control data S23A, and at the same time, compares the given measured code amount S22 and quantization error power S24 with a predetermined value to update the control function. It has an update function.

Next, the operation of FIG. 6 will be described. Input image S
_{When i} is input, the information amount compression means 11 and the quantization means 1
2, encoding means 13, buffer 14, statistic calculation means 2
1 and the code amount calculation means 22 perform almost the same operations as in the first embodiment. When the quantized output data S12 output from the quantizer 12 is supplied to the quantizer error calculator 24, the quantizer error calculator 24 calculates the quantized output data S12 (= The root mean square of E (i) is taken in image units or in units of several blocks, and the quantization error power (= quantization error root mean) S24 (= σ _E
² ) is obtained and output to the quantization control means 23A.

[0024]

[Equation 9] In the quantization control means 23A, first, the statistic calculation means 21.
Correlation coefficients ρ _H and ρ of the statistic S21 given by
_{Using v} , the information amount compression means 1 of the normalized image input data
The energy of the compressed output data according to 1 is calculated. For example, when a block DCT is used as the information amount compression means 11, the quantization control means 23A causes the energy (dispersion) σ _N ² (u, u) of each component of the normalized image input data (σ ² = 1) after DCT. v) is calculated from the above equation (11).
Here, the quantization error power σ of each component after DCT
Expressing _e ² (u, v) as shown in the above equation (12), and given a predetermined average quantization error power σ _e0 ² , σ _em ² was obtained so as to obtain the following equation (17). After that, the code amount b ₀ is obtained from the equation (18).

[0025]

[Equation 10] When a predetermined code amount b ₀ is given, σ _em ² is obtained so as to obtain the equation (17), and then the quantization error power S24 (= σ _e0 ² ) is obtained from the equation (17). Then, the quantization step width Δ (= S23A) is obtained from the following equation (19) and is output to the quantization means 12.

[0026]

[Equation 11] At the same time, the quantization control means 23A causes the code amount calculation means 22
And the code amount b ₁ (bits / pixel) calculated in
The predetermined code amount is compared with each other, and the equation (18) is sequentially updated so that the coefficient k _n is as in the following equation (20).

[0027]

[Equation 12] Further, the quantization control means 23A has the quantization error power S24 (= σ _e1 ² ) calculated by the quantization error calculation means 24.
And σ _e0 ² are compared, and the equation (19) is sequentially updated so that the coefficient l _n becomes the following equation (21).

[0028]

[Equation 13] As described above, in the image coding apparatus of the present embodiment, based on the statistical property of the input image S _i , the code amount of the image becomes a predetermined value or the quality of the image (quantization error power). ) Is a predetermined value and the quantizing control means 23A controls the quantizing means 12 so that the local quantizing quality of the image is uniform, so that a flexible and stable image is obtained. Quality is obtained. That is, when the demand for the image quality is strict, the quality of the image can be intensively controlled, whereas when the demand for the code amount of the image is severe, the code amount of the image can be intensively controlled. Furthermore, since the code amount and the quantization error power of the image are measured and the quantization control means 23A is updated, the control always adapted to the image can be realized. The present invention is not limited to the above embodiment,
For example, each of the constituent blocks in FIG. 1 may be configured by an individual circuit using an integrated circuit or the like, or may be configured by program control using a computer. Further, in order to improve the compression coding accuracy, it is possible to add other functional blocks or the like to the device of FIG.

[0029]

As described in detail above, according to the first aspect of the present invention, based on the statistical property of the input image, the code amount of the image becomes a predetermined value and the quantization of the image is performed. As the quality is controlled by the quantization control means so that it does not change inside the image, stable image quality is obtained and
Quantization can be easily controlled. Further, for example, if the actual code amount of the image is measured and the quantization control means is updated so that the actually measured code amount becomes a predetermined code amount, control with a small error always adapted to the image can be realized.

According to the second and third inventions, the code amount of the input image is set to a predetermined value or the quality (quantization error power) of the image is set based on the statistical property of the input image. Since the quantizing means is controlled so that the value becomes, and there is no unevenness in the local quantizing quality of the image, a flexible and stable image quality can be obtained. That is, when the demand for the image quality is strict, the quality of the image can be intensively controlled, whereas when the demand for the code amount of the image is severe, the code amount of the image can be intensively controlled. Further, since the code amount and the quantization error power of the image are actually measured and the quantization control means is updated, the control always adapted to the image can be realized.

[Brief description of drawings]

FIG. 1 is a configuration block diagram of an image coding apparatus showing a first embodiment of the present invention.

FIG. 2 is a configuration block diagram showing a conventional image encoding device.

FIG. 3 is a diagram for explaining a Midread type linear quantizer.

FIG. 4 is a diagram showing a conventional method of determining a quantization step width.

FIG. 5 is an explanatory diagram of a conventional problem.

FIG. 6 is a configuration block diagram of an image coding apparatus showing a second embodiment of the present invention.

[Explanation of symbols]

 11 Information amount compression means 12 Quantization means 13 Coding means 14 Buffer 21 Statistics amount calculation means 22 Code amount calculation means 23, 23A Quantization control means 24 Quantization error calculation means

Front page continuation (72) Inventor Yoko Harada 1-7-12 Toranomon, Minato-ku, Tokyo Oki Electric Industry Co., Ltd.

Claims (3)

[Claims] 1. An information amount compression means for removing redundancy of an input image, a quantization means for quantizing the image data of which redundancy has been removed by the information amount compression means, and a quantization output data of the quantization means. In an image coding apparatus including a coding unit for coding and a buffer for temporarily storing the image data coded by the coding unit, a statistical property of the image is calculated by calculating statistical properties of the input image. And a code amount calculation unit that calculates the code amount of the image based on the output of the encoding unit, and the code amount of the image becomes a predetermined code amount based on the statistical amount of the image. A control function for controlling the quantizing means,
And a quantization control means having an updating function for updating the control function from the calculated code amount of the image, the image coding device. 2. An information amount compression means for removing redundancy of an input image, a quantization means for quantizing image data of which redundancy has been removed by the information amount compression means, and a quantization output data of the quantization means. In an image coding apparatus including a coding unit for coding and a buffer for temporarily storing the image data coded by the coding unit, a statistical property of the image is calculated by calculating statistical properties of the input image. , A code amount calculation unit that calculates the code amount of the image based on the output of the encoding unit, and a quantum that calculates the quantization error power of the image from the quantized output data of the quantization unit. A quantization error calculation means, a control function for controlling the quantization means so that the quantization quality or code amount of the image becomes a predetermined value based on the statistical amount of the image, and the code of the calculated image From the quantization error power, the image coding apparatus characterized by a quantization control means, provided with an update function for updating the control functions. 3. The image coding process according to claim 2, wherein the control function controls the quantizing means based on the code amount of the image when the code amount of the image is determined, and the quantization quality of the image. The image coding apparatus is characterized in that the quantizing means is controlled based on the quantizing quality when is determined.