JPH06284401A - Picture coder - Google Patents

Abstract

PURPOSE:To control a quantity of a code after coding without causing extreme deterioration in picture quality and limination in frame number of the picture to be coded. CONSTITUTION:Data obtained by converting a quantization matrix corresponding to each frequency component coefficient resulting from picture data subjected to orthogonal transformation at a DC transformation section 1 into an actual quantization step size with a quantization parameter via the DC transformation section 1 and a delay section 6 are quantized by a quantization section 7. Then the coder is provided with a preceding quantization section 2 quantizing the data from the DC transformation section 1 with a predetermined quantization parameter Qc to obtain an optimum quantization parameter, a preceding variable length coding section 3 applying variable length coding to preceding quantization data, a preceding code quantity count section 4 counting the code quantity Bc generated in this case, and a quantization parameter calculation section 5 calculating the quantization parameter based on a relation between a target code quantity Bm and the code quantization Bc.

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 compressing moving images and still images.

[0002]

2. Description of the Related Art As a high-efficiency image coding system, a coding system using orthogonal transformation, quantization and variable length coding is widely used. As the orthogonal transform, generally, a two-dimensional DCT transform of 8 × 8 pixels is often used. DCT
The converted data is decomposed for each spatial frequency component of the image, but a quantization matrix corresponding to each spatial frequency component is prepared in advance. There is a quantization parameter for performing some transformation on this quantization matrix, and the code amount after compression can be controlled by changing this quantization parameter. However, the amount of code after compression greatly depends on the picture pattern of the image.Therefore, when transmitting coded data at a constant rate or recording it in a certain amount of memory, the amount of code must be within the range. It is necessary to control so that the image quality is kept as high as possible.

In the conventional image coding apparatus, feedback control is often used in which the quantization parameter is changed according to the amount of data stored in the transmission buffer.

FIG. 3 shows a conventional image coding apparatus. 11
Is a DCT transform unit for performing DCT transform on an input image divided into blocks of 8 × 8 pixels, 12 is a quantizer for quantizing the data after DCT transform based on a quantization parameter, 13
Is a variable length coding unit for variable length coding the quantized data, 14 is a transmission buffer for the variable length coded data, and the quantization control unit 15 is for storing the amount of data fed back from the transmission buffer 14. The value determines the quantization parameter (change).

[0005]

In the conventional image coding apparatus, feedback control in which the quantization parameter is changed according to the amount of data stored in the transmission buffer 14 is used. In this method, the data storage in the transmission buffer 14 is performed. When the amount becomes large, the quantization parameter is controlled so that the quantization step size becomes coarse in order to suppress the generated code amount. Therefore, the amount of data accumulated in the transmission buffer 14 is extremely smaller than that when the amount is small. There is a problem that the image quality deteriorates.

Further, there is a problem that such feedback control cannot be used when only one frame has to be recorded in a fixed amount of memory like an electronic still camera.

An object of the present invention is to perform image coding in which the code amount after coding can be controlled without causing extreme deterioration in image quality and without being limited by the number of frames of the image to be coded. To provide a device.

[0008]

In order to achieve the above-mentioned object, the present invention is provided by a preceding coding means for preceding-coding image data with a predetermined quantization parameter, and the preceding coding means. Optimum quantization parameter calculation means for calculating the optimum quantization parameter based on the code amount and the target code amount, and quantization for quantizing image data with the optimum quantization parameter obtained by the optimum quantization parameter calculation means. And a means for variable-length coding the data obtained by the quantizing means.

[0009]

According to the present invention, the code amount is predicted to determine the quantization parameter, and the quantization and the variable length coding are performed by this. Therefore, encoding can be performed with stable image quality, control can be performed in frame units, and the image can be used for still image compression.

[0010]

Embodiments of the present invention will now be described with reference to FIGS. FIG. 1 shows an example of the configuration of an image coding apparatus according to the present invention. First, the DCT converter 1
DCT conversion processing is performed on the input image divided into blocks of × 8 pixels. The preceding quantization unit 2 converts the quantization matrix into a quantization step size with a predetermined quantization parameter Qc, and uses this to convert D from the DCT conversion unit 1 into a quantization step size.
Quantize the CT transform coefficients. The preceding variable length coding unit 3 performs variable length coding such as Huffman coding on the data quantized ahead of time by the preceding quantization unit 2. The preceding code amount counting unit 4 counts the code amount Bc of the data coded by the preceding variable length coding unit 3. In the quantization parameter calculation unit 5, the following equation composed of the preset target code amount Bm, the parameters A and B, and the output Bc from the preceding code amount counting unit 4 is used.

[0011]

## EQU00002 ## A quantization parameter Qx that satisfies (Bm / Bc) .Qx ^3/2 + A.Qx + B = 0 is calculated. The delay unit 6 maintains the DC during the calculation of the quantization parameter Qx.
The DCT transform coefficient from the T transform unit 1 is held. The quantizer 7 converts the quantization matrix into a quantization step size using the quantization parameter Qx, and quantizes the DCT transform coefficient with the quantization step size. The variable length coding unit 8 performs variable length coding on the data quantized by the quantizing unit 7 using a Huffman code or the like.

FIG. 2 shows a quantization matrix qm (i,
It is an example in which j) is converted into a quantization step size qs (i, j) by a quantization parameter Qp. The conversion formula is
For example, the following conversion formula is used.

[0013]

## EQU00003 ## qs (i, j) = qm (i, j) .times.Qp / C where C is a constant. In the example shown in FIG. 2, Qp = 1
This is an example when 0 and C = 50.

[0014]

As described above, according to the present invention, the code amount after encoding can be controlled without causing the extreme deterioration of the image quality and without being limited by the number of frames of the image. it can.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration example of an image encoding device of the present invention.

FIG. 2 is a diagram showing an example of a quantization matrix and a quantization step size.

FIG. 3 is a block diagram showing a conventional example.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 DCT conversion unit 2 Preceding quantization unit 3 Preceding variable length encoding unit 4 Preceding code amount counting unit 5 Quantization parameter calculation unit 6 Delay unit 7 Quantization unit

Claims (3)

[Claims] 1. An advance quantization means for performing advance encoding of image data with a predetermined quantization parameter, and an optimum quantization parameter based on a code amount and a target code amount obtained by the preceding encoding means. Optimum quantization parameter calculation means for calculating, quantization means for quantizing image data with the optimum quantization parameter obtained by the optimum quantization parameter calculation means, and variable length code for the data obtained by the quantization means An image coding apparatus, comprising: an encoding unit. 2. The preceding coding means, the preceding quantization unit for performing the preceding quantization of the image data with a predetermined quantization parameter Qc, and the variable-length data for the preceding quantization performed by the preceding quantization unit. It has a preceding variable length coding unit for coding, and a preceding code amount counting unit for counting the code amount Bc generated in the preceding variable length coding unit, and the optimum quantization parameter calculation means has a target code amount Bm. The image coding apparatus according to claim 1, further comprising: a quantization parameter calculation unit that calculates an optimum quantization parameter based on and the code amount Bc. 3. The quantization parameter calculation unit is a ratio (Bm / Bc) between the target code amount Bm and the code amount Bc.
And 2 determined depending on the quantization parameter Qc
Calculating Qx as an optimum quantization parameter that satisfies the following equation (Bm / Bc) · Qx ^3/2 + A · Qx + B = 0, which is composed of two parameters (A and B) The image encoding device according to claim 2.