JP2743037B2 - Encoding method and encoding device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an encoding method and an encoding apparatus for dividing a digital image signal into sub-bands into a plurality of frequency bands, and blocking and encoding the divided image signals in each band.

[0002]

2. Description of the Related Art
As a method for encoding a digital image signal, Japanese Patent Laid-Open No.
No. 253382 and U.S. Pat. No. 4,394,774. Hereinafter, description will be given using these publications as examples.

FIG. 1 is a block diagram showing a configuration of a conventional digital image signal encoding apparatus disclosed in Japanese Patent Application Laid-Open No. 1-253382. The main motion evaluation circuit 101 defines, for each image, a main displacement vector that minimizes the image-to-image difference with respect to the preceding image, and performs a scan conversion circuit 102 and an encoding circuit 10
5 and are applied. The scan conversion circuit 102 forms a two-dimensional block of 8 pixels × 8 lines, and forms a three-dimensional block from four two-dimensional blocks of four continuous images. In each 3D block, 4 2D blocks of 8 pixels × 8 lines are spatially shifted from one image to the next by the displacement vector for each image output from the main motion evaluation circuit 101. Have been. The three-dimensional block output from the scan conversion circuit 102 is orthogonally transformed in the three-dimensional orthogonal transformation circuit 103. The quantization / normalization circuit 104 includes a three-dimensional orthogonal transformation circuit 10
The normalization and quantization of the transform coefficient output from 3 are realized. The normalization operation weights the transform coefficients by multiplying or dividing the transform coefficients by a parameter associated with the filling rate in the speed control memory 106 and the transform coefficients themselves. Quantization converts the normalized value of each transform coefficient represented by a floating point to an integer value.

The output of the quantization and normalization circuit 104 is applied to an encoding circuit 105, which encodes a word between a predefined set of words, indicated as Huffman codes and stored in memory. And another word representing the address of the encoded value, for each non-zero quantum value, to the rate control memory 106. This address is defined with the one-dimensional scan of the three-dimensional block by encoding the length of the sequence of zero values. Such an encoding circuit 105 having a variable length encoding function is a common type. The speed control memory 106 ensures that the output speed is constant. And U.S. Patent Application No. 4,39
No. 4,774 discloses a method for reducing the speed of encoding the sequence of zero values from the speed control memory 106.

FIG. 2 shows another embodiment of the encoding circuit 105, as shown in FIG. 2, which includes m parallel paths each of which receives all the values to be encoded. May be provided. More specifically, these values are applied along a path to each memory circuit, which has two memories that themselves function as flip-flops. For a given three-dimensional block, the value is written to one of these two memories, while the value of the previous block is read in the other memory in an order corresponding to the new scan. FIGS. 3A and 3B show two scanning types of this reading operation. The type shown in FIG. 3 (a) allows the speed of blocks with fixed content to be minimized, while the type shown in FIG. 3 (b) relates to moving picture parts. The output of each of the memory-forming circuits 111a-111m is applied to a combination circuit 112a-112m, which is a coding circuit that implements variable length coding in the manner described in U.S. Pat. No. 4,394,774. Each output of the circuits 112a to 112m is applied to a counting / selecting circuit 113. The circuit counts the number of bits used to encode the block for each scan for each 3D block, and then determines the scan that minimizes the speed of each 3D block, The output of the branch circuit 114 for storing bits from the circuit is controlled, and the transfer of the bit corresponding to the best encoding of the three-dimensional block from the branch circuit 114 to the multiplexing circuit 115 is guaranteed. The multiplexing circuit 115 multiplexes these bits according to the selected scanning index, which must be transmitted to the decoder for reconstruction of the block, and at the start of each group of four pictures. The transmission of the principal displacement vector that determines

In the conventional encoding method of a digital image signal, when one-dimensionally scans the transform coefficients after the three-dimensional orthogonal transform, the stationary portion is of the scanning type shown in FIG. On the other hand, by selecting the scanning type shown in FIG. 3B, the code amount is reduced. However, when subband division and three-dimensional orthogonal transformation are combined, that is, when a digital image signal is divided into a plurality of frequency bands, and then three-dimensional orthogonal transformation is applied to each band component, especially when a high-frequency For components containing
The scanning type shown in FIGS. 3A and 3B has a problem that it is not the best scanning type.

The quantization level of the transform coefficient output from the three-dimensional orthogonal transform circuit 103 is determined by the filling speed of the speed control memory 106 and parameters related to the transform coefficient itself. However, when such quantization is performed, for example, when a flat portion and an edge portion are mixed in one block, there is a problem that noise in the flat portion is very noticeable in the restored image.

In the above-mentioned Japanese Patent Application Laid-Open No. 1-253382, if an available signal is interlacedly scanned,
Before encoding, the image signal is converted into a progressively scanned image signal. Therefore, as shown in FIG. 4, one image is composed in frame units, and a three-dimensional image is obtained by taking a first dimension in the horizontal direction, a second dimension in the vertical direction, and a third dimension in the time direction. And reduces the redundant components of the video signal by performing orthogonal transformation.

By the way, an actual television screen adopts an interlaced scanning method as shown in FIG. This method is a method of preventing flicker without increasing the amount of transmission information for transmitting moving image information. Accordingly, scanning for one screen is completed with half the number of scanning lines in FIG. In the next screen, the line that has not been scanned in the immediately preceding screen is scanned, thereby suppressing a decrease in the vertical resolution of the image. By the interlaced scanning, the number of screens transmitted in the same time becomes twice that in the sequential scanning, so that the occurrence of flicker is suppressed. This roughly scanned screen is called a field, and one frame is composed of two continuous fields as shown in FIG.
The scanning speed is about 60 fields per second in the ational Television System Committee) system.

A conventional digital image signal encoding method is as follows.
Since the three-dimensional block is formed from the image signals of the progressive scanning method, it is not always possible to effectively reduce the redundancy of the image signals of the interlaced scanning method. In particular, if an image signal of the interlaced scanning method having a large motion is subjected to the same encoding as that of the image signal of the progressive scanning method, a two-dimensional block in which the spatial displacement and the temporal displacement are mixed is formed. However, there is a problem that the redundancy of the image signal does not decrease so much.

FIG. 7 shows, for example, IEEE Transactions on Con
sumer Electronics, Vol.34, No.3 (AUGUSUT, 1988) “AN EXPERIMENTAL DIGITAL VCR WITH 40MM DRUM, SINGL
E ACTUATOR AND DCT-BASED BIT-RATE REDUCTION "is a block diagram showing a configuration of a conventional encoding apparatus. In the figure, reference numeral 121 denotes a block which divides an input digital image signal into a plurality of blocks. Circuit
The blocking circuit 121 outputs the image signal of each block to the DCT circuit 122. The DCT circuit 122 includes a blocking circuit 121
DCT for each block of the image signal output from
(Discrete Cosine Transform) and outputs the transform coefficient to the weighting circuit 123. After weighting each transform coefficient, the weighting circuit 123 outputs the weighted transform coefficient to the adaptive quantization circuit 124. The adaptive quantization circuit 124 has a plurality of quantization tables with different quantization step widths, quantizes the weighted transform coefficients with the optimal quantization step width, and sends the quantized transform coefficients to the variable length coding circuit 125. Output. The variable length coding circuit 125 performs variable length coding on the quantized transform coefficients, and outputs the variable length coded data to the buffer memory 126. The buffer memory 126 converts the variable-length coded data into a fixed rate, stores it, and outputs it at a fixed output rate. The buffer controller 127
In order to prevent the buffer memory 126 from overflowing, the quantization step width in the adaptive quantization circuit 124 is switched and the transform coefficient to be encoded in the variable length encoding circuit 125 is selected.

Next, a specific operation will be described. The input digital image signal is composed of, for example, a luminance signal and two color difference signals.
In 1, after being time-divided, the signal is divided into blocks of, for example, 8 pixels × 8 lines and output to the DCT circuit 122.
The DCT circuit 122 subjects the image signal of each block to 8-point discrete cosine transform in the horizontal and vertical directions. First, an image signal is represented by x (i, j) (i, j = 0,1,1).
, 7), a horizontal 8-point DCT is performed by the following equation.

[0013]

(Equation 1)

The converted image signal f (0, j), f
(M, j) is subjected to a vertical 8-point DCT according to the following equation, and the image signal is transformed into a transform coefficient F (m, n) (m, n =
0, 1,..., 7), and the weighting circuit 12
Output to 3.

[0015]

(Equation 2)

Each transform coefficient input to the weighting circuit 123 is subjected to weighting. Specifically, taking advantage of the fact that human vision is dull against high spatial frequencies,
As shown in FIG. 8, a small weighting is performed on a region including a high spatial frequency component, and a large weighting is performed on a region including a low spatial frequency component. Here, the weighting coefficient W (m, n) is represented by the following equation.

[0017]

(Equation 3)

The output of the weighting circuit 123 is quantized by an adaptive quantization circuit 124. An optimal quantization step width is selected in the adaptive quantization circuit 124 based on the transform coefficient in each block and the quantization parameter from the buffer controller 127, and the weighted transform coefficient is determined by the optimal quantization step width. Are quantized.
More specifically, a coarse quantization step width is selected when the video data is a high-contrast rising portion, and a fine quantization step width is selected when the video data is a small-amplitude detail portion.

The quantized transform coefficients are subjected to variable-length encoding in a variable-length encoding circuit 125, and then to a buffer memory.
126. The amount of data stored in the buffer memory 126 is detected by the buffer controller 127 so that the buffer memory 126 does not overflow. The buffer controller 127 determines a quantization parameter according to the amount of data stored in the buffer memory 126, and based on the quantization parameter, an adaptive quantization circuit
In addition to switching the quantization step width at 124, the transform coefficient to be encoded by the variable length encoding circuit 125 is selected according to the data amount. That is, the buffer controller 127
If the amount of data stored in the buffer memory 126 is large, the data compression ratio is increased, and if the amount of data is small, the data compression ratio is adjusted to be lower. Preventing. The data stored in buffer memory 126 is read at a fixed output rate.

In such an encoding device, when a digital image signal is divided into sub-bands and orthogonal transform is performed on blocks of each sub-band, the frequency characteristics of each sub-band are different due to the influence of aliasing of sub-sampling. Therefore, appropriate weighting for each subband is required.

As a method of reducing the coding rate of a digital image signal, a method of periodically thinning out a sampled image signal, that is, a method using subsampling, as disclosed in Japanese Patent Application Laid-Open No. 63-38385, is known. is there. FIGS. 9 and 10 show a transmitting side (recording side) and a receiving side (reproducing side) of an encoding apparatus in which such a method is applied to a color image signal.
FIG. 3 is a block diagram showing the configuration of FIG.

First, the transmitting side will be described with reference to FIG. For example, an NTSC color image signal is input to the input terminal 131. This color image signal is supplied to an A / D converter 132
The digital color image signal in which one sample is quantized to 8 bits at a sampling frequency of, for example, 4 fsc (fsc: color subcarrier frequency) is obtained from the A / D converter 132. The digital color image signal is output to the sub-sampling circuit 133, and the output signal of the sub-sampling circuit 133 is input to the blocking circuit 134. A pre-filter for band limitation is not provided at the previous stage of the sub-sampling circuit 133, so that the high frequency component of the input color image signal is not lost.

In the sub-sampling circuit 133, the digital color image signal is sampled at a sampling frequency of 2 fsc. Further, the input digital color image signal is converted by the blocking circuit 134 into a continuous signal for each two-dimensional block which is a unit of encoding. In this example, one screen is obtained by dividing the screen of one field.
The size of the block is 8 pixels × 4 lines = 32 pixels. FIG. 11 shows one block. In FIG. 11, a solid line indicates a line of an odd field, and a broken line indicates a line of an even field. Note that unlike this example, for example, a three-dimensional block including four two-dimensional regions belonging to each of four frames may be one block. The sub-sampling circuit 133 provided before the blocking circuit 134 allows pixels in the block to be mapped.
As shown in FIG. 12, the pixels are thinned out, and the number of pixels in one block becomes 16 pixels. In FIG. 12, a circle indicates a sub-sampled pixel, and a cross indicates a thinned pixel.

The output signal of the blocking circuit 134 is input to a dynamic range (DR) detection circuit 135 and a delay circuit 136. The DR detection circuit 135 detects the dynamic range DR and the minimum value MIN for each block. The pixel data PD from the delay circuit 136 is output to the subtractor 137, and the subtractor 137
, Pixel data PDI from which the minimum value MIN has been removed is formed.

The quantizing circuit 138 supplies the pixel data P after the sub-sampling and the removal of the minimum value through the subtractor 137.
DI and dynamic range DR are input. Quantization circuit
In 138, the pixel data is adapted to the dynamic range DR.
PDI quantization is performed. From the quantization circuit 138, a code signal DT obtained by converting one pixel data into four bits is obtained.

The code signal DT from the quantization circuit 138 is output to the framing circuit 139. Frame circuit 139
, A dynamic range DR (8 bits) and a minimum value MIN (8 bits) are input as additional codes for each block. The framing circuit 139 performs an error correction coding process on the code signal DT and the above-mentioned additional code, and adds a synchronization signal. Transmission data is obtained at an output terminal 140 of the framing circuit 139, and the transmission data is transmitted to a transmission line such as a digital line. In the case of a digital VTR,
The output signal is sent to a rotary head via a recording amplifier, a rotary transformer, and the like.

Next, the receiving side will be described with reference to FIG. The received data from the input terminal 141 is input to the frame decomposition circuit 142. By the frame decomposition circuit 142,
The code signal DT and the additional codes DR and MIN are separated and error correction processing is performed. The code signal DT and the dynamic range DR are input to the decoding circuit 143.

The decoding circuit 143 includes a quantization circuit 13 on the transmission side.
Performs the reverse of the process of 8. That is, the data after the removal of the 8-bit minimum level is decoded to the representative level, this data is added to the 8-bit minimum value MIN by the adder 144, and the original pixel data is decoded. The output data of the adder 144 is output to the block decomposition circuit 145. The block decomposition circuit 145 converts the decoded data in the order of the blocks into the same order as the scanning of the television signal, contrary to the blocking circuit 134 on the transmission side. The output signal of the block decomposition circuit 145 is output to the interpolation circuit 146. In the interpolation circuit 146, the data of the thinned pixel is interpolated by the surrounding sub-sampling data. A digital color image signal having a sampling frequency of 4 fsc from the interpolation circuit 146 is converted into a D / A converter.
Output to 147. An analog color image signal is obtained at an output terminal 148 of the D / A converter 147. If a pre-filter is not provided on the transmitting side, aliasing distortion may occur at, for example, a steep change in luminance level. Therefore, a circuit for removing the distortion is provided by an interpolation circuit 146.
May be connected to the output side.

In the coding apparatus configured as described above, the coding rate is reduced, but the resolution of the moving image is degraded, and aliasing distortion occurs, and the image quality is largely deteriorated. Is not enough.

FIGS. 13 and 14 show, for example, IEEE Transactions
on Circuits and Systems, Vol.35, No.2.February 1988
`` Sub-Band Coding of Monochrome and Color Image
FIG. 18 is a block diagram illustrating a configuration of a conventional subband division / combination circuit indicated by “s”.

The digital image signal input from the input terminal 151 is input to a horizontal low-pass filter 152 (hereinafter, referred to as a horizontal LPF 152) that limits the horizontal frequency of the image signal. The horizontal LPF 152 is an even-tap filter having a frequency characteristic as shown in FIG. 15A, and its coefficient is represented by h ₁ (n) (n
= 0,..., N−1; N is an even number), and h ₁ (n) = h ₁ (Nn−1), n = 0,. That is, the input image signal is 256 pixels x 25
Assuming six lines, the horizontal LPF 152 outputs a signal represented by the following equation (1) for each line x (n) (n = 1,..., 256) of the image signal.

[0032]

(Equation 4)

On the other hand, the input digital image signal is
Horizontal high-pass filter to limit horizontal frequency of image signal
153 (hereinafter referred to as horizontal HPF 153). Horizontal
The HPF 153 is an even-tap filter having a frequency characteristic as shown in FIG. 15B, and its coefficient h ₂ (n) (n = 0,..., N−1)
Is h ₂ (n) = h ₁ (n) · (−1) ⁿ . Therefore, h ₂ (n) = − h ₂ (Nn−1), n = 0,..., (N / 2) −1. That is, the horizontal HPF 153 outputs a signal represented by the following equation (2) for each line x (n) (n = 1,..., 256) of the image signal.

[0034]

(Equation 5)

The outputs of the horizontal LPF 152 and the horizontal HPF 153 are input to horizontal 2: 1 sub-sampling circuits 154a and 154b for thinning out the pixels in the horizontal direction by 2: 1 so that the number of pixels in the horizontal direction is halved. Thinned out. The outputs of the horizontal 2: 1 sub-sampling circuits 154a and 154b are output to vertical low-pass filters 155a and 155b (hereinafter, vertical LPFs 15) that limit the vertical frequency.
5a, 155b). Vertical LPF 155a, 155b
Is an even-tap filter having a frequency characteristic as shown in FIG. 15A, and if its coefficient is h ₃ (m) (m = 0,..., M-1; M is an even number), h ₃ (m ) = H ₃ (Mm−1), m = 0,..., (M / 2) −1. That is, the vertical LPFs 155a and 155b are
For each column w (m) (m = 1,..., 256) of the image signal output from the horizontal 2: 1 sub-sampling circuit 154a, 154b, a signal represented by the following equation is output.

[0036]

(Equation 6)

The outputs of the vertical LPFs 155a and 155b are
In the vertical 2: 1 sub-sampling circuits 157a and 157c for thinning the vertical pixels by 2: 1, the number of pixels in the vertical direction is 1 /
2 and are output from the output terminals 158a and 158c, respectively. The signal output from the output terminal 158a is a signal of the LL portion of the four frequency bands shown in FIG.
The signal output from the output terminal 158c is the signal of the portion HL in FIG. On the other hand, the outputs of the horizontal 2: 1 sub-sampling circuits 154a and 154b are output from vertical high-pass filters 156a and 156b (hereinafter referred to as vertical HPF
6a, 156b). Vertical HPF 156a, 156b
Is an even-tap filter having a frequency characteristic as shown in FIG. 15B, and its coefficient h ₄ (m) (m = 0,..., M−1) is h ₄ (m) = h ₃ (m ) ・ (-1) ^m . Therefore, h ₄ (m) = − h ₄ (Mm−1), m = 0,..., (M / 2) −1. That is, the vertical HPFs 156a and 156b are respectively
For each column w (m) (m = 1,..., 256) of the image signal output from the horizontal 2: 1 sub-sampling circuit 154a, 154b, a signal represented by the following equation is output.

[0038]

(Equation 7)

The outputs of the vertical HPFs 156a, 156b are
In the vertical 2: 1 sub-sampling circuits 157b and 157d,
The number of pixels in the vertical direction is reduced by half, and each output terminal
Output from 158b and 158d. The signal output from the output terminal 158b is LH out of the four frequency bands shown in FIG.
Is the signal of the part. The signal output from the output terminal 158d is the signal of the portion HH in FIG.

The above is the operation of the sub-band division circuit.
The four signals that have been sub-band divided and output are coded using predictive coding, orthogonal transform, etc., respectively.
Transmitted. On the decoding side, after decoding these signals,
Perform subband synthesis. The operation of the subband synthesizing circuit shown in FIG. 14 is completely opposite to the operation of the subband dividing circuit shown in FIG. The signals output from the output terminals 158a to 158d on the division side are input to the input terminals 159a to 159d, respectively, and 0 is interpolated in the vertical 1: 2 interpolation circuits 160a to 160d, respectively, so that the pixels in the vertical direction are Number doubles.
Output of vertical 1: 2 interpolation circuit 160a, 160c is vertical LP respectively
Input to F 161a and 161b. The vertical LPFs 161a and 161b are filters having exactly the same characteristics as the vertical LPFs 155a and 155b, and each column u ₁ ′ (m) (m) of the image signal output from the vertical 1: 2 interpolation circuit 160a or 160c, respectively. = 1,..., 256).

[0041]

(Equation 8)

On the other hand, the outputs of the vertical 1: 2 interpolation circuits 160b and 160d are input to the vertical HPFs 162a and 162b, respectively. The vertical HPFs 162a and 162b are filters having exactly the same characteristics as the vertical HPFs 156a and 156b.
Each column u ₂ ′ (m) of image signals output from 160b and 160d (m
= 1,..., 256).

[0043]

(Equation 9)

The calculator 163a subtracts the output of the vertical HPF 162a from the output of the vertical LPF 161a, and the calculator 163b calculates the vertical LPF 161b.
The output of the vertical HPF 162b is subtracted from the output of. Arithmetic unit163
The outputs of a and 163b are horizontal 1: 2 interpolation circuits 164a and 164b, respectively.
In 164b, 0 is interpolated to double the number of pixels in the horizontal direction. The output of horizontal 1: 2 interpolation circuit 164a is horizontal LPF1
Entered in 65. The horizontal LPF 165 is a filter having exactly the same characteristics as the horizontal LPF 152, and has a horizontal 1: 2 interpolation circuit.
Each line y ₁ ′ (n) (n =
1,..., 256), a signal represented by the following equation (3) is output.

[0045]

(Equation 10)

On the other hand, the output of the horizontal 1: 2 interpolation circuit 164b is
Input to the horizontal HPF 166. Horizontal HPF 166 is horizontal HPF
A filter with exactly the same characteristics as the 153
Each line y ₂ ′ of the image signal output from the interpolation circuit 164b
(n) For (n = 1,..., 256), a signal represented by the following equation (4) is output.

[0047]

[Equation 11]

The arithmetic unit 167 subtracts the output of the horizontal LPF 166 from the output of the horizontal LPF 165, and outputs the subtracted signal to the output terminal 16
Output from 8.

The conventional subband dividing / synthesizing circuit is configured as described above, but has a problem in filtering at an end point of an image. In other words, the aforementioned document "Sub-
In the "BandCoding of Monochrome and Color Images", a 16-tap filter as shown in Table 1 below is used for both the horizontal and vertical filters. For example, when an input image signal is passed through a horizontal LPF 152,
For each line x (n) (n = 1,..., 256) of the image signal,
There is a problem that the values of x (−6),..., X (0) and x (257),.

[0050]

[Table 1]

Conventionally, as a countermeasure for such a problem, there has been a method in which an input signal x (n) (n = 1,..., 256) to a filter is folded back, connected, and operated. That is, And perform a filter operation.

However, when the image signal is folded and filtered as described above, there is a problem that the result of subband division and synthesis at the end point of the image is not completely restored. In order to specifically illustrate this problem, the first line x (n) (n = 1,..., 256) of the input image signal is represented by x (1) = 16, x (2) = 120, x ( 3) = 130, x (4) =
140, x (5) = 150, x (6) = 160, x (7) = 170, x
(8) = 180, x (9) = 190, x (10) = 200, x (n) = 200,
(N = 11,..., 256) is considered. This is passed through a 16-tap horizontal LPF 152 as shown in Table 1 to provide a horizontal 2: 1 subsampling circuit.
2 154a: When decimating the 1, (1), (5) from the _{equation, y 1 (1) = 59.6} , y 1 (3) = 144.7, y 1 (5) = 152.
6, y ₁ (7) = 175.1, y ₁ (9) = 195.3, y ₁ (11) = 20
0.5, y ₁ (13) = 199.8, y ₁ (15) = 200.0, y ₁ (17) =
199.9… is output. On the other hand, through a horizontal HPF 153,
When passed through the sub-sampling circuit 154b, Equation (2),
From the formula, y ₂ (1) = −37.7, y ₂ (3) = 8.2, y ₂ (5) = −3.
1, y ₂ (7) = 1.0, y ₂ (9) = −1.6, y ₂ (11) = 0.
1, y ₂ (13) = − 0.1, y ₂ (15) = 0.1, y ₂ (17) = 0.0,
… Is output. Subsequent vertical 2: 1 subsampling circuit
Assuming that no distortion occurs in the vertical subband division in 155a to 157d and the vertical subband synthesis by the vertical 1: 2 interpolation circuits 160a to 163b, the arithmetic unit 163a
Is y ₁ (n) (n = 1, 3, 5,..., 255).
The output of b is y ₂ (n) (n = 1, 3, 5,..., 255). here,
The output from the arithmetic units 163a and 163b is looped back in the same manner as in equation (5).

That is, It is. At this time, the output y ₁ ′ (n) (n = −255, −254,..., 511) of the horizontal 1: 2 interpolation circuit 164a and the output y ₂ ′ (n) (n) of the horizontal 1: 2 interpolation circuit 164b = −255, −254,…, 511)
Are Becomes Therefore, the output x ₁ (n) of the horizontal LPF 165 is
(3) from the _{equation, x 1 (1) = 46.3} , x 1 (2) = 77.9, x 1 (3) = 129.
4, x ₁ (4) = 156.1, x ₁ (5) = 151.0, x ₁ (6) = 15
3.2, x ₁ (7) = 169.1, x ₁ (8) = 182.0, x ₁ (9) =
191.3,..., And the output x ₂ (n) of the horizontal HPF 166 is obtained from the equation (4) as x ₂ (1) = 45.5, x ₂ (2) = − 27.4, x ₂ (3) = −
1.8, x ₂ (4) = 11.9, x ₂ (5) = 0.7, x ₂ (6)
_{= -6.0, x 2 (7)} = -0.7, x 2 (8) = 2.1, x 2
(9) = 1.4, ... Therefore, when the output X (n) of the arithmetic unit 167 is rounded off to the decimal point, X (1) = 0, X (2) = 105, X (3) = 131, X
(4) = 144, X (5) = 150, X (6) = 159, X (7)
= 170, X (8) = 180, X (9) = 190,... Near the end points of the image (n = 1,..., 4)
X (n) ≠ x (n), which indicates that the image cannot be accurately restored.

The present invention has been made in view of such circumstances, and has as its main object to provide an encoding method and an encoding apparatus capable of efficiently compressing the data amount of a digital image signal such as a color television signal. And

Another object of the present invention is to provide an encoding method and an encoding method capable of performing optimal one-dimensional scanning when performing three-dimensional orthogonal transformation after subband division and reducing the redundancy of digital image signals. An object of the present invention is to provide an encoding device.

Still another object of the present invention is to provide an encoding method and an encoding apparatus capable of reducing redundancy even in a digital image signal of the interlaced scanning method.

Still another object of the present invention is to provide an encoding method and an encoding apparatus capable of obtaining good image quality on the decoding side even in a flat portion where image quality degradation is conspicuous.

Still another object of the present invention is to provide a coding method capable of performing weighting suitable for frequency characteristics on transform coefficients of each subband block and performing effective information compression. An object of the present invention is to provide an encoding device.

Still another object of the present invention is to provide an encoding method and an encoding apparatus which can reduce the image quality degradation and reduce the encoding rate.

It is still another object of the present invention to provide a band division / synthesis method capable of accurately restoring even an end point of an image.

[0061]

An encoding method according to a first aspect of the present invention divides a digital image signal into sub-bands,
In a method of performing orthogonal transformation of two or more dimensions on each of the divided subbands and one-dimensionally scanning and encoding the obtained transformation coefficients, the position of the transformation coefficient at which one-dimensional scanning is started for each subband It is characterized by changing.

A coding method according to a second invention of the present application is a method for performing coding by dividing a digital image signal into sub-bands and forming each of the divided band components into blocks.
It is characterized in that a part of each of the divided components is decimated and coded at a fixed cycle.

A coding method according to a third invention of the present application is a method for performing coding by dividing a digital image signal into subbands, forming each of the divided band components into blocks, and performing coding.
It is characterized in that it is determined whether or not the block is an effective image block based on the activity of the block, and only the effective image block is encoded.

The encoding method according to the fourth invention of the present application is a method for dividing a digital image signal into sub-bands, performing orthogonal transform on each of the divided band components, quantizing the transform coefficients, and encoding. A quantization step width is determined based on a pixel value of a high-band block before conversion.

In the encoding method according to the fifth invention of the present application, one of an odd field and an even field is passed through an odd tap vertical filter and the other is an even tap vertical filter for an interlaced digital image signal. After passing through, a two-dimensional block is formed in the field, a three-dimensional block is formed by combining odd and even fields in the time direction, and three-dimensional orthogonal transform is performed and encoded.

An encoding method according to a sixth aspect of the present invention is a method for encoding by dividing a digital image signal into sub-bands, performing orthogonal transform on each of the divided band components, and performing weighting on the transform coefficients. , Weighting is performed so that the weighting is continuous in the frequency direction for each subband.

In the band division / combination method according to the seventh invention of the present application, when subband division / combination is performed using an even-tap filter, at the time of division, a folded image is connected to an end point of the image, and then filtering is performed. At the time of synthesis, the folded image is connected to the end point of the image before passing through the low-pass filter, and each pixel value folded at the end point of the image is multiplied by -1 before passing through the high-pass filter. Are connected.

In the band dividing / combining method according to the eighth invention of the present application, when subband dividing / combining is performed using a filter of an odd tap, at the time of dividing and combining, folding back to the end point of the image is performed before filtering. The image is characterized in that an image reduced by one pixel is connected.

[0069] The encoding device according to a ninth aspect of the present invention is di
A coding apparatus for coding a digital image signal,
Divide digital image signal into image signals of multiple subbands
Sub-band division means to divide and image signal of each sub-band
Into two or more dimensional blocks for each pixel
Blocking means and each block
Orthogonal transformation means for performing an intersection transformation to obtain a transformation coefficient;
The transformed coefficients are first-ordered from different starting positions for each subband.
Encoding means for performing original scanning and encoding .

[0070] The encoding device according to a tenth aspect of the present invention is di
A coding apparatus for coding a digital image signal,
Divide digital image signal into image signals of multiple subbands
Sub-band division means to divide and image signal of each sub-band
Means for blocking a plurality of pixels for each of a plurality of pixels;
A means for thinning out some locked blocks and the remaining
Orthogonal transform that performs orthogonal transform on each block to obtain transform coefficients
Means and coding means for coding the obtained transform coefficients.
Characterized in that it comprises.

[0071] The encoding apparatus according to an eleventh aspect of the present invention is di
A coding apparatus for coding a digital image signal,
Divide digital image signal into image signals of multiple subbands
Sub-band division means to divide and image signal of each sub-band
Means for blocking a plurality of pixels for each of a plurality of pixels;
Transform coefficients by applying orthogonal transformation to each locked block
And orthogonal transform means for obtaining
Quantization steps with different quantization step widths based on the values
Means for selecting one quantization table from the
Quantify the transform coefficients obtained based on the selected quantization table.
Encoding means for encoding the data .

[0072] The encoding apparatus according to a twelfth aspect of the present invention is di
A coding apparatus for coding a digital image signal,
Divide digital image signal into image signals of multiple subbands
Sub-band division means to divide and image signal of each sub-band
Means for blocking a plurality of pixels for each of a plurality of pixels;
Transform coefficients by applying orthogonal transformation to each locked block
Orthogonal transform means to obtain the obtained transform coefficients
Encoding means, wherein the orthogonal transformation means comprises:
Of the sub-band block obtained by the
The means for seeking the activity and the required activity
Means for determining whether the block is an effective image block based on the
Means for outputting a conversion coefficient in the effective image block.
Characterized in that it Bei.

[0073]

[0074]

A coding apparatus according to a thirteenth invention of the present application is a coding apparatus for coding a digital image signal of the interlaced scanning system, wherein an odd tap of one of an odd field and an even field of the digital image signal passes. A first vertical filter, a second vertical filter of even taps passing through the other of the odd field and the even field of the digital image signal, and an image signal output from the first and second vertical filters for each of a plurality of pixels. Blocking means for forming a block into three-dimensional blocks, orthogonal transform means for performing a three-dimensional orthogonal transform on each of the blocked blocks to obtain a transform coefficient, and coding means for coding the obtained transform coefficient. It is characterized by having.

According to a fourteenth aspect of the present invention, in the coding apparatus for coding a digital image signal, a subband dividing means for dividing the digital image signal into a plurality of subband image signals, Blocking means for blocking the image signal of the band for each of a plurality of pixels, orthogonal transform means for performing an orthogonal transform on each of the blocked blocks to obtain transform coefficients, and performing weighting on the obtained transform coefficients It is characterized by comprising a weighting means and an encoding means for encoding an output of the weighting means.

The coding apparatus according to the fifteenth invention of the present application is characterized in that the weighting means performs weighting such that weighting for each subband is continuous in the frequency direction.

An encoding apparatus according to a sixteenth aspect of the present invention is an encoding apparatus for encoding a digital image signal, comprising: a sub-band dividing means for dividing the digital image signal into a plurality of sub-band image signals; Blocking means for blocking the image signal of the band for each of a plurality of pixels, orthogonal transform means for performing an orthogonal transform on each of the blocked blocks to obtain a transform coefficient, and a plurality of encoding tables corresponding to each subband And encoding means for encoding the obtained transform coefficients based on an encoding table selected from these. An encoding apparatus according to a seventeenth aspect of the present invention is an encoding apparatus for encoding a digital image signal, comprising: a subband dividing unit that divides the digital image signal into image signals of a plurality of subbands; Blocking means for blocking a signal for each of a plurality of pixels, orthogonal transform means for performing a transform on each of the blocked blocks to obtain a transform coefficient, and weighting means for performing a weighting on the obtained transform coefficient; , A plurality of encoding tables corresponding to each subband, and encoding means for encoding the output of the weighting means based on an encoding table selected from these.

[0079]

According to the first and ninth aspects of the present invention, when each band component of a digital image signal divided into a plurality of frequency bands is subjected to orthogonal transformation of two or more dimensions and encoded, the power is maximized for each band. One-dimensional scanning is started from the conversion coefficient. Then, the code amount is reduced.

According to the second and tenth aspects, of each band component of the digital image signal divided into a plurality of frequency bands, a component in which image quality deterioration is not conspicuous is thinned out periodically. Then,
The coding rate is reduced while keeping the image quality deterioration small.

In the third and twelfth inventions, image blocks in which image quality degradation is not conspicuous are thinned out of each band component of the digital image signal divided into a plurality of frequency bands. Then,
The encoding rate is reduced while keeping the image quality deterioration small.

In the fourth and eleventh inventions, orthogonal transform is performed on each band block of the digital image signal divided and blocked according to the frequency, and obtained based on the pixel values of the high band block before the conversion. The quantization step width when quantizing the transformed coefficients is determined. Based on each pixel value in the block, it is determined whether this block includes a flat portion where image quality degradation is conspicuous or a portion where the change is drastic. Choose a quantization table with a wide quantization step width,
Select a quantization table with a narrow quantization step width for blocks that contain flat parts and parts that change rapidly, and for blocks that contain flat parts but do not contain parts that change rapidly. A quantization table having a medium quantization step width is selected. In this way, in the restored image, good image quality is maintained even in a flat portion.

In the fifth and thirteenth inventions, one of the odd field and the even field is a vertical filter having an odd tap,
The other is passed through an even-tap vertical filter, and these are spring-loaded to form a three-dimensional block. Then, the spatial sampling positions of both fields are aligned, and the digital image signal of the interlaced scanning method has a more accurate signal configuration in the time axis direction, thereby increasing the signal compression ratio.

According to the sixth, fourteenth , and fifteenth inventions, a transform coefficient obtained by orthogonally transforming a digital image signal that has been divided into subbands and divided into blocks according to the frequency is suitable for the frequency characteristic in each subband. Waiting is applied. Moreover, in the sixth and fifteenth inventions, this weighting is smoothly connected between the respective bands.

In the seventh and eighth inventions, the folded image is connected for filtering the image at the time of division, but the connected portion not transmitted is reproduced at the time of synthesis.
Therefore, even at the end points of the image, the original image is correctly restored.

In all the inventions except the fifth and thirteenth inventions, the digital image signal is divided into sub-bands, and each sub-band is divided into blocks. Therefore, the compression coding can be performed at a higher level than in the case where the subband division is not performed.

[0087]

Embodiments of the present invention will be specifically described below.

FIGS. 17 and 18 are block diagrams showing the configuration of an example of the encoding apparatus according to the present invention. FIG. 17 shows the configuration on the code side, and FIG. 18 shows the configuration on the decoding side. In FIG. 17, reference numeral 1 denotes an input terminal of an NTSC color television signal, and an NTSC decoder 2 separates a luminance signal (Y signal) and color signals (I, Q signals) of the color television signal. The separated Y, I, Q signals are output to a matrix circuit 3, which converts these signals into a luminance signal (Y signal) and a color difference signal (RY, BY signal). These signals are output to the A / D converter 4. A
The / D converter 4 converts the Y, RY, and BY signals into digital signals, and outputs these to the subband dividing circuit 5. The sub-band dividing circuit 5 divides each of the Y, RY, and BY signals into four sub-bands according to the frequency,
Each sub-band is divided into blocks, and the blocks are output to the three-dimensional orthogonal transform circuit 6. The three-dimensional orthogonal transform circuit 6 applies a discrete cosine transform (Discrete Cosine Transf) to each block.
om: DCT) to obtain transform coefficients, and output the obtained transform coefficients to the encoder 7. The encoder 7 quantizes and encodes the input transform coefficient, and outputs the encoded data to an output terminal 8.
Output via.

In FIG. 18, reference numerals 11 to 18 denote constituent members on the decoder side. Reference numeral 11 denotes an input terminal to which the data encoded as described above is inputted. A decoder 12 connected to the input terminal 11 restores the encoded data into three-dimensional data, and converts the data into three-dimensional data. Output to the dimension inverse orthogonal transform circuit 13. 3
The dimensional inverse orthogonal transform circuit 13 performs an inverse DCT on the three-dimensional data. The sub-band synthesis circuit 14 calculates the 3
The dimensional data is synthesized, returned to the original field image, and digital Y, RY, and BY signals are output to the D / A converter 15. The D / A converter 15 converts these signals into analog signals and outputs them to the matrix circuit 16. The matrix circuit 16 converts analog Y, RY, and BY signals into Y, I, and Q signals and outputs the signals to the NTSC encoder 17. The NTSC encoder 17 converts these Y, I, and Q signals into NT signals.
The signal is converted into an SC color television signal and output via an output terminal 18.

FIG. 19 is a block diagram showing the internal configuration of the subband dividing circuit 5. In the figure, 21, 22, and 23 indicate A /
Y signal, RY signal, BY output from the D converter 4
This is a signal input terminal. The input terminals 21, 22, and 23 are connected to a Y signal sub-band dividing circuit 24, an RY signal sub-band dividing circuit 25, and a BY signal sub-band dividing circuit 26, respectively. The internal configuration of each of these division circuits 24, 25, and 26 is the same, and FIG.
Only the internal configuration of 24 is shown. The Y signal sub-band dividing circuit 24 includes a vertical low-pass filter (vertical LPF) 241
And vertical high-pass filter (vertical HPF) 242 and vertical LP
Vertical 2: 1 sub-sampling circuits 243a and 243b for sampling the output from the F 241 and the vertical HPF 242 so that the number of pixels in the vertical direction is halved, and horizontal low-pass filters (horizontal LPFs) 244a and 244b. And horizontal high pass filters (horizontal HPF) 245a, 245b, and horizontal LPF 244a, horizontal HPF 245a,
Horizontal 2: sampling the output from the horizontal LPF 244b and the horizontal HPF 245b so that the number of pixels in the horizontal direction is halved.
It comprises one sub-sampling circuit 246a, 246b, 246c, 246d. The three-dimensional data output from each of the division circuits 24, 25, 26 is stored in the memory 27.

FIG. 20 is a block diagram showing the internal configuration of the subband synthesis circuit 14. In the figure, reference numeral 32 denotes a memory for storing three-dimensional data output from the three-dimensional inverse orthogonal transform circuit 13 via the input terminal 31. The memory 32 includes a Y signal sub-band synthesizing circuit 33, an RY signal sub-band synthesizing circuit 34, B
-Y signal sub-band synthesis circuits 35 are connected to each other. The internal configuration of each of these combining circuits 33, 34, and 35 is the same, and FIG. 20 shows only the internal configuration of Y signal sub-band combining circuit 33. Y signal subband synthesis circuit
33 is a horizontal 1: 2 interpolation circuit 332a, 332b, 332c, which doubles the number of pixels in the horizontal direction by interpolating 0 as a value.
332d, horizontal LPF 333a, 333b, horizontal HPF 334a, 334b
And subtracters 335a and 335b, and a vertical 1: 2 interpolation circuit that doubles the number of pixels in the vertical direction by interpolating 0 as a value
336a and 336b, a vertical LPF 337, a vertical HPF 338, and a subtractor 339. The Y signal, the RY signal, and the BY signal are output from each of the combining circuits 33, 34, and 35 to a D / A converter.
Output to 15 respectively.

Next, the operation will be described.

Generally, in order to compress an image signal, a luminance signal and a chrominance signal are often handled independently. Therefore, the NTSC color television signal input from the input terminal 1 is
The SC decoder 2 separates the signal into a Y signal and I and Q signals. The matrix circuit 3 converts the signal into a Y signal and RY and BY signals. The A / D converter 4 converts the signal into a digital signal. Convert. The sampling frequency at this time is 13.5 M for the Y signal.
Hz, RY and BY signals are 6.75 MHz. Therefore, NT
In the case of the SC color television signal, the number of effective samples of one horizontal line is, for example, 704 for the Y signal and 352 for the RY and BY signals, and one field is composed of 262.5 horizontal lines. Of these, for example, 240
The line is output as one field. In the sub-band dividing circuit 5, each field of the Y signal, the RY signal, and the BY signal is divided into a plurality of frequency bands. For example, the Y signal is divided into four frequency bands LL, HL, LH, and HH as shown in FIG. 21, and the RY signal and the BY signal are divided as shown in FIG.
Divide into four frequency bands of LL, HL, LH, HH.

The operation of the sub-band division circuit 5 will be described with reference to FIG. The Y signal input from the input terminal 21 is Y
In the signal sub-band division circuit 24, the signal is divided into four bands. The input Y signal is band-limited by a vertical LPF 241 having frequency characteristics as shown in FIG. 23, and then the number of pixels in the vertical direction is reduced to 1/2 in a vertical 2: 1 sub-sampling circuit 243a. The output of the vertical 2: 1 sub-sampling circuit 243a is passed through a horizontal LPF 244a having a frequency characteristic as shown in FIG. 25. In the horizontal 2: 1 sub-sampling circuit 246a, the number of pixels in the horizontal direction is reduced to 1/2. I will The output of the horizontal 2: 1 sub-sampling circuit 246a is a signal in the LL band in FIG. 21, and the number of pixels is 1/4 of the input signal. This signal is called the LL band of the Y signal. On the other hand, the output of the vertical 2: 1 sub-sampling circuit 243a is also input to the horizontal HPF 245a having frequency characteristics as shown in FIG. 26, and the horizontal 2: 1 sub-sampling circuit 246b reduces the number of pixels in the horizontal direction to １ ／. Is thinned out. The output of the horizontal 2: 1 sub-sampling circuit 246b is a signal in the HL band of FIG. 21, and the number of pixels is 1/4 of the input signal. This signal is called the HL band of the Y signal.

Incidentally, the Y signal input from the input terminal 21 is also input to the vertical HPF 242 having frequency characteristics as shown in FIG. The output of the vertical HPF 242 is reduced by 1/1 in the vertical 2: 1 sub-sampling circuit 243b.
Thinned out to 2. Vertical 2: 1 sub-sampling circuit 243b
The output of the horizontal LPF 244b with frequency characteristics as shown in Figure 25
, The horizontal 2: 1 sub-sampling circuit 246c reduces the number of pixels in the horizontal direction to に. The output of the horizontal 2: 1 sub-sampling circuit 246c is a signal in the LH band in FIG. 21, and the number of pixels is 1/4 of the input signal. This signal is called the LH band of the Y signal. On the other hand, the output of the vertical 2: 1 sub-sampling circuit 243b is also input to a horizontal HPF 245b having a frequency characteristic as shown in FIG. 26, and the horizontal 2: 1 sub-sampling circuit 246d reduces the number of pixels in the horizontal direction to ２. Is thinned out. The output of the horizontal 2: 1 sub-sampling circuit 246d is a signal in the HH band of FIG. 21, and the number of pixels is １ ／ of the input signal. This signal is called the HH band of the Y signal.

As described above, the Y signal sub-band dividing circuit 24
, The Y signal is divided into four bands LL, HL, LH, HH, and each subband component is output.

The RY signal input from the input terminal 22 is R
In the −Y signal sub-band dividing circuit 25, the signal is divided into four bands LL, HL, LH, and HH as shown in FIG. Also,
The BY signal input from the input terminal 23 is divided into four bands LL, HL, LH, and HH as shown in FIG. The operation of the RY signal sub-band dividing circuit 25 and the BY signal sub-band dividing circuit 26 is the same as the operation of the Y signal sub-band dividing circuit 24.

The LL band, HL band, LH band, and HH band of the Y signal output from the Y signal
RY output from the RY signal subband division circuit 25
LL band, HL band, LH band, HH band and B of signal
The LL band, HL band, LH band, and HH band of the BY signal output from the -Y signal sub-band division circuit 26 are input to the memory 27, where eight fields are accumulated. memory
27 collects the three bands into a three-dimensional block composed of a plurality of adjacent pixels and outputs them in block units until the next eight fields are accumulated. For example, adjacent 8
In each line, adjacent pixels of 8 pixels per line are formed as a two-dimensional block, and adjacent two-dimensional blocks of eight consecutive fields are combined to form a three-dimensional block having a size of 8 pixels × 8 lines × 8 fields. The Y signal, RY signal, and BY signal that are three-dimensionally blocked are output from the memory 27 in the order of LL band, HL band, LH band, and HH band.

Each block output from the subband dividing circuit 5 is subjected to a three-dimensional DCT by a three-dimensional orthogonal transform circuit 6. The transform coefficients output from the three-dimensional orthogonal transform circuit 6 are encoded by the encoder 7 and one-dimensionally scanned, and for each non-zero value, its value and a sequence of zero values up to that value are calculated. The length is variable-length coded by a two-dimensional Huffman code or the like, and output from the output terminal 8. This variable length code is
A different variable length code table may be used for each band, or the same one may be used.

After performing a three-dimensional DCT on each block (one block is 8 pixels × 8 lines × 8 fields) of the Y band of a Y signal of a certain natural moving image, the block is quantized to 10 bits and converted. The root mean square (RMS) of the coefficients d (i, j, k) was determined. With reference to the result of the RMS, a description will be given of the difference in the start position of the one-dimensional scan in each subband block.

Tables 2 and 3 show the RMS of the conversion coefficient in the LL band of the Y signal.

[0102]

[Table 2]

[0103]

[Table 3]

From the results in Tables 2 and 3, in the LL band, when viewed from a block in a two-dimensional plane, (i, j) = (0, 0)
Power is concentrated around the area. Therefore, in the LL band, as shown in FIG. 27 (a), one-dimensional scanning from a horizontal and vertical low-order sequence conversion coefficient to a high-order sequence conversion coefficient is suitable. For a three-dimensional block, scanning such a two-dimensional block requires 8
Repeat several times.

Tables 4 and 5 show the RMS of the conversion coefficient in the HL band of the Y signal.

[0106]

[Table 4]

[0107]

[Table 5]

From the results in Tables 4 and 5, in the HL band, when viewed from a block in a two-dimensional plane, (i, j) = (7, 0)
Power is concentrated around the area. This is because after passing through the HPF 245a having a frequency characteristic as shown in FIG. 26, sampling is performed at 2: 1. Figure 28 shows the HL band.
As shown in (1), the signal is a signal folded in the horizontal direction. For example, a signal of 6.75 MHz becomes a signal of 0 MHz. As a result,
Although the power of the transform coefficients after DCT should originally concentrate on the lower-order sequence, in such a case, since the signal is a signal folded in the horizontal direction, the power concentrates on the higher-order sequence only in the horizontal direction. Therefore, in the HL band, FIG.
As shown in (b), one-dimensional scanning from the upper right transform coefficient to the lower left transform coefficient in the figure where the horizontal direction is a high-order sequence and the vertical direction is a low-order sequence is suitable.

Tables 6 and 7 show the RMS of the conversion coefficient in the LH band of the Y signal.

[0110]

[Table 6]

[0111]

[Table 7]

From the results shown in Tables 6 and 7, in the LH band, when viewed from a block in a two-dimensional plane, (i, j) = (0, 7)
Power is concentrated around the area. This is because the LH band is a signal folded in the vertical direction. Therefore,
In the LH band, as shown in FIG. 27C, the one-dimensional scan from the lower left transform coefficient to the upper right transform coefficient in the figure where the horizontal direction is a low-order sequence and the vertical direction is a high-order sequence is performed. Are suitable.

Tables 8 and 9 show the RMS of the conversion coefficient in the HH band of the Y signal.

[0114]

[Table 8]

[0115]

[Table 9]

From the results in Tables 8 and 9, in the HH band, when viewed from a block in a two-dimensional plane, (i, j) = (7, 7)
Power is concentrated around the area. Therefore, in the HH band, as shown in FIG. 27 (d), one-dimensional scanning is performed from a transform coefficient having a high-order sequence in both the horizontal and vertical directions to a transform coefficient having a low-order sequence in both the horizontal and vertical directions. Are suitable.

Next, the operation on the decoding side will be described.

The decoding side from the input terminal 11 to the output terminal 18 follows a completely opposite process from the code side from the input terminal 1 to the output terminal 8. The data input from the input terminal 11 is returned to the original three-dimensional data form at the decoder 12, and subjected to inverse DCT transform at the three-dimensional inverse orthogonal transform circuit 13. The three-dimensional block subjected to the inverse orthogonal transform is used as a subband synthesis circuit.
At 14, the four subbands are combined and output one field at a time.

The operation of the subband synthesizing circuit 14 is completely opposite to the operation of the subband dividing circuit 5. The three-dimensional block input from the input terminal 31 is stored in the memory 32 for eight fields. The memory 32 transfers the LL, HL, LH, and HH bands of the Y signal to the Y signal subband synthesizing circuit 33 and stores the LL, HL, LH, and HH of the RY signal until the next eight fields are accumulated.
The band is input to the RY signal sub-band synthesizing circuit 34, and the LL, HL, LH, and HH bands of the BY signal are input to the BY signal sub-band synthesizing circuit 35 one by one.

Here, the operation of the Y signal sub-band synthesizing circuit 33 will be described. Y output from memory 32
The LL band of the signal is input to a horizontal LPF 333a having frequency characteristics as shown in FIG. 25 after doubling the number of pixels in the horizontal direction by interpolating 0 in a horizontal 1: 2 interpolation circuit 332a. The HL band of the Y signal output from the memory 32 is interpolated by a horizontal 1: 2 interpolation circuit 332b to 0, and then input to a horizontal HPF 334a having a frequency characteristic as shown in FIG. The subtractor 335a outputs the horizontal HPF from the output of the horizontal LPF 333a.
Subtract the output of 334a. The output of the subtractor 335a is vertical 1:
The number of pixels in the vertical direction is doubled by interpolating 0 in the two interpolation circuit 336a, and then input to the vertical LPF 337 having frequency characteristics as shown in FIG. On the other hand, the LH band of the Y signal output from the memory 32 is a horizontal 1: 2 interpolation circuit 33.
After doubling the number of pixels in the horizontal direction by interpolating 0 in 2c, a horizontal LP having frequency characteristics as shown in FIG.
Input to F 333b. Also, the Y output from the memory 32
The HH band of the signal is obtained by a horizontal 1: 2 interpolation circuit 332d.
After horizontal interpolation, horizontal HP with frequency characteristics as shown in Fig. 26
Entered in F 334b. The subtractor 335b subtracts the output of the horizontal HPF 334b from the output of the horizontal LPF 333b. The output of the subtractor 335b is input to a vertical HPF 338 having frequency characteristics as shown in FIG. 24 after the number of pixels in the vertical direction is doubled by interpolating 0 in a vertical 1: 2 interpolation circuit 336b. You. The subtractor 339 subtracts the output of the vertical HPF 338 from the output of the vertical LPF 337 and outputs the result. The operation of the RY signal sub-band synthesizing circuit 34 and the BY signal sub-band synthesizing circuit 35
The operation is the same as that in the above-described Y signal sub-band combining circuit 33.

As described above, the sub-band combining circuit 14
The Y, RY, and BY signals output from
After being converted into an analog signal by a D / A converter 15 and further converted into a Y signal, an I signal and a Q signal by a matrix circuit 16, the signal is converted into an NTSC color television signal by an NTSC encoder 17, and is output from an output terminal 18. Is output.

In the above embodiment, as shown in FIG. 27, a case where so-called “zigzag scan” is performed is shown. However, the one-dimensional scanning of the encoder 7 is low in both horizontal and vertical directions for the LL band. From the conversion coefficient, which is the next sequence, HL
For the band, from the transform coefficients corresponding to the horizontal high-order sequence and the vertical low-order sequence, for the LH band,
It is important to start one-dimensional scanning from a conversion coefficient corresponding to a vertical low-order sequence and a vertical high-order sequence, and to a HH band from a conversion coefficient which is a high-order sequence in both the horizontal and vertical directions. Therefore, for example, one-dimensional vertical scanning as shown in FIG. 29A for the HL band and FIG. 29B for the HH band may be performed.

In the above embodiment, scanning of a three-dimensional block of 8 × 8 × 8 is obtained by repeating scanning of a two-dimensional block of 8 × 8 as shown in FIG. 27 eight times. Was. That is, in the case of the LL band, scanning was performed as shown in FIG. However, as shown in FIG. 30B, the moving image may be scanned in the time direction first.

Next, another embodiment of the present invention will be described. This embodiment is an example in which components of a certain sub-band after sub-band division are not subjected to orthogonal transformation but are thinned out.

Each block output from the sub-band division circuit 5 is orthogonally transformed by the three-dimensional orthogonal transformation circuit 6, and the Y signal is subjected to LL band and HL band as shown in FIG.
Only the LH band and the LH band are orthogonally transformed, and as for the RY signal and the BY signal, only the LL band is orthogonally transformed as shown in FIG. The transform coefficients that are thus orthogonally transformed and output from the three-dimensional orthogonal transform circuit 6 are encoded by the encoder 7
After the encoding, a one-dimensional scan is performed, and for each non-zero value, a predefined Huffman code indicating the value and a Huffman code indicating the length of the sequence of zero values up to the value Is output from the output terminal 8.

Here, an example in which encoding and decoding are performed using a certain sample image will be described. The LL band, HL band, LH band, and HH band obtained by dividing the sample image into 4 parts in the horizontal and vertical directions are 8 pixels x 8 lines x 8
After constructing the field into a three-dimensional block, performing three-dimensional orthogonal transform and quantizing, the original image is restored by performing three-dimensional inverse orthogonal transform and subband synthesis. Table 10 shows the results of encoding four frames of such sample images. The S / N ratio is calculated by the following equation (6).

[0127]

[Table 10]

[0128]

(Equation 12)

In the method 1, the image obtained by dividing the sub-band is
After performing dimensional orthogonal transform, the obtained transform coefficients were multiplied by appropriate weights (1 to 0.4) according to the frequency band, and then one-dimensional scanning was performed to perform Huffman coding. Method 2 is an embodiment of the present invention, in which the HH band of the Y signal and the HL signal of the RY signal and the BY signal are used.
Band, LH band and HH band were decimated and coded.
Looking at each band, the main components are almost concentrated in the LL band, and the energy decreases in the order of the LH band, the HL band, and the HH band. In the method 2 of the present invention, the total bit rate is reduced by 10% or more compared to the method 1,
For the Y signal and the BY signal, the total bit rate is reduced by almost 40%. In the method 2, the S / N ratio is 1 compared to the method 1.
Although it has been deteriorated by about 2 dB, deterioration of the reproduced image can hardly be visually confirmed.

In the above-described embodiment, an example in which the HH band of the Y signal and the LH band, HL band, and HH band of the RY signal and the BY signal are thinned out in each field has been described. Each band may be thinned out at a constant period as shown in FIG. FIG. 33 shows an example in which each band of the Y signal is thinned out in an 8-field cycle.
4, 7, 8 fields LH band and HL band, 3,
The HH bands of 4, 5, 6, 7, and 8 fields are thinned out. Since the sensitivity of the eyes decreases for moving images, it is difficult to notice deterioration even when high-frequency components are periodically thinned out. Since all components are encoded in the first and second fields, the resolution and S / N ratio of the still image of this frame are sufficient.

Next, still another embodiment of the present invention will be described. This embodiment is an example in which it is determined whether or not an image block is an effective image block, and only the effective image block is encoded.

FIG. 34 is a block diagram showing the internal configuration of the three-dimensional orthogonal transform circuit 6. The three-dimensional orthogonal transform circuit 6 includes an orthogonal transform circuit 42 that performs a DCT on each block output from the subband division circuit 5 via the input terminal 41 to obtain a transform coefficient, and a magnitude of an image signal in each block. A determination value calculating circuit 43 for calculating a determination value A for determining whether or not the block is an effective image block based on the determination result
And a determination reference value B for determining an effective image block
Criterion value table 45 as a memory that stores
And a determinator 44 that compares the criterion value A from the criterion value calculation circuit 43 with the criterion value B from the criterion value table 45 to determine whether the block is an effective image block. Then, the obtained transform coefficient is output to the encoder 7 via the output terminal 46, and the obtained determination result is output to the encoder 7 via the output terminal 47.

Each block output from the subband dividing circuit 5 is subjected to a three-dimensional DCT by a three-dimensional orthogonal transform circuit 6. The operation of the three-dimensional orthogonal transformation circuit 6 is shown in FIG.
This will now be described with reference to FIG.

The Y signal, RY input from the input terminal 41
Each block of the signal and the BY signal is subjected to three-dimensional DCT by the orthogonal transform circuit 42, and the transform coefficient is output to the output terminal 46.
Is output to the encoder 7 via the
Entered in 43. In the judgment value calculation circuit 43, the Y signal, R-
Based on the variance of the image signals in each block of the Y signal and the BY signal, a determination value A for determining whether or not the block is an effective image block is calculated. Assuming that a variance value in each block is σ _AC ² , the determination value calculation circuit 43 calculates A represented by the following equation as a determination value, and outputs the determination value A to the determiner 44. Where A = σ _AC ^2/512, variance sigma _AC ² is calculated from the following expression.

[0135]

(Equation 13)

The determination reference value B corresponding to each of the Y signal, the RY signal, and the BY signal for determining whether the image block is an effective image block is stored in a determination reference value table as a memory.
It is stored in 45 and output to the determiner 44 as needed. The determinator 44 compares the criterion value A output from the criterion value calculation circuit 43 with the criterion value B output from the criterion value table 45. If A ≧ B, this block is an effective image block. On the other hand, if A <B, it is determined that this block is not a valid image block. And
The determiner 44 outputs the determination result thus obtained to the encoder 7 via the output terminal 47.

In the encoder 7, based on the determination result,
The three-dimensional DCT coefficients are quantized and encoded. If the three-dimensional orthogonal transform circuit 6 determines that the block is a valid image block, the encoder 7 outputs block information indicating that the block is a valid image block, and outputs the three-dimensional DCT coefficient of this block. It is quantized and encoded as it is. If the three-dimensional orthogonal transform circuit 6 determines that the block is not a valid image block, the encoder 7 determines that this block is a block of the LL band of the Y signal, the RY signal, or the BY signal. Block information indicating that the block is not a valid image block is output, and only the DC component of the three-dimensional DCT coefficient is quantized and encoded. If the block is a block of the HL band, LH band, or HH band of the Y signal, the RY signal, or the BY signal, the encoder 7
Outputs only block information indicating that this block is not a valid image block, and does not quantize three-dimensional DCT coefficients.

Here, an example in which encoding and decoding are performed using a certain sample image will be described. The LL band, HL band, LH band, and HH band obtained by dividing the sample image into 4 parts in the horizontal and vertical directions are 8 pixels x 8 lines x 8
After constructing the field into a three-dimensional block, performing three-dimensional orthogonal transform and quantizing, the original image is restored by performing three-dimensional inverse orthogonal transform and subband synthesis. Table 11 shows the result of encoding four frames of such a sample image. The S / N ratio is calculated by the above equation (6).

[0139]

[Table 11]

The method 1 is a conventional method for coding without determining a valid image block. Method 2 is an embodiment of the present invention, in which an effective image block is determined, and if it is not an effective image block, all transform coefficients except for the DC component of the LL band are not encoded. The determination reference value B used for determining an effective image block is 0 in the LL and LH bands of the Y signal, 0.2 in the HL and HH bands, and 0, LH, and LH in the LL band of the RY signal and the BY signal. In the HL and HH bands
0.2.

From the results shown in Table 11, it can be seen that the total bit rate is reduced by about 17% in the scheme 2 of the present invention as compared with the scheme 1. In the method 2, compared to the method 1, the S / N
Although the ratio is deteriorated by about 2 to 3 dB, the deterioration of the reproduced image is hardly confirmed by visual observation, so that it is not a problem.

In the above-described embodiment, whether or not an effective image block is determined is based on the variance σ _AC ² of each block. However, the effective image block is determined based on the maximum value or dynamic range of each block. It is also conceivable to determine whether or not this is the case. In addition, for the blocks that are not valid image blocks, the components other than the DC component of the LL band were not encoded, but for the other LH, HL, and HH bands,
Only the DC component may be encoded.

Next, still another embodiment of the present invention will be described. FIG. 35 and FIG. 36 are block diagrams showing the configuration on the encoding side and decoding side in this embodiment. In FIG. 35 showing the code side, 90 is an input terminal of an image signal, and 50 is an analog / digital converter (hereinafter, referred to as an A / D converter). The sub-band dividing circuit 51 divides the digitized image signal into four bands according to the frequency, and
Output to The blocking circuit 52 is a sub-band division circuit
Each band signal output from 51 is 8 pixels x 8
It is configured into blocks of lines × 8 fields, and output to the orthogonal transformation circuit 53 and the determination circuit 58 in order. The orthogonal transformation circuit 53
The orthogonal transform is performed on each block output from the blocking circuit 52, and the obtained transform coefficients are output to the quantization circuit 56.
The quantization circuit 56 has a plurality of quantization tables having different quantization step widths. The determination circuit 58 is a block circuit
An optimal quantization table is selected by using the high-frequency component block output from 52, and the selected quantization table is output to the quantization circuit 56. The quantization circuit 56 quantizes the transform coefficient output from the orthogonal transform circuit 53 according to the quantization table selected by the determination circuit 58, and outputs the quantized transform coefficient to the encoding circuit 57. The coding circuit 57 is a quantization circuit
Encode 56 outputs.

In FIG. 36 showing the decoding side, reference numeral 91 denotes an input terminal for data encoded as described above. Input terminal 91
The decoding circuit 61 connected to the decoder performs an inverse conversion of the encoding circuit 57 and outputs decoded data to the inverse quantization circuit 62. The inverse quantization circuit 62 expands the output of the decoding circuit 61 in accordance with the quantization table, and converts the inverse-quantized transform coefficient into an inverse orthogonal transform circuit.
Output to 63. The inverse orthogonal transform circuit 63 performs an inverse orthogonal transform on the output of the inverse quantization circuit 62, and outputs the data after the inverse orthogonal transform to the memory 64. The memory 64 stores the output of the inverse orthogonal transform circuit 63 for eight fields. The subband combining circuit 65 combines the components output from the memory 64, and outputs the combined data to a digital / analog converter (hereinafter, referred to as a D / A converter) 66. The D / A converter 66 converts the digital image signal output from the sub-band synthesizing circuit 65 into an analog signal.

FIG. 37 shows an example of the configuration of the sub-band division circuit 51. The sub-band dividing circuit 51 includes an A / D converter 50
Horizontal LPF 511 that passes digital image signals output from
A horizontal HPF 512, horizontal 2: 1 sub-sampling circuits 513a and 513b for thinning the number of pixels in the horizontal direction by half, and a vertical LPF 51
4a, 514b, vertical HPFs 515a, 515b, and a vertical 2: 1 subsampling circuit 516a, 5 for thinning the number of pixels in the vertical direction by half.
16b, 516c, and 516d.

FIG. 38 shows an example of the configuration of the subband synthesis circuit 65. The sub-band synthesis circuit 65 includes a vertical 1: 2 interpolation circuit 651a, 651b, 651c, 651d that doubles the number of pixels in the vertical direction by interpolating 0, and a vertical LPF 652a, 652b.
A vertical HPF 653a, 653b, adders 654a, 654b, a horizontal 1: 2 interpolator 655a, 655b that doubles the number of pixels in the horizontal direction by interpolating 0, a horizontal LPF 656, and a horizontal LPF 656.
An HPF 657 and an adder 658 are provided.

Next, the operation will be described. The analog image signal input from the input terminal 90 is converted into a digital signal in the A / D converter 50. The sampling frequency at this time is, for example, 1 if the input signal is a luminance signal.
The frequency is 3.5 MHz, and the color difference signal is 6.75 MHz. Therefore, 1
The number of effective samples of a horizontal line is, for example, 70
4. The color difference signal becomes 352, and one field is composed of 262.5 lines. Among them, for example, 240 lines are output as an effective line as one field. The sub-band dividing circuit 51 divides each field of the digital image signal output from the A / D converter 50 into a plurality of frequency bands. For example, if the sampling frequency is 13.5MH
In the case of z, the frequency band is divided into four frequency bands of LL, LH, HL, and HH as shown in FIG.

The operation of the subband dividing circuit 51 will be described with reference to FIG. The input digital image signal is
After the band is limited in the horizontal LPF 511 having a frequency characteristic like 25, the number of pixels in the horizontal direction is halved in the horizontal 2: 1 sub-sampling circuit 513a. The digital image signal is also input to a horizontal HPF 512 having a frequency characteristic as shown in FIG. The output of the horizontal HPF 512 is thinned out in the horizontal 2: 1 sub-sampling circuit 513b by half the number of pixels in the horizontal direction. The outputs of the horizontal 2: 1 sub-sampling circuits 513a and 513b are band-limited by vertical LPFs 514a and 514b having frequency characteristics as shown in FIG.
In 16c, the number of pixels in the vertical direction is decimated by half.
On the other hand, the outputs of the horizontal 2: 1 sub-sampling circuits 513a and 513b are vertical HPFs having frequency characteristics as shown in FIG.
After being also input to 515a and 515b and band-limited, the vertical 2: 1 sub-sampling circuits 516b and 516d reduce the number of pixels in the vertical direction by half. Where vertical 2: 1
The output of the sub-sampling circuit 516a is the LL band of FIG. 21, and the output of the vertical 2: 1 sub-sampling circuit 516b is the LL band of FIG.
Vertical 2: 1 sub-sampling circuit 51
The output of 6c is the HL band of FIG. 21, and the output of the vertical 2: 1 sub-sampling circuit 516d is the HH band of FIG.

The four outputs of the sub-band dividing circuit 51 are formed into blocks of 8 pixels × 8 lines × 8 fields in the blocking circuit 52, and blocks of four components corresponding to the same position are continuously output. Is done. These blocks are input to the orthogonal transformation circuit 53, where they are subjected to orthogonal transformation. As the orthogonal transform, for example, a three-dimensional DCT is used.

On the other hand, 4 output from the blocking circuit 52
The LH band and the HL band of the one component block are input to the determination circuit 58. The determination circuit 58 detects whether or not there is a flat portion and a rapidly changing portion at the corresponding position on the screen from the pixel values of the blocks of the LH band and the HL band, and accordingly, the four corresponding to the same position are detected. Determine the quantization table for the block of components.

FIG. 39 shows an example of the configuration of the decision circuit 58. In the figure, reference numeral 92 denotes an input terminal for inputting an LH band and an HL band block among the four components output from the blocking circuit 52. The judgment circuit 58 further divides each input block into a plurality of sub-blocks, and calculates the maximum absolute value of the pixel value for each sub-block output from the sub-block formation circuit 581. An arithmetic unit 582 to be obtained, a minimum value detector 583 for obtaining the minimum value of the value output from the arithmetic unit 582, a maximum value detector 584 for obtaining the maximum value of the value output from the arithmetic unit 582, and a predetermined value α , 585a, 585b, 585c, 5
85d, a switch 586a for selecting and outputting the outputs of the constant generators 585a and 585b, a switch 586b for selecting and outputting the outputs of the constant generators 585c and 585d, and an output of the minimum value detector 583 and the switch 586a. A comparator 587a for comparing the output, a comparator 587b for comparing the output of the maximum value detector 584 with the output of the switch 586b, and a quantization circuit according to the output of the comparator 587a, 587b.
Control signal generator 58 for determining the quantization table in 56
8 and

Next, the operation will be described. The blocks of the LH band and the HL band are sequentially input to the input terminal 92 from the blocking circuit 52. Each input block is divided into 32 sub-blocks of 4 pixels × 4 lines by dividing the input block into four in each two-dimensional plane as shown in FIG. The arithmetic unit 582 obtains the maximum absolute value for each sub-block. That is, if the pixel value of the sub-block is represented by s (i, j) (i, j = 0,1,2,3), the arithmetic unit 582 obtains the following equation (7) for each sub-block and outputs I do. Max ｛| s (i, j) |; i, j = 0,1,2,3｝ ... (7)

Since the LH band and the HL band are considered to represent vertical and horizontal edges, respectively, the output of the arithmetic unit 582 expressed by the equation (7) indicates the degree of change of each sub-block. Is shown. That is, the arithmetic unit 582
Is very small, the sub-block is flat, and if it is large, the sub-block contains drastic changes.

The output of the arithmetic unit 582 is input to the minimum value detector 583 and the maximum value detector 584. The minimum value detector 583 is 1
The minimum value of the value of Expression (7) for 32 sub-blocks corresponding to the block is obtained. That is, the minimum value detector 583 outputs the minimum value of the 32 values continuously output from the arithmetic unit 582. On the other hand, the maximum value detector 584 obtains the maximum value of the value of the expression (7) for 32 sub-blocks corresponding to one block. That is, the maximum value detector 584 outputs the maximum value of the 32 values continuously output from the arithmetic unit 582. Therefore, when the output of the minimum value detector 583 is very small, it indicates that the block includes a flat portion, and when the output of the maximum value detector 584 is large, it indicates that the block includes a drastic change.

The constant generators 585a, 585b, 585c, 585d output predetermined values α, β, γ, δ, respectively. Switch 586a
Selects the output of the constant generator 585a when the value output from the minimum value detector 583 is the value of the LH band, and selects and outputs the output of the constant generator 585b when the value output from the minimum value detector 583 is the value of the HL band. . When the value output from the maximum value detector 584 is a value in the LH band, the switch 586b operates as a constant generator 58.
The output of 5c is selected, and when the value is in the HL band, the output of the constant generator 585d is selected and output. The comparator 587a outputs "1" when the output of the minimum value detector 583 is smaller than the output of the switch 586a, and outputs "0" otherwise. The output of the maximum value detector 584 is a switch 58
"1" is output when it is larger than the output of 6b, and "0" is output otherwise.

That is, the comparator 587a is the minimum value detector 583
When the value for the LH band is output from, if this value is smaller than the predetermined value α, it is determined that the corresponding block includes a flat portion in the vertical direction, and “1” is output. Outputs "0". Similarly, when the value for the HL band is output, the comparator 587a determines that this block includes a flat portion in the horizontal direction and outputs “1” when this value is smaller than the predetermined value β. In the case of, "0" is output. The predetermined values α and β are the LH band,
For the HL band sub-block, it is determined in advance from the range of values taken by equation (7).

The comparator 587b outputs LH from the maximum value detector 584.
When a value for a band is output, this value is equal to a predetermined value γ.
If it is larger, it is determined that the corresponding block contains a large change in the vertical direction, and “1” is output. Otherwise, “0” is output. Similarly, in the case of the HL band, when the output of the maximum value detector 584 is larger than the predetermined value δ, it is determined that this block includes a large change in the horizontal direction and “1” is determined.
And outputs "0" otherwise. The predetermined values γ and δ are determined in advance for the LH band and HL band sub-blocks from the range of values given by equation (7).

Since the LH band block and the HL band block are input to the input terminal 92 in this order, the comparator 587
a, 587b are, respectively, the judgments va,
After outputting vb, the judgments ha and hb in the horizontal direction are output.

The outputs of the comparators 587a and 587b are control signal generators.
Entered in 588. The control signal generator 588 uses the two values va, ha continuously output from the comparator 587a and the two values vb, hb continuously output from the comparator 587b to quantize the quantization circuit 56. Determine the table and output the control signal.
That is, the comparator 587a first outputs a determination va as to whether there is a flat portion in the vertical direction, and then outputs a determination ha as to whether there is a flat portion in the horizontal direction.
The comparator 587b first outputs a determination vb as to whether or not a large change in the vertical direction is included, and then outputs a determination hb as to whether or not it includes a large change in the horizontal direction. , A quantization table is determined. For example, when the quantization circuit 56 holds three types of quantization tables, the one with the smallest quantization step width is a high-rate quantization table, and the one with a medium quantization step width is a medium-rate quantization table. The quantization table and the one having the coarsest quantization step width will be referred to as a low-rate quantization table. At this time, if va = 0 and ha = 0, a low-rate quantization table is selected because there is no flat portion in which image quality deterioration is conspicuous in this block. On the other hand, va = 1 or ha
If = 1, there is a flat part in this block.
Then, (va = 1 or ha = 1) and (vb = 1 or hb
In the case of = 1), a high-rate quantization table is selected because the corresponding block includes a flat portion and a portion that changes rapidly, and image quality deterioration is conspicuous. Also, (va = 1 or ha =
1) If vb = 0 and hb = 0, the corresponding block has a flat portion, but does not have a portion that changes drastically. Therefore, a medium-rate quantization table is selected. The control signal generator 588 outputs which quantization table is selected to the quantization circuit 56. The above is the operation of the determination circuit 58.

The quantizing circuit 56 quantizes the transform coefficients output from the orthogonal transform circuit 53 according to the output of the decision circuit 58. At this time, the determination circuit 58 determines a quantization table from the LH band block and the HL band block corresponding to the same position, and outputs a control signal.
Applies the control signal output from the determination circuit 58 to blocks of four bands (LL, LH, HL, HH) corresponding to the same position. For example, if the output of the decision circuit 58 means “select a high-rate quantization table”, the transform coefficient of the LL band is the largest of the three types of quantization tables for the LL band. Is quantized using a fine quantization table, and the transform coefficient of the LH band is LH
Quantization is performed using the quantization table having the smallest quantization step width among the three types of quantization tables for bands,
HL band conversion coefficients are quantized using the quantization table with the smallest quantization step width among the three types of HL band quantization tables, and HH band conversion coefficients are three types of HH band quantization. The quantization is performed using the quantization table having the smallest quantization step width among the tables.

Here, the quantization tables for the respective bands may be the same or different from each other.
Further, these quantization tables may be uniform quantization tables having a fixed quantization step width, or may be non-linear quantization tables in which the quantization step width is not constant.

The transform coefficient quantized by the quantization circuit 56 is
The data is input to the encoding circuit 57 and encoded by one-dimensional scanning of a three-dimensional block and variable-length encoding. Encoding circuit 57
Also encodes and outputs an index indicating the quantization table used in the quantization circuit 56.

The decoding side follows a process completely opposite to the encoding.
That is, the data encoded as described above is input to the input terminal 91, and is decoded by the decoding circuit 61 into an index representing the original three-dimensional transform coefficient and the quantization table. The transform coefficient is inversely quantized in the inverse quantization circuit 62 according to an index representing a quantization table. The transform coefficient output from the inverse quantization circuit 62 is an inverse orthogonal transform circuit.
At 63, an inverse orthogonal transform is performed. Inverse orthogonal transformation 8
A block of pixels × 8 lines × 8 fields is stored in the memory 64. The memory 64 stores blocks for eight fields, and outputs each component one field at a time until the next eight fields are stored. The signals of the respective components output from the memory 64 are combined in the sub-band combining circuit 65.

The operation of the subband synthesizing circuit 65 will be described with reference to FIG. LL band, LH output from memory 64
The signals of the band, the HL band, and the HH band are each interpolated by 0 in the vertical 1: 2 interpolation circuit 651a, 651b, 651c, 651d, and the number of pixels in the vertical direction is doubled. Outputs of the vertical 1: 2 interpolation circuits 651a and 651c are band-limited in vertical LPFs 652a and 652b having frequency characteristics as shown in FIG. On the other hand, the outputs of the vertical 1: 2 interpolation circuits 651b and 651d are
Vertical HPF 653a, 6 with frequency characteristics as shown in Fig. 24
Bandwidth is limited at 53b. The output of the vertical LPF 652a and the output of the vertical HPF 653a are added in an adder 654a.
The output of the vertical LPF 652b and the output of the vertical HPF 653b are added in an adder 654b. The outputs of the adders 654a and 654b are interpolated with 0 in horizontal 1: 2 interpolation circuits 655a and 655b, respectively, and the number of pixels in the horizontal direction is doubled. The output of the horizontal 1: 2 interpolation circuit 655a is band-limited by a horizontal LPF 656 having frequency characteristics as shown in FIG. The output of the horizontal 1: 2 interpolation circuit 655b is band-limited by a horizontal HPF 657 having frequency characteristics as shown in FIG. Horizontal LPF 656
The output of the horizontal HPF 657 is added to the output of the horizontal HPF 657 and output.

As described above, the digital image signal synthesized by the sub-band synthesizing circuit 65 is converted into an analog signal by the D / A converter 66 and output.

In the decision circuit 58, the control signal generator 588
Is a signal va indicating whether or not a vertical flat portion is included.
And a signal ha indicating whether or not a flat portion is included in the horizontal direction.
And a signal vb indicating whether there is a large change in the vertical direction.
And, the quantization table is selected from the signal hb indicating whether or not there is a large change in the horizontal direction, especially when the input image signal is an interlaced scanning method, there are very few flat portions only in the vertical direction, Also, because it is not visually noticeable,
The presence of a flat portion may be determined based on whether or not a flat portion is included in the horizontal direction. That is, using only three signals, ha, vb, and hb, for example, ha = 1 and (vb = 1 or hb =
If 1), it is determined that a flat portion and a portion with a great change are mixed in the corresponding block, and a high-rate quantization table is selected. If ha = 1 and (vb = 0 and hb = 0), the corresponding block Has a flat portion but no sharply changing portion, so a medium-rate quantization table is selected. If ha = 0, it is determined that there is no flat portion, and a low-rate quantization table is selected.

In the above-described example, the arithmetic unit 582 in the determination circuit 58 obtains the maximum value of the absolute value of the pixel value in each sub-block. It works. Another configuration example of the judgment circuit 58 is shown in FIG. In the figure, the same reference numerals as those in FIG. 39 indicate the same parts. 93a is the input terminal of the LH band, 93b is the input terminal of the HL band, 589a and 589b are absolute value calculation circuits, 590 is an adder, 591 is the maximum value detector in the sub block that detects the maximum value in the sub block, 592 Is a control signal generator.

Next, the operation will be described. The LH band block output from the blocking circuit 52 has an input terminal 93a.
And the HL band block is input from an input terminal 93b, and is converted into an absolute value in absolute value calculation circuits 589a and 589b, respectively. In the absolute value calculation circuits 589a and 589b, the LH band block and the HL band block that have been converted into absolute values are added in an adder 590, and further divided in a subblock forming circuit 581 into a plurality of subblocks. The sub-block dividing circuit 581 divides one block into 32 sub-blocks as shown in FIG. The maximum value detector 591 in the sub-block outputs the maximum value in each sub-block. That is, the pixel value of each sub-block output from the sub-block forming circuit 581 is represented by s (i, j) (i, j =
(0,1,2,3), the maximum value detector 591 in the sub-block outputs the following equation (8) for each sub-block. Max ｛s (i, j); i, j = 0,1,2,3… (8)

Since the output of the adder 590 indicates a horizontal or vertical edge, if the value of the maximum value detector 591 in the sub-block, that is, the value of the expression (8) is very small, the sub-block is not affected. Indicates a flat portion, and when large, indicates that the sub-block includes a drastic change.

The output of the maximum value detector 591 in the sub-block is input to the minimum value detector 583 and the maximum value detector 584. The minimum value detector 583 finds the minimum value of the value of the expression (8) for 32 sub-blocks corresponding to one block. That is, the minimum value detector 583 outputs the minimum value of the 32 values continuously output from the maximum value detector 591 in the sub-block. On the other hand, the maximum value detector 584 corresponds to one block
The maximum value of the value of the expression (8) is obtained for 32 sub-blocks. That is, the maximum value detector 584 outputs the maximum value of the 32 values continuously output from the maximum value detector 591 in the sub-block. Therefore, when the output of the minimum value detector 583 is very small, it indicates that the block includes a flat portion, and when the output of the maximum value detector 584 is large, it indicates that the block includes a sharp change. I have.

The control signal generator 592 includes the minimum value detector
According to the output of 583 and the output of the maximum value detector 584, it is determined which quantization table to select. For example, a relatively small value α ′ and a relatively large value β ′ are determined in advance, and when the output d1 of the minimum value detector 583 satisfies d1 ≧ α ′, a low-rate quantization table is selected, and d1 < If α ′ and the output d2 of the maximum value detector 584 satisfies d2 ≦ β ′, select a medium rate quantization table. If d1 <α ′ and d2> β ′, select a high rate quantization table. Select a table. Or figure
The quantization table may be switched according to a two-dimensional chart such as 42. In this case, even if there are three or more quantization tables, it can be easily applied.

In the above-described embodiment, the quantization table is selected by using the LH band and the HL band among the four components divided into sub-bands.
A quantization table can be selected using the HL band and the HH band. FIG. 43 shows a configuration example in this case. In the figure, the same reference numerals as those in FIG. 35 indicate the same parts. 58a is the LH band, HL band, and HH band output from the blocking circuit 52.
This is a determination circuit that determines a quantization table using a block of a band.

FIG. 44 shows a configuration example of the decision circuit 58a. In the figure, the same reference numerals as those in FIG. 39 denote the same components. The LH band, HL band, and HH band blocks output from the blocking circuit 52 are sequentially input from the input terminal 92a, and are divided into 32 sub-blocks by the sub-blocking circuit 581. The arithmetic unit 582 calculates the maximum value of the absolute value of the pixel value in each sub-block. The output of the arithmetic unit 582 is input to the minimum value detector 583 and the maximum value detector 584. The minimum value detector 583 obtains the minimum value of the output of the arithmetic unit 582 for 32 sub-blocks corresponding to one block. The maximum value detector 584 obtains the maximum value of the output of the arithmetic unit 582 for 32 sub-blocks corresponding to one block. The operations of the sub-block circuit 581, the arithmetic unit 582, the minimum value detector 583, and the maximum value detector 584 are the same as those in the above-described embodiment.

Constant generators 585a, 585b, 585e, 585c, 585d, 585
f outputs predetermined values α, β, ε, γ, δ, and θ, respectively. A switch 593a selects a predetermined value α of the constant generator 585a when the value output from the minimum value detector 583 is the value of the LH band, and selects a predetermined value α of the constant generator 585b when the value output from the minimum value detector 583 is the value of the HL band. The value β is selected, and when the value is in the HH band, the predetermined value ε of the constant generator 585e is selected and output.

593b is also a switch and has a maximum value detector 584.
When the value output from is the value of the LH band, the predetermined value γ of the constant generator 585c is selected, and when the value output is the value of the HL band, the predetermined value δ of the constant generator 585d is selected. In some cases, a predetermined value θ of the constant generator 585f is selected and output.

The comparator 587a outputs "1" when the output of the minimum value detector 583 is smaller than the output of the switch 593a, and outputs "0" otherwise. The comparator 587b outputs "1" when the output of the maximum value detector 584 is larger than the output of the switch 593b, and outputs "0" otherwise.
That is, similarly to the above-described embodiment, the comparator 587a outputs "1" when the corresponding block has a flat portion, and the comparator 587b outputs "1" when there is a drastic change in the corresponding block.

Each block of the LH band, the HL band, and the HH band is input from the input terminal 92a in this order.
587a and 587b first output the judgments ha and hb for the LH band block, then output the judgments va and vb for the HL band block, and finally output the judgments da and db for the HH band block I do. These comparators 587a, 587b
The three consecutive outputs are input to the control signal generator 588a, and the quantization table is determined.

For example, when the quantization circuit 56 holds three quantization tables, a quantization table with a small quantization step width is replaced with a high-rate quantization table and a quantization table with a medium quantization step width. Is called a medium-rate quantization table, and a quantization table with a coarse quantization step width is called a low-rate quantization table. At this time,
If ha = 0 and va = 0, it is determined that there is no flat portion, and a low-rate quantization table is selected. ha = 1 or va = 1
In the case of, since there is a flat portion, hb = 1 or vb = 1 or
If db = 1, there is a drastic change, so a high-rate quantization table is selected. Otherwise, a medium-rate quantization table is selected.

Although the quantization table is selected using the LH band, HL band, and HH band in the above-described embodiment, the same configuration can be applied to the case where the determination is performed using only one of these components. .

Further, in each of the above embodiments, the subblock dividing circuit 581 divides one block into 32 subblocks as shown in FIG. 41, but the size of the subblock is limited to 4 × 4. Not something, for example, horizontal 4
4 × 2 pixels, 2 pixels in vertical direction, 2 fields in time direction
It may be configured as a × 2 three-dimensional sub-block, and the size may be set according to the size of the block or the hardware configuration. Also, the size of one block as a unit of the orthogonal transformation does not need to be limited to 8 × 8 × 8, and the same effect can be obtained for three-dimensional blocks and two-dimensional blocks of arbitrary sizes.

Further, in each of the above embodiments, the case where the input signal is divided into four sub-bands has been described. For example, when the input signal is divided into two only in the horizontal direction,
The same applies to the case where the low-frequency component is divided into seven bands by repeatedly dividing the sub-band into four, and the quantization table can be selected using the high-frequency component. That is, in general, even when subbands are divided into a number of bands, the same effect can be obtained by detecting whether there is a flat portion and a portion that changes rapidly in a block using a high-frequency component, and selecting a quantization table. Can be obtained.

Next, still another embodiment of the present invention will be described. In this example, one of an odd-numbered field and an even-numbered field is passed through an odd-tap vertical filter and the other is passed through an even-tap vertical filter in an input interlaced scanning digital image signal.

FIG. 45 is a block diagram showing the internal configuration of subband dividing circuit 5 in such an example. In the figure, portions denoted by the same reference numerals as those in FIG. 19 indicate the same members, and therefore, description thereof will be omitted. In the figure, reference numerals 28, 29, and 30 denote a Y signal sub-band division circuit, an RY signal sub-band division circuit, and a BY signal sub-band division circuit in the sub-band division circuit 5, respectively. It is. The Y signal sub-band dividing circuit shown below
The configuration of 28 will be described. The vertical LPF connected to the input terminal 21 is divided into a vertical LPF 241a and a vertical LPF 241b. A switch 39 is provided between the vertical LPFs 241a and 241b and the vertical 2: 1 sub-sampling circuit 243a. The output terminal of the vertical LPF 241a is connected to one terminal a of the switch 39, and the output terminal of the vertical LPF 241b is connected to the other terminal b of the switch 39. The vertical HPF connected to the input terminal 21 is the vertical HPF 242a and the vertical HPF
242b. Vertical HPF 242a, Vertical HPF 242b
A switch 40 is provided between the switch 40 and the vertical 2: 1 sub-sampling circuit 243b. One terminal a of the switch 40
Is connected to the output terminal of the vertical HPF 242a, and the other terminal b of the switch 40 is connected to the output terminal of the vertical HPF 242b.

FIG. 46 is a block diagram showing an internal configuration of subband combining circuit 14 in such an example. In the figure, portions denoted by the same reference numerals as those in FIG. 20 indicate the same members, and therefore, description thereof will be omitted. In the figure, reference numerals 36, 37, and 38 denote a Y signal subband synthesis circuit, an RY signal subband synthesis circuit, and a BY signal subband synthesis circuit in the subband synthesis circuit 14, respectively. It is. The Y signal sub-band synthesis circuit shown below
The configuration of 36 will be described. The vertical LPF connected to the vertical 1: 2 interpolation circuit 336a is divided into a vertical LPF 337a and a vertical LPF 337b. A switch 48 is provided between the vertical LPFs 337a and 337b and the subtractor 339. The output terminal of the vertical LPF 337a is connected to one terminal a of the switch 48, and the output terminal of the vertical LPF 337b is connected to the other terminal b of the switch 48. A vertical 1: 2 interpolation circuit
The vertical HPF connected to 336b is vertical HPF 338a and vertical HPF
338b. Vertical HPF 338a, Vertical HPF 338b
A switch 49 is provided between and the subtractor 339. The output terminal of the vertical HPF 338a is connected to one terminal a of the switch 49, and the vertical HPF 338a is connected to the other terminal b of the switch 49.
The output terminal of F 338b is connected.

Next, the operation will be described. Input terminal 21
Is input to the Y signal sub-band dividing circuit 28.
Is divided into four bands. Hereinafter, the operation of the Y signal sub-band division circuit 28 will be described in detail.
The Y signal input from the input terminal 21 is a vertical LPF 241a, 24
Passed through 1b. The switch 39 switches between the terminal a and the terminal b for each field and outputs the result. For example, the terminal a is selected for an odd field, and the terminal b is selected for an even field. Therefore, in the case of the odd field, the Y signal that has passed through the vertical LPF 241a of the odd tap is output, and in the case of the even field, the Y signal that has passed through the vertical LPF 241b of the even tap is output. Figure 47
As shown in (1), the spatial sampling positions of the odd field and the even field are the same. The subsequent operation is the same as the example shown in FIG. Also, input terminal 21
Is also passed through the vertical HPFs 242a and 242b. The switch 40 switches between the terminal a and the terminal b for each field and outputs the result. For example, the terminal a is selected for an odd field, and the terminal b is selected for an even field. Therefore, in the case of the odd field, the Y signal that has passed through the vertical HPF 242a of the odd tap is output, and in the case of the even field, the Y signal that has passed through the vertical HPF 242b of the even tap is output. As shown in FIG. 47, the odd sampling field and the even sampling field have the same spatial sampling position. Note that the subsequent operations are
This is the same as the example shown in FIG. The RY signal sub-band dividing circuit 29 and the BY signal sub-band dividing circuit 30 perform the same operation.

FIG. 48 shows the memory 27 in such an example.
, Four pixels in the horizontal direction, four pixels in the vertical direction, and eight fields in the time direction are collectively output as a 4 × 4 × 8 three-dimensional block. Each component has the same spatial sampling position in both the odd field and the even field. Therefore, even if the odd and even fields are mixed to form a three-dimensional block, the pixel in the block is Are arranged in a grid pattern.

Next, the operation on the reproducing side will be described. L
The operation up to the vertical 1: 2 interpolation circuit 336a in the L and HL bands is the same as the example shown in FIG. The output from the vertical 1: 2 interpolation circuit 336a is input to a vertical LPF 337a having odd taps and a vertical LPF 337b having even taps. Then, in the case of an odd field, the terminal a of the switch 48 is selected, and the signal passing through the vertical LPF 337a of the odd tap is output to the subtractor 339. On the other hand, in the case of the even field, the terminal b of the switch 48 is selected, and the signal passing through the vertical LPF 337b of the even tap is output to the subtractor 339. A vertical 1: 2 interpolation circuit 33 in the LH and HH bands is used.
Since the operations up to 6b are the same as those in the example shown in FIG. 20, the description will be omitted. The output from the vertical 1: 2 interpolation circuit 336b is input to a vertical HPF 338a of an odd tap and a vertical HPF 338b of an even tap. And when it is an odd field, switch
The terminal a of 49 is selected, and the signal passing through the odd-numbered tap vertical HPF 338a is output to the subtractor 339. On the other hand, in the case of the even field, the terminal b of the switch 49 is selected, and the signal passing through the vertical HPF 338b of the even tap is output to the subtractor 339. In the subtractor 339, the output of the switch 49 is subtracted from the output of the switch 48, and the original Y signal of one field is output. The RY signal sub-band synthesis circuit 3
7. The operation of the BY signal sub-band synthesizing circuit 38 is the same as the operation of the Y signal sub-band synthesizing circuit 36 described above.

In such an example, after the spatial sampling positions of the odd and even fields are aligned, the two fields are spring-loaded to form a three-dimensional block and perform three-dimensional orthogonal transformation. Information compression can be significantly performed on image signals.

Next, still another embodiment of the present invention will be described. This embodiment is an example in which the way of weighting is changed in each divided subband. FIG. 49 is a block diagram showing the configuration on the code side in such an embodiment. In the figure, portions denoted by the same reference numerals as in FIG. 35 indicate the same members.

The digital image signal is frequency-separated in two stages in the horizontal and vertical directions by the sub-band dividing circuit 51 to become four types of sub-band signals. The obtained sub-band signals are output to the blocking circuits 52a, 52b, 52c, 52d, respectively. The blocking circuits 52a, 52b, 52c, and 52d block each of a plurality of pixels, and output the obtained blocks to the orthogonal transform circuits 53a, 53b, 53c, and 53d. Orthogonal transformation circuit 53
a, 53b, 53c, 53d apply, for example, a three-dimensional DCT to each block, and convert the obtained transform coefficients into weighting circuits 54a, 54a, 5d.
Output to 4b, 54c, 54d. Waiting circuits 54a, 54b,
54c and 54d are based on the frequency characteristics of each subband,
The transform coefficients output from the orthogonal transform circuits 53a, 53b, 53c, and 53d are weighted, and the weighted transform coefficients are output to the variable-length coding circuits 55a, 55b, 55c, and 55d. Each of the variable length coding circuits 55a, 55b, 55c, 55d performs variable length coding on the output of the weighting circuits 54a, 54b, 54c, 54d.

Next, the operation will be described.

Each sub-band signal output from the sub-band dividing circuit 51 is divided into a plurality of pixels, for example, every 8 pixels × 8 lines × 8 fields in the blocking circuits 52a, 52b, 52c, 52d. . Then, blocks X ₀ (i, j, k) (i: horizontal direction, j: vertical direction, k: time direction, i, j, k =
0, 1,..., 7) are divided into respective blocking circuits 52a, 52b, 52c,
The signal is output from 52d to the orthogonal transform circuits 53a, 53b, 53c, 53d. Each of these blocks includes an orthogonal transformation circuit 53a, 53b, 5
In 3c and 53d, a three-dimensional DCT is performed, and 8 × 8 × 8 transform coefficients X (i, j, k) (i, j, k = 0, 1, 1)
.., 7) are weighting circuits 54a, 54b, 54c, 54d.
Output to The transform coefficients output from the orthogonal transform circuits 53a, 53b, 53c, 53d to the respective weighting circuits 54a, 54b, 54c, 54d are subjected to weighting. The waiting circuit 54a corresponds to the LL band, and the waiting circuit 54b
Corresponds to the LH band, the waiting circuit 54c corresponds to the HL band, and the waiting circuit 54d corresponds to the HH band. The output of each weighting circuit 54a, 54b, 54c, 54d is output to the corresponding variable length coding circuit 55a, 55b, 55c, 55d.
Is subjected to variable length coding.

Here, each of the weighting circuits 54a, 54b,
The operation in 54c and 54d and the size of the waiting will be described.

From the results of Tables 2 and 3 described above, in the case of the LL band, when viewed from a block in a two-dimensional plane, (i, j) =
Since power is concentrated around (0, 0), in the LL band, as shown in FIG. 50 (a), a large weighting is applied to the upper left transform coefficient, which is a horizontal and vertical low-order sequence, It is suitable to apply a small weighting to the transform coefficients that are higher-order sequences.

From the results in Tables 6 and 7 described above, in the LH band, when viewed from a block in a two-dimensional plane, (i, j) =
Since power is concentrated around (0, 7), in the LH band, as shown in FIG. 50 (b), the lower left sequence where the horizontal direction is a low-order sequence and the vertical direction is a higher-order sequence It is suitable to apply a large weighting to the transform coefficients and apply a small weighting to the transform coefficients having a high-order sequence in the horizontal direction and a low-order sequence in the vertical direction.

From the results of Tables 4 and 5 described above, in the HL band, when viewed in blocks in a two-dimensional plane, (i, j) =
Since power is concentrated around (7,0), in the HL band, as shown in FIG. 50 (c), the upper right is the higher-order sequence in the horizontal direction and the lower-order sequence in the vertical direction. It is suitable to apply a large weighting to the transform coefficients and apply a small weighting to the transform coefficients having a low-order sequence in the horizontal direction and a high-order sequence in the vertical direction.

From the results in Tables 8 and 9 described above, in the HH band, when viewed from a block in a two-dimensional plane, (i, j) =
Since power is concentrated around (7, 7), in the HH band, as shown in FIG. 50 (d), a large weighting is applied to the lower right transform coefficient in which the horizontal and vertical directions are both higher-order sequences. It is suitable to apply a small weighting to a transform coefficient having a low-order sequence in both the horizontal and vertical directions.

Looking at the numerical values in each table in the time direction, the power is concentrated on the plane of k = 0 in any band, and the power decreases as k becomes higher.
It is suitable to apply a large weighting to a low-order transform coefficient with k and a small weighting to a high-order k transform coefficient. Furthermore, since human vision is dull to high spatial frequencies, a large weighting is applied to the LL band where low spatial frequencies are concentrated, and a small weighting is applied to the HH band where high spatial frequencies are concentrated. Are suitable.

In consideration of the above, and considering that the weighting for the four sub-bands is smoothly connected in the horizontal and vertical frequency directions, the weighting coefficient in each sub-band is as follows.

The weighting coefficient W _LL for the LL band
(I, j, k) may be set as in the following equation.

[0201]

[Equation 14]

Similarly, weighting coefficients W _LH (i, j, k) _, for the LH band, HL band, and HH band
W _HL (i, j, k) _and W _HH (i, j, k) may be set as follows.

[0203]

(Equation 15)

[0204]

(Equation 16)

[0205]

[Equation 17]

By performing the above-mentioned weighting, effective weighting can be applied to each of the sub-bands.

In the above embodiment, a three-dimensional DCT is used as the orthogonal transform, but a two-dimensional DCT may be used. In this case, the RMS of the transform coefficient of the two-dimensional DCT shows the same tendency as the RMS in the two-dimensional plane of the three-dimensional DCT.
Weighting coefficients W _LL (i, j) _, W _LH (i, j) _, W _HL (i, j) for LH band, HL band, and HH band
_, W _HH (i, j) may be set as in the following equations.

[0208]

(Equation 18)

[0209]

[Equation 19]

[0210]

(Equation 20)

[0211]

(Equation 21)

Hereinafter, the band division / combination method of the present invention will be described. FIGS. 51 and 52 show a sub-band dividing / combining circuit when the horizontal frequency band is divided into two. In FIG. 51 showing the configuration of the sub-band division circuit, reference numeral 70 denotes an input terminal for a digital image signal. The sub-band division circuit is
A wrapped image adding circuit 71 that wraps the image input from the end point at the end point and adds the wrapped image, and an even tap horizontal that limits the horizontal frequency of the image output from the wrapped image adding circuit 71
LPF 72, horizontal HPF 73 and horizontal LPF 72, horizontal HPF 7 respectively
Horizontal 2: thinning out the number of horizontal pixels of the output of 3 to 2: 1
And one sub-sampling circuit 74a, 74b. 75a and 75b are output terminals of the horizontal 2: 1 sub-sampling circuits 74a and 74b, respectively.

In FIG. 52 showing the configuration of the sub-band synthesizing circuit, reference numerals 76a and 76b denote input terminals of each band divided as described above. The subband synthesizing circuit adds a wrapped image adding circuit 77 that wraps the image input from the input terminal 76a at the end point and adds the wrapped image to the image input from the input terminal 76b. Return image addition circuit 78 and return image addition circuit 7
The horizontal 1: 2 interpolation circuits 79a and 79b for interpolating 0 in the image output from 7, 78 to double the number of pixels in the horizontal direction and the horizontal frequency of the output of the horizontal 1: 2 interpolation circuit 79a are limited. A horizontal LPF 80 with an even tap, a horizontal HPF 81 with an even tap that limits the horizontal frequency of the output of the horizontal 1: 2 interpolation circuit 79b, and a calculator that subtracts the output of the horizontal HPF 81 from the output of the horizontal LPF 80
82 and. Reference numeral 83 denotes an output terminal of the calculator 82.

Next, the operation will be described. FIG. 53 is a conceptual diagram of an output signal from each member for explaining this operation. In the image signal input from the input terminal 70, a folded image is added to an end point of the image in the folded image adding circuit 71. For example, when the image signal is an image of 704 pixels × 240 lines, the signal x (n) (n = 1,
., 704), a folded image is added to the end point of the image as shown in the following equation (9) (FIG. 53 (b)).

[0215]

The output of the folded image adding circuit 71 is horizontal.
Input to LPF 72 and horizontal HPF 73. The horizontal LPF 72 is a low-pass filter having an even-tap linear phase, and its coefficient h ₁ (n) (n = 0,..., N1 −1; N1 is an even number) is h ₁ (n) = h ₁ (N _{1 -n-1) n = 0} , ..., a _{(n 1/2) -1 ...} (10). That is, the horizontal LPF 72 outputs the signal x of each line.
In response to (n), a signal represented by the following equation (11) is output (FIG. 53).
(C)).

[0219]

(Equation 22)

On the other hand, the horizontal HPF 73 is a high-pass filter having a linear phase of even taps, and its coefficient h ₂ (n) (n = 0,
..., N ₂ -1; N ₂ is an even number) _{is, h 2 (n) = -h} 2 (N 2 -n-1) n = 0, ..., a _{(N 2/2) -1 ...} (12) Become. That is, the horizontal HPF 73 outputs the signal x of each line.
In response to (n), the signal shown in the following equation (13) is output (see FIG. 53).
(C)).

[0219]

(Equation 23)

The outputs of the horizontal LPF 72 and the horizontal HPF 73 are output from the output terminals 75a and 75b, respectively, in the horizontal 2: 1 sub-sampling circuits 74a and 74b, in which the number of pixels in the horizontal direction is thinned to half. It is output (FIG. 53 (d)). The above is the operation of the subband division circuit. The two signals output by sub-band division are respectively coded using predictive coding, orthogonal transform, and the like. After decoding these signals, the decoding side performs subband synthesis.

The signals of the respective sub-bands divided as described above are input to the input terminals 76a and 76b. That is,
Assuming that no distortion occurs in the encoding / decoding after the subband division, the signals input from the input terminals 76a and 76b are the same as the signals output from the output terminals 75a and 75b, respectively. Therefore, each line of the image input from the input terminals 76a and 76b is y ₁ (n), y ₂ (n) (n = 1,
3,5,…, 703). The folded image addition circuit 77
The folded image is added to the end point of the image input from the input terminal 76a. That is, the following equation (14) is used.

[0222]

The subband dividing circuit does not output the added folded image portion. However, if the horizontal LPF 72 also calculates the folded image portion, the following expression is obtained from the expressions (9) and (11). become.

[0224]

(Equation 24)

[0225]

(Equation 25)

Therefore, if a folded image is added in the folded image adding circuit 77 as shown in Expression (14), the subband dividing circuit can transmit the added folded image portion without transmitting the portion of the added folded image. , Can be accurately restored. This folded image addition circuit 77
Is interpolated with 0 in the horizontal 1: 2 interpolation circuit 79a, and the number of pixels in the horizontal direction is doubled. That is,
Each line of the image output from the horizontal 1: 2 interpolation circuit 79a is as follows, as shown in FIG.

[0227]

The output of the horizontal 1: 2 interpolation circuit 79a is the horizontal LP
Entered in F80. The horizontal LPF 80 is an even-tap low-pass filter whose coefficient is represented by g ₁ (n) (n = 0,..., N ₃ −
1: N ₃ is an even number) to the horizontal 1: 2 each line y ₁ of image output from the interpolation circuit 79a '(n) (n = -703, -702,
, 1407), the signal shown in the following equation is output (FIG. 53).
(G)).

[0229]

(Equation 26)

On the other hand, the folded image adding circuit 78 adds a value obtained by multiplying the folded image by -1 to the end point of the input image from the input terminal 76b. That is, the input terminal 76
Each line y ₂ (n) of the image input from b (n = 1,3,5, ...,
703), the following equation (15) is used.

[0231]

Here, as in the case of the folded image adding circuit 77, if it is assumed that the horizontal HPF 73 also calculates the folded image portion in the sub-band dividing circuit, the following equation (9) is obtained.
From equation (13), the following is obtained.

[0233]

[Equation 27]

[0234]

[Equation 28]

Therefore, if the aliased image is added in the aliased image adding circuit 78 as shown in the expression (15), the subband dividing circuit does not need to transmit the aliased image portion, and the subband synthesizing circuit can accurately calculate the aliased image. Can be restored. The output of the folded image adding circuit 78 is interpolated with 0 in a horizontal 1: 2 interpolation circuit 79b to double the number of pixels in the horizontal direction. As shown in FIG. , As follows.

[0236]

The output of the horizontal 1: 2 interpolation circuit 79b is the horizontal HP
Entered in F81. The horizontal HPF 81 is an even-tap high-pass filter whose coefficient is represented by g ₂ (n) (n = 0,..., N ₄ −
1: N ₄ is an even number) to the horizontal 1: 2 interpolation circuit each line y ₂ of image output from 79b '(n) (n = -703, -702,
, 1407), the following signals are output.

[0238]

(Equation 29)

The arithmetic unit 82 calculates the horizontal HP from the output of the horizontal LPF 80.
The output of F81 is subtracted and output from the output terminal 83.

Next, in order to specifically show the effect of this embodiment, the horizontal LPF 72, horizontal HPF 73, horizontal LPF
Suppose that the F80 and the horizontal HPF 81 are the 16-tap filters in Table 1 and the first line x (n) (n =
1, ..., 704) becomes x (1) = 16, x (2) = 120, x (3) = 130, x (4) =
140, x (5) = 150, x (6) = 160, x (7) = 170, x
(8) = 180, x (9) = 190, x (10) = 200, x (n) = 200,
(n = 11, 12, ..., 704) is considered. At this time, the signal output from the output terminal 75a _{is, y 1 (1) = 59.6} , y 1 (3) = 144.7, y 1 (5) = 152.
6, y ₁ (7) = 175.1, y ₁ (9) = 195.3, y ₁ (11) = 20
0.5, y ₁ (13) = 199.8, y ₁ (15) = 200.0, y ₁ (17) =
199.9, ... The signal output from the output terminal 75b _{is, y 2 (1) = -37.7} , y 2 (3) = 8.2, y 2 (5) =
−3.1, y ₂ (7) = 1.0, y ₂ (9) = −1.6, y ₂
(11) = 0.1, y ₂ (13) = − 0.1, y ₂ (15) = 0.1, y ₂
(17) = 0.0,... The output terminal 75a, the output of 75b, synthetic side of the input terminal 76a, the input to 76 b, the output of the horizontal LPF 80 _{is, x 1 (1) = 46.3} , x 1 (2) = 77.9, x 1 (3) = 129.
4, x ₁ (4) = 156.1, x ₁ (5) = 151.0, x ₁ (6) = 15
3.2, x ₁ (7) = 169.1, x ₁ (8) = 182.0, x ₁ (9) =
191.3, ..., and the output of the horizontal HPF 81 _{is, x 2 (1) = 30.5} , x 2 (2) = -42.0, x 2 (3) = -
0.5, x ₂ (4) = 16.3, x ₂ (5) = 1.2, x ₂ (6)
= −6.7, x ₂ (7) = − 0.8, x ₂ (8) = 2.1, x ₂
(9) = 1.4, ...

Therefore, the output X (n) of the arithmetic unit 82 is obtained by X (1) = 16, X (2) = 120, X (3) = 130, X (4)
= 140, X (5) = 150, X (6) = 160, X (7) = 170,
X (8) = 180, X (9) = 190,..., X (n) = x (n) even near the end point of the image, and the original image can be accurately restored.

In the above embodiment, the folded image is added before the 0 interpolation.
The same configuration can be applied to the case where a folded image is added after interpolation of.

Further, in the above embodiment, the case where the filter is an even-numbered tap has been described, but a filter having an odd-numbered tap can be similarly configured. FIGS. 54 and 55 show configuration examples in the case of using an odd-tap filter. As shown in FIG. 56B, the image input from the input terminal 70 is connected to an image obtained by reducing one pixel from the image turned back to the end point, as shown in FIG.
6) It becomes like the formula.

[0244]

The horizontal LPF 85 is a low-pass filter having odd-numbered taps having a linear phase, and has coefficients h ₁ (n) (n = 0,..., N
₁ -1; N ₁ is an odd number), then h ₁ (n) = h ₁ (N ₁ -n-1) n = 0,..., (N ₁ -1) / 2. That is, the horizontal LPF 85 is a signal x (n) of each line.
Then, the following equation (17) is output.

[0246]

[Equation 30]

The horizontal 2: 1 sub-sampling circuit 74a thins the number of horizontal pixels of the output of the horizontal LPF 85 by half, and
₁ (1), y ₁ (3), y ₁ (5)... Are output from the output terminal 75a (FIG. 56 (c)). On the other hand, the horizontal HPF 86 is a high-pass filter having an odd number of taps having a linear phase, and its coefficient is represented by h _2.
(n) (n = 0, ..., N ₂ -1; N ₂ is an odd number), h ₂ (n) = h ₂ (N ₂ -n-1) n = 0, ..., (N ₂ -1 ) / 2. That is, the horizontal HPF 86 outputs the following equation (18) for the signal x (n) of each line.

[0248]

(Equation 31)

The horizontal 2: 1 sub-sampling circuit 74c has a horizontal
The number of pixels in the horizontal direction of the output of the HPF 86 is halved, and y ₂
(2), y ₂ (4), y ₂ (6), etc. are output from the output terminal 75b (FIG. 56 (d)).

In the subband synthesizing circuit, the signals of the respective bands output from the output terminals 75a and 75b on the division side are input from the input terminals 76a and 76b, respectively. That is, the lines of the image input from the input terminals 76a and 76b are respectively y ₁ (n) (n = 1, 3, 5,..., 703) and y ₂ (n) (n = 2, 4, 6 ,
..., 704). The horizontal 1: 2 interpolation circuits 79a and 79c are
Each of the images input from the input terminals 76a and 76b has 0
To double the number of pixels in the horizontal direction. Therefore, the outputs y ₁ '(n) and y ₂ ' (n) of the horizontal 1: 2 interpolation circuits 79a and 79c
Are as follows.

[0251]

The folded image addition circuits 87a and 87b connect the end points of the images output from the horizontal 1: 2 interpolation circuits 79a and 79c, respectively, with the reduced image reduced by one pixel. That is, the following equations (19) and (20) are used.
(FIGS. 56 (e) and (f)).

[0253]

The sub-band division circuit does not transmit the added folded image portion, but if the horizontal LPF 85, the horizontal HPF
Assuming that F 86 also calculates the folded image part,
From equations (16) and (17), the following is obtained.

[0255]

(Equation 32)

Similarly, y ₁ (704 + n) = y ₁ (704-n) (n = 1, 2,..., 704). Further, the following expression is similarly established from the expressions (16) and (18).

[0257]

Therefore, in the folded image adding circuits 87a and 87b, if the folded image is added to the end point of the image as in the equations (19) and (20), the folded image that is not transmitted and added in the subband dividing circuit is obtained. Parts can also be accurately restored. Outputs of the folded image adding circuits 87a and 87b are input to a horizontal LPF 88 and a horizontal HPF 89, respectively. The horizontal LPF 88 is an odd-tap low-pass filter whose coefficient is g ₁ (n) (n = 0,..., N ₃ -1; N ₃ is an odd number).
Then, the following equation is output for each line y ₁ ′ (n) of the image output from the folded image adding circuit 87a.

[0259]

[Equation 33]

On the other hand, the horizontal HPF 89 is a high-pass filter having odd taps, and its coefficient is represented by g ₂ (n) (n = 0,..., N ⁴ −).
1; N ⁴ is an odd number).
The following equation is output for each line y ₂ ′ (n) of the output image.

[0261]

(Equation 34)

The computing unit 82 outputs the output of the horizontal LPF 88 and the horizontal HPF
The output of the output terminal 83 is added to the output of the output terminal 89 and output from the output terminal 83.

In the above-described embodiment, the configuration is such that a folded image is added after interpolating 0. However, the same configuration can be applied when interpolating 0 after adding a folded image.

In each of the above embodiments,
Although the sub-band division for dividing the horizontal frequency band into two has been described, it is obvious that the same configuration can be applied to the case where the vertical frequency band is divided into two. In addition, the horizontal frequency band is set to 2
By continuously connecting the sub-band dividing circuit for dividing and the sub-band dividing circuit for dividing the vertical frequency band into two,
The present invention is also applicable to a case where a two-dimensional frequency band is divided into four parts.
Further, it is clear that the present invention can be applied to a case where a circuit for dividing the sub-band is configured more finely by continuously connecting a plurality of sub-band dividing circuits.

In the above embodiment, all the folded images are added. However, it is apparent that only the number of pixels required for filtering needs to be actually added.

[0266]

According to the first and ninth aspects of the present invention, the position of the conversion coefficient for starting one-dimensional scanning is changed according to the type of band, and scanning is performed from the conversion coefficient with the maximum power.
The redundancy of the digital image signal can be reduced as compared with the related art.

According to the second and tenth aspects of the present invention, a part of each divided band is thinned out at a constant period and encoded, so that the image quality is less deteriorated and the encoding rate can be reduced.

According to the third and twelfth inventions, only the effective image blocks are coded.
The coding rate can be reduced.

According to the fourth and eleventh aspects, it is determined whether or not the block of the original image includes a flat portion and a portion where the change is sharp.
Since the quantization table is selected based on this determination result, good image quality can be maintained even in a flat portion in the restored image.

According to the fifth and thirteenth aspects, after the spatial sampling positions of the odd and even fields are aligned, the two fields are spring-loaded to form a three-dimensional block and perform the three-dimensional orthogonal transformation. Remarkable information compression can also be performed on digital image signals of the system.

According to the sixth, fourteenth and fifteenth inventions, weighting is applied to the transform coefficients according to the frequency characteristics of each subband. In particular, in the sixth and fifteenth inventions, the weighting is performed so that the subbands are continuously connected. Therefore, effective information compression can be performed.

According to the seventh and eighth inventions, the portion of the folded image which is not transmitted and added for filtering the end point of the image can be accurately reproduced in the subband synthesizing circuit, so that the image can be reproduced even near the end point of the image. It is possible to restore accurately.

In all of the inventions other than the fifth and thirteenth inventions, since coding is performed after dividing into blocks after subband division, more efficient coding can be performed as compared with the case where subband division is not performed.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a conventional digital image signal encoding device.

FIG. 2 is a block diagram showing an internal configuration of an encoding circuit in FIG.

FIG. 3 is a conceptual diagram showing scanning in the encoding circuit shown in FIG. 2;

FIG. 4 is a conceptual diagram of a conventional three-dimensional block of a progressive scanning system.

FIG. 5 is a principle diagram of an interlaced scanning method of a television screen.

FIG. 6 is a conceptual diagram showing the relationship between fields and frames in a standard television signal.

FIG. 7 is a block diagram illustrating a configuration of a conventional encoding device.

FIG. 8 is an explanatory diagram illustrating weighting of a weighting circuit in a conventional encoding device.

FIG. 9 is a block diagram illustrating a configuration on a transmission side of a conventional encoding device.

FIG. 10 is a block diagram illustrating a configuration of a receiving side of a conventional encoding device.

FIG. 11 is a diagram illustrating an example of sampling in a conventional encoding device.

FIG. 12 is a diagram illustrating an example of subsampling in a conventional encoding device.

FIG. 13 is a block diagram showing a configuration of a conventional subband division circuit.

FIG. 14 is a block diagram illustrating a configuration of a conventional subband synthesis circuit.

FIG. 15 is a frequency characteristic diagram of a filter used in FIGS.

FIG. 16 is a conceptual diagram showing band division of the sub-band division circuit shown in FIG.

FIG. 17 is a block diagram showing a configuration on the encoding side of the encoding device according to the present invention.

FIG. 18 is a block diagram illustrating a configuration on a decoding side of the encoding device according to the present invention.

FIG. 19 is a block diagram showing an internal configuration of a sub-band division circuit in FIG.

20 is a block diagram illustrating an internal configuration of a subband synthesis circuit in FIG.

FIG. 21 is a diagram illustrating subband division of a Y signal according to the present invention.

FIG. 22 is a diagram illustrating subband division of an RY signal and a BY signal according to the present invention.

FIG. 23 is a diagram illustrating frequency characteristics of a vertical LPF in the encoding device of the present invention.

FIG. 24 is a diagram showing a frequency characteristic of a vertical HPF in the encoding device of the present invention.

FIG. 25 is a diagram showing a frequency characteristic of a horizontal LPF in the encoding device of the present invention.

FIG. 26 is a diagram showing frequency characteristics of a horizontal HPF in the encoding device of the present invention.

FIG. 27 is a diagram showing one-dimensional scanning in the present invention.

FIG. 28 is an explanatory diagram of the folding operation of the frequency of the horizontal HPF in the encoding device of the present invention.

FIG. 29 is a diagram showing another one-dimensional scan in the present invention.

FIG. 30 is a diagram showing another scan of the three-dimensional block according to the present invention.

FIG. 31 is a diagram illustrating subbands of a Y signal to be encoded in the present invention.

FIG. 32 shows an RY signal, B to be encoded in the present invention.
FIG. 5 is a diagram illustrating subbands of a −Y signal.

FIG. 33 is a diagram illustrating another example of a sub-band of a Y signal to be encoded in the present invention.

FIG. 34 is a block diagram illustrating an internal configuration of a three-dimensional orthogonal transform circuit according to the present invention.

FIG. 35 is a block diagram showing a configuration on the encoding side of the encoding device according to the present invention.

FIG. 36 is a block diagram illustrating a configuration on a decoding side of the encoding device according to the present invention.

FIG. 37 is a block diagram showing an internal configuration of a sub-band division circuit in FIG. 35.

FIG. 38 is a block diagram showing an internal configuration of a subband combining circuit in FIG. 36.

FIG. 39 is a block diagram showing an internal configuration of a determination circuit in FIG. 35;

40 is a block diagram showing another internal configuration of the determination circuit in FIG. 35.

FIG. 41 is a conceptual diagram illustrating the operation of the sub-blocking circuit in FIGS. 39 and 40.

FIG. 42 is a conceptual diagram for explaining the operation of the control signal generator in FIGS. 39 and 40.

FIG. 43 is a block diagram showing a configuration on the encoding side of the encoding device according to the present invention.

FIG. 44 is a block diagram showing an internal configuration of a determination circuit in FIG. 43.

FIG. 45 is a block diagram showing another internal configuration of the subband division circuit in FIG. 17;

FIG. 46 is a block diagram showing another internal configuration of the subband synthesis circuit in FIG. 18;

FIG. 47 is a diagram showing a spatial sampling position in the present invention.

FIG. 48 is a conceptual diagram showing a configuration of a three-dimensional block according to the present invention.

FIG. 49 is a block diagram showing a configuration on the coding side of a coding device according to the present invention.

FIG. 50 is a diagram illustrating the waiting of the waiting circuit in FIG. 49;

FIG. 51 is a block diagram illustrating a configuration of a subband division circuit according to the present invention.

FIG. 52 is a block diagram illustrating a configuration of a subband synthesis circuit according to the present invention.

FIG. 53 is a conceptual diagram for explaining the operation of the subband division / combination circuit shown in FIGS. 51 and 52.

FIG. 54 is a block diagram showing another configuration of the subband dividing circuit according to the present invention.

FIG. 55 is a block diagram showing another configuration of the subband synthesis circuit according to the present invention.

FIG. 56 is a conceptual diagram for explaining the operation of the subband division / combination circuit shown in FIGS. 54 and 55.

[Explanation of symbols]

 5 Sub-band division circuit 6 Three-dimensional orthogonal transformation circuit 7 Encoder 42 Orthogonal transformation circuit 43 Judgment value calculation circuit 44 Judgment device 51 Sub-band division circuit 52 Blocking circuit 52a Blocking circuit 52b Blocking circuit 52c Blocking circuit 52d Blocking Circuit 53 Orthogonal transform circuit 53a Orthogonal transform circuit 53b Orthogonal transform circuit 53c Orthogonal transform circuit 53d Orthogonal transform circuit 54a Waiting circuit 54b Waiting circuit 54c Waiting circuit 54d Waiting circuit 55a Variable length encoding circuit 55b Variable length encoding circuit 55c Variable length encoding Circuit 55d Variable-length encoding circuit 56 Quantizing circuit 57 Encoding circuit 58 Judgment circuit 58a Judgment circuit 71 Folded image addition circuit 72 Horizontal LPF 73 Horizontal HPF 74a Horizontal 2: 1 sub-sampling circuit 74b Horizontal 2: 1 sub-sampling circuit 74c Horizontal 2: 1 sub-sampling circuit 77 Return image addition circuit 78 Return image addition Road 79a Horizontal 1: 2 interpolation circuit 79b Horizontal 1: 2 interpolation circuit 79c Horizontal 1: 2 interpolation circuit 80 Horizontal LPF 81 Horizontal HPF 84 Returned image addition circuit 85 Horizontal LPF 86 Horizontal HPF 87a Returned image addition circuit 87b Returned image addition circuit 88 Horizontal LPF 89 Horizontal HPF 581 Sub-block circuit 582 Computing unit 583 Minimum value detector 584 Maximum value detector 587a Comparator 587b Comparator 588 Control signal generator 588a Control signal generator 589a Absolute value computing circuit 589b Absolute value computing circuit 591 Maximum value detector in sub-block 592 Control signal generator

 ──────────────────────────────────────────────────続 き Continued on the front page (31) Priority claim number Japanese Patent Application No. 2-189929 (32) Priority date Hei 2 (July 16, 1990) (33) Priority claim country Japan (JP) (31) Priority Claim number Japanese Patent Application No. 2-191662 (32) Priority Date Hei 2 (1990) July 17 (33) Priority claiming country Japan (JP) (31) Priority claim number Japanese Patent Application No. 2-243555 (32) Priority Japan 2 (1990) September 12 (33) Priority claiming country Japan (JP) (31) Priority claim number Japanese Patent Application No. 3-79210 (32) Priority date Hei 3 (1991) April 11 (33) ) Priority country Japan (JP) (56) References Patent 2026614 (JP, C2) IEEE PROC. ICASSP (1990) (A TECHNIQUE FOR THE EFFICIENT C ODING OF THE UPPER BANDS IN SUBBAND CODING OF IMAGES) IEICE Technical Report, SS E88-176 61-66 IEEE GLOBCOM '88 (1988) p. 743-749 (COMPATI BLE HDTV CODING FOR BROADBAND ISDN)

Claims (17)

(57) [Claims] 1. An encoding method for encoding a digital image signal, the method comprising: dividing the digital image signal into image signals of a plurality of sub-bands; And a step of subjecting each of the blocked blocks to orthogonal transform to obtain transform coefficients; and a step of one-dimensionally scanning and encoding the obtained transform coefficients. A coding method characterized by changing a start position of the transform coefficient in the one-dimensional scanning depending on a band. 2. An encoding method for encoding a digital image signal, comprising the steps of: dividing the digital image signal into image signals of a plurality of sub-bands; and dividing the image signal of each sub-band into a plurality of pixels. Steps and
A step of thinning out some of the blocks that have been blocked,
An encoding method comprising: performing orthogonal transform on each of the remaining blocks to obtain transform coefficients; and encoding the obtained transform coefficients. 3. An encoding method for encoding a digital image signal, wherein the digital image signal is divided into a plurality of sub-band image signals, and each sub-band image signal is divided into a plurality of pixels. Steps and
Determining the activity of the image signal in each of the blocked blocks; determining whether each block is a valid image block based on the activity; and performing orthogonal transformation on each of the blocked blocks to perform conversion. A coding method comprising: obtaining a coefficient; and coding only a transform coefficient in an effective image block. 4. An encoding method for quantizing and encoding a digital image signal, wherein the digital image signal is divided into a plurality of sub-band image signals, and each sub-band image signal is divided into a plurality of pixels. Blocking, performing orthogonal transformation on each of the blocked blocks to obtain a transform coefficient, and performing one quantization from a plurality of quantization tables having different quantization step widths based on pixel values in the high-band block. An encoding method, comprising: selecting an encoding table; and quantizing and encoding the obtained transform coefficients according to the selected quantization table. 5. An encoding method for encoding a digital image signal of an interlaced scanning system, wherein one of an odd field and an even field of the digital image signal is passed through a vertical filter having odd taps, and Passing the other of the odd and even fields of the signal through a vertical filter with even taps, and forming horizontal and vertical two-dimensional blocks in each field with output signals from both vertical filters; Spring forming two-dimensional blocks for a plurality of fields in the time direction to form a three-dimensional block; applying orthogonal transformation to the formed three-dimensional block to obtain a transform coefficient; and encoding the obtained transform coefficient A coding method comprising: 6. An encoding method for encoding a digital image signal, wherein the digital image signal is divided into image signals of a plurality of sub-bands, and the image signal of each sub-band is divided into a plurality of pixels. Steps and
Applying orthogonal transform to each of the block-formed blocks to obtain transform coefficients; applying weighting to the obtained transform coefficients so that weighting for each subband is continuous in the frequency direction; and Encoding the applied transform coefficients. 7. A sub-band splitting / synthesizing method for dividing a digital image signal into sub-bands and synthesizing the divided digital image signals, wherein an image signal folded at an end point of the digital image signal is connected to the end point. And a step of passing the image signal to which the aliased image signal is connected through a first low-pass filter having even taps, and decimating the output from the first low-pass filter 2: 1 to form a first low-pass filter.
Obtaining an output signal, passing the image signal to which the aliased image signal is connected through a first high-pass filter having even taps, and decimating the output from the first high-pass filter by 2: 1 to obtain a second output. Obtaining an output signal; connecting an image signal folded at an end point of the first output signal to the end point; interpolating 0 to the first output signal to which the folded image signal is connected and doubling the output signal. Setting the number of pixels, passing the 0-interpolated image signal through a second low-pass filter of even-numbered taps, folding the image signal at the end point of the second output signal and multiplying each pixel value by -1 To the end point; interpolating 0 to the second output signal to which the folded image signal is connected to obtain twice the number of pixels; and converting the image signal after 0 interpolation to the second height of the even tap. Pass the bandpass filter Step a, the second sub-band dividing and synthesizing method characterized by a step of combining the output of said second high-pass filter and the output of the low pass filter. 8. A sub-band dividing / combining method for dividing a digital image signal into sub-bands and combining the divided digital image signals, wherein the image signal obtained by reducing one pixel from an image signal folded at an end point of the digital image signal. To the end point, passing the image signal to which the folded image signal is connected through a first low-pass filter having odd taps, and thinning out the output from the first low-pass filter 2: 1. Obtaining a first output signal from the first high-pass filter, passing the image signal to which the folded image signal is connected through a first high-pass filter having odd taps, and thinning out the output from the first high-pass filter 2: 1. Obtaining a second output signal, interpolating 0 to the first output signal to double the number of pixels, and wrapping around the end point of the first output signal after the 0 interpolation. Connecting the image signal obtained by reducing the image signal by one pixel to the end point, passing the image signal to which the folded image signal is connected through a second low-pass filter having odd taps, and outputting the second output signal to the second output signal. 0
To a pixel number twice as many as the number of pixels, and connecting an image signal obtained by reducing one pixel from an image signal folded at an end point of the second output signal after the 0 interpolation to the end point,
Passing the image signal to which the aliased image signal is connected through a second high-pass filter having odd taps, and combining the output of the second low-pass filter and the output of the second high-pass filter. A subband division / combination method characterized by having: 9. An encoding apparatus for encoding a digital image signal, comprising: a sub-band dividing means for dividing the digital image signal into image signals of a plurality of sub-bands; Blocking means for blocking into blocks of two or more dimensions, orthogonal transform means for performing an orthogonal transform on each of the blocked blocks to obtain transform coefficients, and obtaining the obtained transform coefficients from different start positions for each subband. A coding unit for performing one-dimensional scanning and coding. 10. An encoding apparatus for encoding a digital image signal, comprising: a sub-band dividing unit for dividing the digital image signal into image signals of a plurality of sub-bands; Blocking means for blocking, means for thinning out some of the blocks that have been blocked, orthogonal transformation means for performing orthogonal transformation on each of the remaining blocks to obtain transformation coefficients, and a code for encoding the obtained transformation coefficients Encoding means comprising encoding means. 11. An encoding apparatus for encoding a digital image signal, comprising: a sub-band dividing means for dividing the digital image signal into image signals of a plurality of sub-bands; Blocking means for blocking, orthogonal transformation means for performing an orthogonal transformation on each of the blocked blocks to obtain a transformation coefficient, and a plurality of quantization tables having different quantization step widths based on pixel values in a high-band block An encoding apparatus comprising: means for selecting one quantization table from, and encoding means for quantizing and encoding a transform coefficient obtained based on the selected quantization table. 12. An encoding apparatus for encoding a digital image signal, comprising: a sub-band dividing means for dividing the digital image signal into image signals of a plurality of sub-bands; Blocking means for blocking, orthogonal transform means for performing an orthogonal transform on each of the blocked blocks to obtain a transform coefficient, and coding means for coding the obtained transform coefficient, the orthogonal transform means Means for determining the activity of a sub-band block obtained by the blocking means, means for determining whether or not the image block is an effective image block based on the obtained activity, and means for outputting a transform coefficient in the effective image block. An encoding device, comprising: 13. An interlaced scanning digital image signal.
An encoding device for encoding the digital image.
One of the odd and even fields of the image signal
A first vertical filter of odd taps passing through the
Of the odd and even fields of the image signal
A second vertical filter of even taps passing through the other;
The image signals output from the first and second vertical filters are
Blocking method to block into 3D blocks for each pixel
And stage, three-dimensional orthogonal transformation on each block which is blocked
And a coding means for coding the obtained transform coefficients. 14. An encoding apparatus for encoding a digital image signal, comprising: a sub-band dividing means for dividing the digital image signal into image signals of a plurality of sub-bands; Block
Blocking means and orthogonal to each block
Orthogonal transform means for performing a transform to obtain a transform coefficient;
Weighting to apply weighting to conversion factor
Encoding means for encoding the output of the weighting means . 15. Waiting for each subband
So that the weighting is continuous in the frequency direction.
2. The method according to claim 1, wherein the means performs weighting.
5. The encoding device according to 4 . 16. An encoding apparatus for encoding a digital image signal, comprising: a sub-band dividing means for dividing the digital image signal into image signals of a plurality of sub-bands; Blocking means for performing blocking , orthogonal transform means for performing an orthogonal transform on each of the blocked blocks to obtain transform coefficients,
It has a plurality of encoding tables according to the
Encoding means for encoding a transposition coefficient based on an encoding table selected from the above . 17. A code for encoding a digital image signal.
A plurality of sub-images.
Sub-band dividing means for dividing into band image signals;
Block image signal of sub-band into multiple pixels
Blocking means and orthogonal to each block
Orthogonal transform means for performing a transform to obtain a transform coefficient;
Weighting to apply weighting to conversion factor
Means and a plurality of encoding tables corresponding to each subband.
And select the output of the waiting means from these
Encoding means for encoding based on the encoded table
An encoding device comprising: