JPH06334985A - Picture encoding device - Google Patents

Abstract

PURPOSE:To efficiently encode a picture signal in a form adapted to visual characteristics. CONSTITUTION:A transformed coefficient S2 obtained by DCT transformation of the picture signal is weighted with space resolution characteristics, which are different by the center part and the peripheral part of the picture, in accordance with a control signal MD by an adaptive weighting part 3, and a generated transformed coefficient S3 is subjected to quantization, variable length code encoding, and run length code encoding by a quantizing part 4 and a Huffman code encoding part 5 to generate encoded data S5 whose information volume is considerably compressed. Consequently, picture encoding of high compression efficiency and less degradation in picture quality is realized because the picture is encoded with the spatial resolution characteristic adapted to visual characteristics in accordance with the area of the picture.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image coding apparatus for coding an image signal with high efficiency, and more particularly to an image coding apparatus suitable for efficiently compressing an information amount in a form adapted to visual characteristics.

[0002]

2. Description of the Related Art Generally, a digitized image signal has an enormous amount of information. Therefore, there is a problem in terms of speed and cost in communicating the image signal as it is through a transmission line or recording it in a storage medium. Therefore, paying attention to the redundancy of the image signal,
High-efficiency coding is performed to compress the information amount to several tenths.

In this high efficiency coding, orthogonal transform coding,
Predictive coding, Huffman coding, and various coding schemes by combining these have been devised. These encoding methods focus on the statistical properties of the image and remove redundancy in the horizontal, vertical, and temporal directions contained in the image signal to achieve efficient compression of the amount of information. Then, the amount of information required for transmission / accumulation can be compressed to about several M to several tens of Mbits.

[0004]

In the above-mentioned high efficiency coding, the coding efficiency is poor depending on the image or the image quality deterioration due to the coding is mainly performed because the coding is based on the statistical property of the image. Has a problem that it becomes conspicuous.

An object of the present invention is to provide an image coding apparatus which has a high compression efficiency of information amount and has little image quality deterioration due to coding.

[0006]

In order to achieve the above object, the present invention is provided with means for generating a plurality of types of image signal series having different horizontal and vertical spatial resolutions based on an image signal. Then, an image signal sequence having a high spatial resolution is selected in the central region of the television screen, and an image signal sequence having a low spatial resolution is selected in the peripheral region, and high efficiency encoding is performed using these image signal sequences.

Further, the generation of the image signal series is realized by the weighting process of different characteristics for the transform coefficient obtained by performing the discrete cosine transform of the image signal.

[0008]

The principle of the present invention is shown in FIG. FIG. 3A shows an example of the spatial resolution characteristic of the image on the television screen according to the present invention. Generally, in television images, program production is performed in such a manner that a main image occupies almost the central area of the screen. Therefore, the viewer's viewpoint is also concentrated in the central area of the screen, and the deterioration of the spatial resolution and the deterioration of the image quality in this area are visually conspicuous. On the other hand, in the peripheral area of the television screen, even if the spatial resolution is slightly lowered or the image quality is slightly deteriorated, it is hardly noticeable visually and is inconspicuous.

Therefore, the code is encoded using the image signal series of mode F having a high spatial resolution in the central area of the screen where the visual characteristics are sensitive and mode M and mode L in which the spatial resolution gradually decreases in the peripheral area of the screen. The converted image can obtain the same overall image quality as that of an image in which the entire screen is encoded with the image signal sequence of mode F. On the other hand, mode M,
The L image signal sequence can be encoded with a smaller number of bits as compared with the mode F. Therefore, according to the present invention, it is possible to realize image coding with the same overall image quality as compared with the conventional technique, the total number of coded bits is small, and the compression efficiency is high.

Further, in the present invention, FIG.
As shown in (d), the transform coefficients obtained by performing the discrete cosine transform of the image signal are weighted with different characteristics in modes F, M, and L to generate image signal sequences having different spatial resolution characteristics. The transform coefficient of the discrete cosine transform corresponds to the direct current component to the high frequency component in the horizontal / vertical frequency domain. Therefore, by performing a weighting process on these transform coefficients, it is possible to easily generate transform coefficients equivalent to the transform coefficients of the discrete cosine transform for the image signal series with the limited horizontal / vertical spatial resolution characteristics. That is, horizontal
Complex processing such as frequency band limitation by a vertical two-dimensional filter is not necessary, and conversion coefficients for image signal sequences having different spatial resolutions can be generated by simple signal processing that simply weights conversion coefficients.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment according to the present invention will be described with reference to the block diagram shown in FIG. This is suitable for video coding in which discrete cosine transform coding and Huffman coding are combined.

Image signals in the form of components (luminance signal Y, color difference signals C _r , C _b ) are input to the image preprocessing unit 1,
Format conversion such as digitization and frame synthesis,
Signal processing such as block division is performed to generate, for example, a block signal S1 of 8 × 8 pixels. Further, a mode signal MD that specifies the spatial resolution characteristic of the block is generated from the position on the frame.

The DCT calculator 2 performs a matrix calculation using an 8 × 8 discrete cosine transform matrix to generate a transform coefficient S2. Then, in the adaptive weighting unit 3, the mode signal MD
The conversion coefficient is weighted in any one of modes F, M, and L according to the above, and a conversion coefficient corresponding to an image signal sequence having a predetermined spatial resolution characteristic is generated. The quantizer 4 quantizes the transform coefficient to transform it into a fixed-length code S4. The quantization characteristic is controlled by the control signal CT of the encoding control unit 7. The Huffman coding unit 5 performs variable length coding processing in which a code having a short code length is assigned to a code having a high occurrence probability, and a variable length code S5 is generated. The variable length coding parameters are controlled by the control signal CT. Then, it is input to the buffer unit 6. On the other hand, a signal is read from the buffer unit 6 at a constant speed and a code S6 having a constant bit rate is output. Further, the information indicating the state of the buffer capacity is input to the coding control unit 7, and the control signal CT for controlling the coding parameter is generated based on this signal. The transmission path coding unit 8 performs predetermined signal processing such as packetization and addition of error correction code to generate a coded bit stream signal BS.

On the other hand, in the decoding section, signal processing reverse to that in the encoding section is performed to decode the image signal. Transmission line decoding unit 9
Then, the code error generated in the transmission path is corrected and the constant bit rate code S6 is decoded. And the buffer unit 1
0 and the decoding control unit 11. Decryption control unit 1
1, the control signal DT for controlling the reading of the signal from the buffer unit, the decoding parameter, etc. based on this signal.
To generate. The variable-length code S5 read from the buffer unit 10 is the fixed-length code S5 originally used by the Huffman decoding unit 12.
Decrypt to 4. The inverse quantization unit 13 performs inverse quantization processing and decodes the transform coefficient S3. Then, the IDCT calculation unit 14 performs a matrix calculation using a discrete cosine transform inverse matrix to decode the block signal S1 of 8 × 8 pixels. The code error correction unit 15 outputs a signal S1 'which has been subjected to error correction processing for replacing a block signal containing a code error with a correction signal having a strong correlation (for example, a block signal of the previous frame). Then, the image post-processing unit 16 performs conversion into a predetermined image format and conversion into an analog signal, and an image signal (luminance signal Y, color difference signal C in component form.
Decode _r , C _b ).

An embodiment of this main block section will be described below.

FIG. 3 is a circuit diagram of an embodiment of the image preprocessing section 1. The luminance signal Y and the color difference signals C _r and C _b are converted into digital signals by the A / D converter 17. Then, the color difference signal is subjected to a process of thinning out the sampling points in the horizontal and vertical directions by 1/2 in the sub-sampling unit 18. The frame synthesizing unit 19 synthesizes, for example, the signals of the first and second fields of interlaced scanning to create the signal of the frame shown in FIG. The signal of the frame is divided into macroblocks, and each macroblock has 8 × 8 pixel blocks Y ₀ , Y ₁ , Y ₂ , Y ₃ and 8 × 8 pixel color difference signals C _r , C for the luminance signal. It consists of blocks of _b . Then, discrete cosine transform is performed in units of this block. In the block sequence generator 20, these Y ₀ , Y ₁ , Y ₂ , Y ₃ , C _r ,
A block signal sequence S1 of C _b is generated. Further, the weighting mode setting unit 21 detects the position of each macroblock in the frame, and the mode F having a high spatial resolution in the central region, the mode M having a low spatial resolution in the peripheral region,
A mode signal MD designating L is generated.

FIG. 4 is a circuit diagram of an embodiment of the DCT calculator 2. In the discrete cosine transform, the transform coefficient f (u,
v) (u: horizontal, v: vertical) is a product operation of conversion coefficients fu and fv in the horizontal and vertical directions, that is, f (u, v) = fu
-It can be generated by the calculation of fv.

The fu calculation unit 22 performs horizontal calculation using the conversion coefficient fu. The rearrangement unit 23 rearranges the signals in order to perform calculation in the vertical direction, and the fv calculation unit 2
In step 4, the calculation with the conversion coefficient fv is performed to obtain the conversion coefficient S2.
(F (u, v)) is generated.

FIG. 5 is a diagram showing an embodiment of the adaptive weighting unit 3. Generally, in the transform coefficient of the discrete cosine transform, main components are concentrated in the horizontal and vertical low bands. Therefore, the zigzag scanning unit 25 arranges the conversion coefficients in the zigzag scanning order as shown in FIG. Weighting and weighting units 26, 27,
At 28, the conversion coefficients are weighted with the characteristics shown in (c), (d) and (e) of the same figure, and the conversion coefficients corresponding to the spatial resolutions of the modes F, M and L are generated. By this weighting process, the characteristics from the mode F with high spatial resolution to the mode L with low spatial resolution are realized. The switch 29 is connected to the terminals a, b, and c in modes F, M, and L, respectively, by the mode signal MD, and generates a conversion coefficient S3 having a predetermined spatial resolution characteristic.

FIG. 6 is a circuit diagram of an embodiment of the quantizer 4. This is the case where there are two types of quantization characteristics. The quantizing circuits 30 and 31 quantize the transform coefficient with different characteristics and convert it into a fixed-length code. The switch 32 is
According to the quantization mode QM of the control signal CT, the signal S4 is output by connecting to either of the terminals a and b. In addition, the functions of the quantizing circuits 30 and 31 and the switch 32 are collectively described as
It can also be realized in the form of a table lookup by ROM.

FIG. 7 is a circuit diagram of an embodiment of the Huffman encoder 5. The variable length coding unit 33 converts the fixed length code of the transform coefficient into the variable length code VLC. The run-length coding unit 34 generates a code RLC by run-length coding the number of zero-valued transform coefficients. In addition, the coding parameter generation unit 35 generates information CP of coding parameters necessary for decoding. Then, the multiplexing unit 36 time-divisionally multiplexes these signals to generate a variable-length code sequence S5. FIG. 11B shows an example of the configuration of the code sequence S5 in the macroblock layer (described in FIG. 3). Each macroblock is composed of an MB header part and a block data part. The MB header part includes information CP of encoding parameters,
That is, it is formed by a code indicating a synchronous code, a coding type, a quantization type, a coding block pattern, and the like. Further, the block data part is formed by the codes VLC and RLC.

FIG. 8 is a circuit diagram of an embodiment of the image post-processing section 16. The decoded block signal S1 'is input to the frame reproduction unit 37, arranged in a predetermined block area and reproduced as a frame signal. The sample point interpolation unit 38 performs sample point interpolation processing on the color difference signal components to decode the signal of the original sampling structure. The format conversion unit 39 performs image format conversion for converting a frame signal into an interlaced scanning signal, for example. And D
The / A converter 40 converts the analog signal and decodes the image signal in the component form of the luminance signal Y and the color difference signals C _r and C _b .

In the case of a composite type image signal, for example, an NTSC television signal, the present embodiment can be applied by the configuration of the image pre-processing section and post-processing section as described below.

FIG. 9 is a circuit diagram of an embodiment of the image preprocessor 1 for the composite signal. The NTSC television signal VS is sampled by the A / D converter 17 at a frequency four times as high as the color subcarrier _fsc and converted into a digital signal. The YC separation unit 41 separates the luminance signal Y and the color signal C by a two-dimensional YC separation process, for example. Then, the color difference demodulation unit 42 performs synchronous detection using the color sub-carrier wave f _sc ,
The color difference signals C _r and C _b are demodulated. The sub-sampling unit 18 thins out the sampling points to 1/2 in the horizontal and vertical directions. The frame synthesizing unit 19 synthesizes the signals of the first and second fields of interlaced scanning to form a frame signal. The block sequence generation unit 20 divides the frame signal into the luminance signal Y ₀ , as shown in FIG.
_{Y 1, Y 2, Y 3} , block signal S1 of the color difference signal C _r, C _b
To generate. On the other hand, the weighting mode setting unit 21 detects a block position in the frame and generates a mode signal MD that specifies a mode F having a high spatial resolution in the central region and a mode M or L having a low spatial resolution in the peripheral region. To do.

FIG. 10 is a circuit diagram of an embodiment of the image post-processing section 16 for the composite signal. The decoded block signal S1 'is input to the frame reproduction unit 37, arranged in a predetermined block area and reproduced as a frame signal. The sample point interpolation unit 38 performs sample point interpolation processing on the color difference signal components to decode the signal of the original sampling structure. The format conversion unit 39 scan-converts the frame signal into the original interlaced scanning signal. Then, the color difference signals C _r and C _b are _subjected to quadrature amplitude modulation by the color subcarrier f _sc in the color difference converter 43 to generate the color signal C. Process part 4
In step 4, the color signal C is added to the luminance signal Y, and a predetermined signal such as a synchronizing signal or a burst signal is added. And D / A
The converter 40 converts the signal into an analog signal and decodes a composite NTSC television signal.

FIG. 11 is a circuit diagram of an embodiment of the code error correction unit 15. The decoded block signal S1 is input to the terminal a of the switch 45. On the other hand, to the terminal b, for example, the block signal of one frame before generated by the correction signal generation unit 46 is input as the correction signal ECS. And
By the error block flag signal EBF of the control signal DT,
A switch is connected to the terminal a in the block having no code error and to the terminal b in the block containing the code error to generate the block signal S1 'which has been subjected to the error correction processing for replacing with the correction signal. This signal S1 'is also input to the modification signal generator 46 and used as a modification signal in the next frame.

As described above, according to the present embodiment, it is possible to realize an image signal encoding apparatus which is suitable for visual characteristics, has high compression efficiency, and has little image quality deterioration due to encoding.

Next, regarding the second embodiment of the present invention,
This will be described with reference to the block diagram shown in FIG. 12 and the explanatory diagram of the encoding process shown in FIG. This embodiment is based on CCITT Recommendation H.264. It is suitable for an encoding device compliant with the H.261 video encoding system.

First, the description of the encoding process in this embodiment will be given with reference to FIG. The figure (a) shows the parameter of the encoding process with respect to each encoding mode, and the figure (b) shows the form of the predictive encoding process. In the I-picture coding mode, in-frame DCT conversion coding is performed as in the first embodiment. That is, different weighting processing is performed on the transform coefficient of the scattered cosine transform depending on the area of the block, the transform coefficient of the spatial resolution characteristic corresponding to the modes F, M, and L is generated, and this is Huffman-encoded. . On the other hand, in the P picture coding mode, motion compensation interframe predictive coding and DCT transform coding are performed.

That is, as shown in FIG. 3B, the difference between the prediction signal generated by subjecting the signal of the previous frame to motion compensation processing and the signal of the current frame is calculated to extract the prediction error signal.
Then, the prediction error signal is subjected to discrete cosine transform, and the transform coefficient is Huffman coded. Therefore, in the P picture, the weighting process is performed with, for example, a through characteristic (all transform coefficients have a weight of 1).

In the coding section shown in FIG. 12, the switches 47 and 53 are connected to the terminal a in the I picture coding mode and to the terminal b in the P picture coding mode to perform the coding process.

The image signals in the component form of the luminance signal Y and the color difference signals C _r , C _b are input to the image preprocessing unit 1 and subjected to signal processing such as digitization and frame synthesis, and 8 × 8.
A block signal series S1 of pixels is generated. The subtractor 48 subtracts the prediction signal PS from the signal S1 to generate a prediction error signal PE. Further, the motion vector detection unit 52 extracts the motion vector MV required for the motion compensation process by a method such as block matching.

The DCT calculator 2 performs a matrix calculation using an 8 × 8 discrete cosine transform matrix to generate a transform coefficient S2. In the case of I picture, the adaptive weighting unit 3 performs weighting processing according to the mode signal MD,
Transform coefficients corresponding to the M and L spatial resolution characteristics are generated. On the other hand, in the case of a P picture, the signal S has a through characteristic.
Generates the same signal as 2. Then, in the quantizer 4,
The transform coefficient is quantized and converted into a fixed-length code S4. The Huffman coding unit 5 performs variable length coding processing in which a code having a short code length is assigned to a code having a high probability of occurrence, coding modes, coding parameters, motion vector information, and the like. Multiplexing is performed for division to generate a variable-length code sequence signal S5. Then, this signal is written in the buffer unit 6. On the other hand, a signal is read from the buffer unit 6 at a constant bit rate, packetized by the transmission path coding unit 8,
Predetermined signal processing such as addition of an error correction code is performed to generate a coded bitstream signal BS. In addition, the coding control unit 7 generates a control signal CT that controls coding parameters based on the buffer capacity information of the buffer unit 6.

Further, in order to generate the prediction signal, the fixed-length code S4 is decoded by the dequantization unit 13 and the IDCT calculation unit 14, and the output signal of the switch 53 is added by the addition unit 49 to obtain the original signal. The block signal S1 is decoded and input to the frame memory 50. Then, the output signal (one frame delay signal) of the frame memory 50 is input to the motion compensation prediction signal generation unit 51, and the motion compensation process is performed based on the motion vector information MV to generate the prediction signal PS.

On the other hand, in the decoding section, the coded bit stream signal BS is input to the transmission path decoding section 9 and a predetermined decoding process such as correction of code error is performed. Then, the decoded signal S6 is input to the buffer unit 10 and the decoding control unit 11. The decoding control unit 11 detects the information of the coding mode and the coding parameter from the signal S6 and generates the control signal DT necessary for the decoding process such as the reading from the buffer unit and the quantization type. Output signal S5 of buffer unit 10
Is input to the Huffman decoding unit 12 and decodes the original fixed-length code sequence S4 and the motion vector information MV.
The inverse quantization unit 13 performs inverse quantization processing, and transform coefficient S3
To decrypt. Then, the IDCT operation unit 14 performs an operation using a discrete cosine transform inverse matrix.

This output signal is a block signal when the coding mode is I picture and a prediction error signal is decoded when it is P picture. Then, in the P picture, the addition unit 49 adds the prediction signal PS to decode the block signal. The switch 54 is connected to the terminals a and b for I picture and P picture, respectively, and the decoded block signal S1
Is output. The code error correction unit 15 performs an error correction process of replacing a signal of a block including a code error with a correction signal (for example, a signal one frame before).

One of the output signals S1 'is input to the image post-processing unit 16 and subjected to predetermined signal processing such as image format conversion and conversion into analog signals to obtain a luminance signal Y and color difference signals C _r , C _b. The image signal in the component form of is decoded. On the other hand, the output signal (1 frame delay signal) of the frame memory 50 is input to the motion compensation prediction signal generation unit 51,
A motion compensation process is performed using the motion vector information MV to generate a prediction signal PS.

Each block section in this embodiment is
It can be realized with the same configuration as that of the first embodiment. Further, for the image signal in the composite form, this embodiment can be applied as it is by adopting the configurations of the image preprocessing unit in FIG. 9 and the image postprocessing unit in FIG.

As described above, according to the present embodiment, CCITT Recommendation H.264 is used in a form that is compatible with visual characteristics, has little deterioration in image quality due to encoding, and has high compression efficiency. 261
It is possible to realize an encoding device conforming to the video encoding method of.

Next, regarding the third embodiment of the present invention,
This will be described with reference to FIGS. 14 to 16. This is suitable for an encoding device for video encoding conforming to the moving image encoding standard MPEG for storage media. FIG. 14 is a block diagram of an encoding unit, FIG. 15 is an explanatory diagram of encoding processing, and FIG. FIG. 3 is a block diagram of a decoding unit.

First, the outline of the encoding process will be described with reference to FIG. As shown in FIG. 9A, the coding mode is I
There are three types: picture, P picture, and B picture. I
The picture is the DCT in the frame as in the first embodiment.
Transform coding is performed. Then, the weighted transform coefficients of the modes F, M, and L having different spatial resolution characteristics according to the area of the block in the frame are Huffman-encoded. As with the second embodiment, the P-picture is a DC-coded prediction error signal for motion-compensated interframe forward prediction coding shown in FIG.
T conversion coding and Huffman coding are performed. For the B picture, prediction error signals of forward, backward, and forward / backward average predictive coding between motion-compensated frames are subjected to DCT transform coding and Huffman coding. Then, in the P-picture and B-picture modes, the weighting process is performed with, for example, a through characteristic (all transform coefficients have a weight of 1).

As shown in FIG. 14, in the encoding unit, the image signals in the component form of the luminance signal Y and the color difference signals C _r and C _b are input to the image preprocessing unit 1 and are subjected to predetermined processing such as digitization and frame synthesis. Signal processing is performed to generate a block signal series S1 of 8 × 8 pixels. Further, the mode signal MD that specifies the spatial resolution characteristic according to the block area is generated.
The subtraction unit 48 subtracts the prediction signal PS from the signal S1,
The prediction error signal PE is generated. In addition, the motion vector detection unit 52 has motion vector information M necessary for motion compensation processing.
V is extracted by a block matching method or the like.

In the I picture coding mode, the switches 47 and 53 are connected to the terminal a, and to the terminal b for the P picture and B picture. Further, the switch 58 is connected to any one of the terminals C, P and B, and performs encoding processing.

The DCT calculator 2 performs a matrix calculation using an 8 × 8 discrete cosine transform matrix to generate a transform coefficient S2. The adaptive weighting unit 3 performs weighting processing on the I picture in accordance with the mode signal MD, and sets the modes F, M,
A conversion coefficient corresponding to the spatial resolution characteristic of L is generated. On the other hand, the P picture and the B picture output the same signal as the input due to the through characteristic. Then, the quantizer 4 quantizes the transform coefficient to generate a fixed-length code S4. Also,
The Huffman coding unit 5 assigns a code having a short code length to a code having a high occurrence probability, and a variable-length coding process,
The coding mode, coding parameters, motion vector information, etc. are coded, and these are time-division multiplexed and output. This signal is input to the buffer unit 6, and the signal read out at the constant bit rate is packetized by the transmission line coding unit 8.
Predetermined signal processing such as addition of an error correction code is performed to generate a coded bitstream signal BS. The coding control unit 7 also generates a control signal CT necessary for controlling coding parameters based on the information on the buffer capacity.

The fixed length code S4 is an inverse quantizer 13, I
The DCT operation unit 14 decodes it, and the addition unit 49 switches 5
The original block signal sequence S1 is decoded by adding it to the output of 3 and input to the frame memory 50. Then, the backward prediction signal generation unit 55, the front-back average prediction signal generation unit 53, and the forward prediction signal generation unit 57 each perform motion compensation processing on the output signal of the frame memory using the motion vector information MV to generate a prediction signal. The switch 58 selects one of them according to the control signal CT and outputs the prediction signal PS.

On the other hand, in the decoding section, as shown in FIG.
The encoded bit stream signal is input to the transmission path decoding unit 9 and subjected to predetermined decoding processing such as correction of code error. Then, the decoded signal S6 is input to the buffer unit 10 and the decoding control unit 11. The decoding control unit 11 detects the information of the coding mode and the coding parameter from the signal S6, and generates the control signal DT necessary for the decoding control such as the reading control of the buffer unit and the quantization type.

The Huffman decoding unit 12 decodes the fixed-length code and the motion vector information MV.
The inverse quantization unit 13 inversely quantizes the signal S4 and decodes the transform coefficient S3. The IDCT calculation unit 14 performs matrix calculation using a discrete cosine transform inverse matrix. This output is a block signal series in the I picture coding mode and a prediction error signal in the P and B pictures. Then, in P and B pictures, the prediction signal PS is added by the adder 49 to decode the block signal sequence. The switch 54 is connected to the terminal a for the I picture and the terminal b for the P and B pictures to obtain the decoded block signal sequence S1. In the code error correction unit 15,
The signal of the block containing the code error is modified by the modified signal (for example, 1
Perform error correction processing that replaces with the signal before the frame). Then, the image post-processing unit 16 performs predetermined signal processing such as image format conversion and conversion into an analog signal to decode the image signal in the component form of the luminance signal Y and the color difference signals C _r and C _b .

On the other hand, the output of the frame memory 50 is the backward prediction signal generator 55, the front-back average prediction signal generator 56,
The prediction signal is input to the forward prediction signal generation unit 57 and motion compensation processing is performed using the motion vector information MV to generate a prediction signal. The switch 58 is connected to the terminal C in the P picture,
In the B picture, it is connected to predetermined terminals a, b, and c to generate the prediction signal PS.

Each block section in this embodiment is
It can be realized with the same configuration as that of the first embodiment. Further, the present embodiment can be applied to a composite type image signal by the configuration of the image preprocessing unit in FIG. 9 and the image postprocessing unit in FIG. 10.

As described above, according to the present embodiment, the moving image coding standard MPEG for storage media, which conforms to visual characteristics, has little deterioration in image quality due to coding, and has high compression efficiency.
It is possible to realize an encoding device conforming to the video encoding method of.

Next, FIG. 17 shows an example of area distribution in the weighting mode in the present invention. The same figure (a) (b) (c)
Indicates an aspect ratio of 4: 3, and (d), (e), and (f) indicate areas of each mode on a television screen having an aspect ratio of 16: 9. Horizontal and vertical spatial resolutions are mode F, mode M,
It becomes worse in the order of Mode L. In either case, the mode F having a high spatial resolution is assigned to the central area of the screen.
Then, an intermediate mode M is allocated as the position approaches the peripheral portion of the screen, and a mode L having the lowest spatial resolution is allocated in the peripheral region. The ratio of the area occupied by these modes also depends on the transmission bit rate. Generally, in a system with a high bit rate, the ratio of mode F is high.
In a system with a low bit rate, a small ratio is realized.

Further, FIG. 17 shows the case where the spatial resolution is changed stepwise, but it is also conceivable that the spatial resolution is continuously changed from the central area to the peripheral area of the screen.

[0053]

As described above, according to the present invention, it is possible to provide an image coded information apparatus which has high compression efficiency of information amount and which performs video coding adapted to visual characteristics.

[Brief description of drawings]

FIG. 1 is a block diagram of a first embodiment of the present invention.

FIG. 2 is an explanatory diagram of the principle of the present invention.

FIG. 3 is a block diagram of an embodiment of an image preprocessing unit.

FIG. 4 is a block diagram of an embodiment of a DCT calculation unit.

FIG. 5 is an explanatory diagram of an example of an adaptive weighting unit.

FIG. 6 is a block diagram of an embodiment of a quantizer.

FIG. 7 is a block diagram of an embodiment of a Huffman encoding unit.

FIG. 8 is a block diagram of an embodiment of an image post-processing unit.

FIG. 9 is a block diagram of an embodiment of an image preprocessing unit for a composite signal.

FIG. 10 is a block diagram of an embodiment of an image post-processing unit for composite signals.

FIG. 11 is a block diagram of an embodiment of a code error correction unit.

FIG. 12 is a block diagram of a second embodiment of the present invention.

FIG. 13 is an explanatory diagram of encoding processing according to the second embodiment.

FIG. 14 is a block diagram of an encoding unit according to a third embodiment of the present invention.

FIG. 15 is an explanatory diagram of encoding processing according to the third embodiment.

FIG. 16 is a block diagram of a decoding unit according to a third embodiment of the present invention.

FIG. 17 is an explanatory diagram of an example of area distribution in a weighting mode.

[Explanation of symbols]

1 ... Image preprocessing unit, 2 ... DCT calculation unit, 3 ... Adaptive weighting unit, 4 ... Quantization unit, 5 ... Huffman coding unit, 6, 10
... buffer section, 7 ... coding control section, 8 ... transmission path coding section, 9 ... transmission path decoding section, 11 ... decoding control section, 12 ...
Huffman decoding unit, 13 ... Inverse quantization unit, 14 ... IDCT
Calculation unit, 15 ... Code error correction unit, 16 ... Image post-processing unit.

Claims (6)

[Claims] 1. An image coding apparatus for coding an image signal with high efficiency, comprising means for generating a plurality of types of image signal sequences having different horizontal and vertical spatial resolutions, and means for detecting an image area on a television screen. An image signal sequence having a high spatial resolution is selected in the central portion of the television screen and an image signal sequence having a low spatial resolution is selected in the peripheral portion of the television screen, and the image region is highly efficient with respect to the selected image signal sequence. An image encoding device characterized by performing encoding. 2. An image coding apparatus according to claim 1, wherein a plurality of types of image signal sequences having different horizontal and vertical spatial resolutions are generated with different weights for transform coefficients obtained by discrete cosine transform of the image signal. . 3. The image coding apparatus according to claim 1, wherein the high-efficiency coding of the image signal is video coding in which orthogonal transform coding and Huffman coding are combined. 4. The image coding apparatus according to claim 1, wherein the high-efficiency coding of the image signal is video coding in which orthogonal transform coding and motion compensation predictive coding Huffman coding are combined. 5. The high efficiency encoding of an image signal according to claim 1 or 2, according to CCITT Recommendation H.264. An image encoding apparatus that is video encoding compliant with H.261. 6. The high-efficiency coding of an image signal according to claim 1 or 2, wherein ISO / IEC JTC1 / SC29 / WG1 is used.
1. An image coding apparatus that is video coding conforming to the moving picture coding standard for storage media of 1 (MPEG).