JPH11146403A - Device and method for video signal coding and video signal coding program storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To attain reduction of the amount of arithmetic processing at the time of coding by generating an n×n discrete cosine transformation(DCT) coefficient from an n×n picture element block. SOLUTION: A motion detection means 18 compares a uncompensated reference picture accumulated in a frame memory 174 with an object picture, detects a distance of motion between both of them as a motion vector and outputs this to a variable length coding means(VLC) 15 and a motion compensation means 175. A subtracter 12 obtains the object picture and a compensated reference picture outputted from the motion compensation means 175, takes a difference between both and outputs the result as difference data to an n×n DCT means 13. The n×n DCT means 13 divides the difference data into a 4×4 picture element block unit, a DCT processing is performed by taking the 4×4 picture element as a unit, a 4×4 DCT coefficient is converted into, this is further converted into a 8×8 DCT coefficient and is outputted to a quantization means 14.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video signal coding apparatus, a video signal coding method, and a video signal coding program recording medium, and more particularly to a video signal coding process involving enlargement of an input image or change in resolution. The present invention relates to a video signal encoding device, a video signal encoding method, and an information recording medium on which a program for encoding a video signal is recorded.

[0002]

2. Description of the Related Art When a moving picture is encoded by a personal computer in real time, the amount of computation becomes enormous. Therefore, when the above processing is performed by software, the processing is simplified. At this time, measures are taken to suppress an increase in the amount of calculation processing by compressing the size of the image before the encoding processing. However, at the stage of decoding a moving image, a moving image having the size before encoding is obtained, and thus the above-described enlarged moving image needs to be enlarged.

FIG. 4 is a block diagram showing the configuration of a conventional moving picture coding apparatus. In the figure, 400
Is an encoding device that receives a video input 40 and encodes it. Reference numeral 41 denotes a reordering unit, which uses the video input 40 as a target image. Reference numeral 42 denotes a subtractor which calculates a difference between the reference image and the target image, and sets the result as difference data. 43
Is a DCT means for performing DCT transformation of the difference data to generate DCT coefficients. 44 is a quantization means,
The DCT coefficient is quantized. Reference numeral 45 denotes a variable length coding (VLC: Variable Length Code) unit, which performs variable length coding of the motion vector and the quantized DCT coefficient, and outputs the result as a code output 46 as a code output 46.
00 is output to the outside. 47 is a decoding means, which includes an inverse quantization means 471, an inverse DCT means 472, an adder 473,
It is composed of a frame memory 474 and a motion compensating means 475, and detects a motion vector and generates a reference image which is an image used for generating the difference data.
Reference numeral 471 denotes an inverse quantization unit that inversely quantizes the DCT coefficient quantized by the quantization unit 44. 472 is reverse D
CT means for decoding difference data from the inversely quantized DCT coefficients. An adder 473 generates a new motion-compensated (uncompensated) reference image from the decoded difference data and the compensated reference image input from the motion compensation unit 475. A frame memory 474 stores the uncompensated reference image. 4
Reference numeral 75 denotes a motion compensating unit that performs motion compensation on the uncompensated reference image stored in the frame memory 474 using the motion vector, and generates a new compensated reference image.
Reference numeral 48 denotes a motion detection unit which detects a motion vector used for motion compensation of the reference image and the variable length coding. The operation of the moving picture coding apparatus according to the related art configured as described above will be described below.

As a moving image compression method, MPEG (Mo
Predictive coding and Discrete Cosine Transform (DCT: Discret
e Cosine Transformation) is used.

[0005] According to the hybrid coding method, a moving image is first compressed by predictive coding as a first step. When the rearranging means 41 receives the image signal 40,
The image signal 40 is rearranged in the encoding order. Further, the rearranging unit 41 arranges the image signals rearranged in the encoding order into a format obtained by dividing the image signals into 8 × 8 pixel blocks, and uses these as target images for the subtractor 42 and the motion detecting unit 48. Output. The motion detection means 48
The uncompensated reference image stored in the frame memory 474 is compared with the target image, the distance of the motion between the two is detected as a motion vector, and this is output to the VLC means 45 and the motion compensation means 475. I do. The subtractor 42 calculates a difference between the compensated reference image output from the motion compensating unit 475 and the target image, and outputs the difference to the DCT unit 43 as difference data.

Next, as a second stage, a compression operation by the discrete cosine transform (DCT) is performed. DCT means 4
3 receives the difference data, and receives 8 ×
A DCT process is performed for each of the eight pixel blocks to generate a DCT coefficient, and this is output to the quantization means 44. Quantization means 4
4 quantizes the DCT coefficient, and quantizes the quantized DCT.
The coefficients are output to the VLC means 45 and the decoding means 47 as coefficients. Variable Length Coding (VLC: Variable Length Co)
de) means 45, upon obtaining the quantized DCT coefficients and the motion vector, performs variable length coding, and outputs the obtained signal as a code output 46 to the outside of the device.

On the other hand, the quantized DCT coefficients input to the decoding means 47 are used for generating a reference image. The operation performed by the decoding means 47 is as follows. The inverse quantization means 471 inversely quantizes the quantized DCT coefficient,
The DCT coefficient is decoded and output to the inverse DCT means 472.
The inverse DCT means 472 applies the inverse DCT coefficient to the decoded DCT coefficient.
The difference data is decoded by performing the T processing and output to the adder 473. The adder 473 adds the compensated reference image output from the motion compensation unit 475 and the difference data, and stores the result as a new uncompensated reference image in the frame memory 474. The motion compensating unit 475 obtains the uncompensated reference image from the frame memory 474 and the motion vector from the motion detecting unit 48, performs motion compensation on the uncompensated reference image using the motion vector, and obtains a new Subtractor 4 as compensated reference image
2 and output to the adder 473.

For the sake of simplicity, the above operation has been described on the assumption that the compensated reference image has been input to the motion detecting means 48 and the adder 42 in advance from the initial operation of the apparatus. The actual relationship between the reference image and the compensated reference image is as follows. That is, the signal flow between the adder 473, the frame memory 474, and the motion compensating means 475 forms a closed loop and operates like a feedback circuit. Plays a role in correcting the uncompensated reference image. Therefore, once the encoding device 400 is activated, the decoding means 4
In 7, the role is changed from the one in which the same reference image is not motion-compensated to the one in which the same reference image is motion-compensated every time the video input 40 is processed.

In the encoding operation, there is a case where the moving image data is encoded only from the image information without performing the predictive encoding. In that case, in the above operation,
Only operations of DCT processing, quantization, and variable length coding are performed.

Next, the operation of a conventional decoding apparatus according to the prior art, which does not perform enlargement processing, will be described. FIG. 5 is a block diagram showing a configuration of a conventional decoding device according to the above-mentioned conventional technique. In the figure, variable length decoding (VLD: Variable L
When the code input 51 is input, the length decoding unit 52 performs variable-length decoding on the code input 51, decodes the quantized DCT coefficients and motion vectors, and sends them to the inverse quantization unit 53 and the motion compensation unit 55, respectively. Output. Next, the inverse quantization means 53 decodes the DCT coefficients from the quantized DCT coefficients and outputs the decoded DCT coefficients to the inverse DCT means 54. Further DC
The T means 54 generates difference data from the decoded DCT coefficients and outputs the difference data to the motion compensation means 55. Motion compensation means 5
5 decodes a moving image by adding the motion vector and the difference data, and outputs the decoded image as a video output 56 to the outside of the device. As described above, in decoding, a moving image is decoded by performing a process reverse to that of encoding. Note that, for compressed data that has not been subjected to predictive encoding, the inverse DC
The output of the T means 54 directly becomes a moving image and is output as a video output 56.

In the above-described operations of the encoding device and the decoding device according to the conventional technology, the processing operation is performed in units of 8 × 8 pixel blocks for both encoding and decoding. The image size matches before and after decoding.

Next, the operation of a conventional decoding apparatus having a function of enlarging an image size will be described. FIG.
FIG. 2 is a block diagram showing a configuration of a conventional decoding device having a function of enlarging the image size. In the figure, 6
Reference numeral 4 denotes an enlarged block generation unit that performs an enlargement process of a DCT coefficient. Except for having the enlarged block generation means 64, the configuration of the device is the same as that of the decoding device according to the general example described above. The operation of the conventional decoding device having the function of enlarging the image size configured as described above will be described below.

The operation until the code input 61 is processed by the VLD means 62 and the inverse quantization means 63 to generate the 8 × 8 DCT coefficients is the same as that of the above-mentioned ordinary decoding apparatus. When the 8 × 8 DCT coefficient is obtained, the enlarged block generation unit 64 uses the 8 × 8 DCT coefficient as a low-frequency component as it is, and n ₁ × n ₂ (n ₁ , n
₂ are both integers greater than 8) to generate new DCT coefficients. FIG. 7 shows the 8 × 8 DCT coefficient and the n
FIG. 4 is a conceptual diagram showing a comparison with a ₁ × n ₂ DCT coefficient. As shown in the figure, the configuration of the n ₁ × n ₂ DCT coefficients, the left upper end portion of the n ₁ × n ₂ DCT coefficients n ₁ row n ₂ columns matrixes constituting the entire said 8 × 8 DCT coefficients It is stored as a low-frequency component, and a high-frequency component in which all elements are 0 is added to the remaining area.

Next, the inverse DCT means 65 calculates the above n ₁ × n
_2. Decode n ₁ × n ₂ pixel blocks from DCT coefficients.
Motion compensation means 66, n _1/8 times the motion vector in the vertical direction, which is decoded by the reference image and the VLD unit 62, and a new motion vector enlarged n _2/8 times in the transverse direction, the n
Motion compensation is performed using the ₁ × n ₂ pixel block, and a video output 67 is obtained. At this time, in each pixel block,
N _2/8-pixel size in the vertical direction n _1/8 times in the transverse direction
Because of the doubling, the decoded image is also enlarged to the above size. Note that, for compressed data that has not been subjected to predictive encoding, the output of the inverse DCT means 65 directly becomes pixel data, and then becomes the video output 67.

As an image enlarging method other than the above-described decoding apparatus having the enlarged block generating means 64, there is a method using cell repeat which has a function of enlarging an image twice in both the vertical and horizontal directions. Cell repeat, which uses a normal decoding device, is a method of enlarging an image by displaying each pixel constituting the image four times. FIG. 8 is a conceptual diagram illustrating the principle of image enlargement by cell repeat. As shown in the drawing, each pixel constituting the original image isotropically increases by one pixel after the cell repeat is performed.

[0016]

As described above, in the method of enlarging a moving image by the conventional technique, the enlarging process is performed not at the encoding stage but at the decoding stage. Also, in the above-described method of enlarging a moving image according to the conventional technique, the size of the DCT coefficient itself is enlarged by adding 0 to the high frequency side of the 8 × 8 DCT coefficient by the enlarged block generation means 64, and On the other hand, the image is enlarged by performing an inverse DCT process. In this method, the DCT means 43 used in the encoding device and the inverse DCT means 6 used in the decoding means are used.
5 differs in the order of the pixel block to be handled. This means that the order of the inverse DCT means 65 must be increased in the decoding operation, which means that the amount of arithmetic processing at the time of decoding is increased and the number of frames of images that can be decoded per second is reduced. Cause problems. In addition to the above-described enlargement by the inverse DCT processing in the decoding, there is also an enlargement method by the cell repeat described above. In this case, distortion appears in the 2 × 2 pixel block generated by the cell repeat. However, there is a problem that image quality is deteriorated.

The present invention has been made in view of such a problem, and in order to reduce the load on the decoding means, a video signal coding apparatus, a video signal coding method, and a video signal coding method for increasing the image size on the coding side. It is an object to provide a video signal encoding program medium. It is another object of the present invention to provide a video signal encoding device, a video signal encoding method, and a video signal encoding program medium that can encode an image while changing the resolution of the image.

[0018]

According to a first aspect of the present invention, there is provided a video signal encoding apparatus for encoding a video signal, the video signal being encoded in the encoding order. A first reference image, which is a reference image that is not motion-compensated, is obtained from a reordering unit that is a target image that is an image to be rearranged and encoded, and a decoding unit that will be described later. The second reference image, which is a motion-compensated reference image, is obtained from a motion detection unit that detects the result as a motion vector in comparison with the target image, and a decoding unit. Images and
A subtractor that calculates a difference from the target image and outputs the difference data as difference data;
(n <8 integer) divided into pixel blocks.
n × n DCT coefficients generated from n pixel blocks
nDCT means and quantizing the n × nDCT coefficients,
Quantizing means for obtaining a quantized n × n DCT coefficient, and variable length coding (VLC) means for obtaining the motion vector and the quantized n × n DCT coefficient, performing variable length coding, and forming a coded signal And the motion vector and the quantized n × n DCT coefficient, using the first and second
And decoding means for decoding the reference image.

According to a second aspect of the present invention, there is provided the video signal encoding apparatus according to the first aspect, wherein the motion detecting means includes the first reference image and an image to be encoded. N × n (n is an integer of 0 <n <8)
This is to detect a motion vector for each pixel block.

According to a third aspect of the present invention, in the video signal encoding apparatus according to the first or second aspect, the n × n DCT means converts the n × n DCT coefficient into a low-frequency component thereof. The new 8 × 8 DCT coefficient is newly generated.

According to a fourth aspect of the present invention, in the video signal encoding apparatus according to the third aspect, the quantizing means reduces the amount of the 8 × 8 DCT coefficient when quantizing the 8 × 8 DCT coefficient. Only the n × n DCT coefficients, which are the domain components, are quantized, and all the remaining components are set to 0.

According to a fifth aspect of the present invention, in the video signal encoding apparatus of the fourth aspect, the VLC means performs variable length encoding on the 8 × 8 DCT coefficient. It is.

According to a sixth aspect of the present invention, in the video signal encoding apparatus according to the fifth aspect, the decoding means operates using the quantized 8 × 8 DCT coefficient as an operation unit. Is what you do.

According to a seventh aspect of the present invention, in the video signal encoding apparatus according to any one of the first to sixth aspects, the motion detecting means converts the target image to 8 / n (n Is an integer of 0 <n <8), and further includes a size conversion unit for converting the target image converted by the size conversion unit and the first reference image. , To detect the motion vector.

According to an eighth aspect of the present invention, in the video signal encoding apparatus according to any one of the first to seventh aspects, an image to be encoded is multiplied by n / 8 (where n is It further comprises a thinning-out means for reducing the number to 0 <n <8.

In a video signal encoding method according to a ninth aspect of the present invention, in the video signal encoding method for encoding a video signal, the video images are rearranged in the order of encoding, and the video image is an image to be encoded. A sorting step as an image;
A motion detecting step of obtaining a first reference image which is an output of a decoding step to be described later, which is a non-motion-compensated reference image, comparing the obtained first reference image with the target image, and detecting a result as a motion vector; And a second reference image, which is a motion-compensated reference image, which is an output of the decoding step, and obtains a difference between the second reference image and the target image. A subtraction step of outputting the difference data as n × n (n is 0
<N <8) The pixel block is divided into
Generate n × n DCT coefficients from × n pixel blocks
× nDCT step, a quantization step of quantizing the n × nDCT coefficient to obtain a quantized n × nDCT coefficient, and acquiring the motion vector and the quantized n × nDCT coefficient to obtain a variable length coding And a decoding step of decoding the first and second reference images using the motion vector and the quantized n × n DCT coefficients. It is provided with.

A video signal encoding program recording medium according to claim 10 is a video signal encoding program recording medium recording a video signal encoding program for encoding a video signal, wherein the video signal encoding program comprises: The video image is rearranged in the order of encoding, a rearranging step as a target image which is an image to be encoded, and a first image which is a motion-compensated reference image which is an output of a decoding step described later. Obtaining a reference image, comparing the obtained image with the target image, and detecting the result as a motion vector; and a second reference, which is a motion-compensated reference image, which is an output of the decoding step. A subtraction step of obtaining a difference between the second reference image and the target image and outputting the difference as difference data; Is the difference between data n × n (n 0 <n <
N × nD that divides the pixel block into n) pixel blocks and further generates n × n DCT coefficients from the n × n pixel blocks.
CT step and quantizing the n × n DCT coefficients,
Variable length coding (VLC) in which a quantization step for obtaining a quantized n × n DCT coefficient, the motion vector, and the quantized n × n DCT coefficient are obtained, variable length coding is performed, and a coded signal is obtained. A video signal encoding program comprising a step and a decoding step of decoding the first and second reference images using the motion vector and the quantized n × n DCT coefficients is recorded. .

[0028]

(Embodiment 1) A video signal encoding apparatus according to Embodiment 1 of the present invention performs a moving image enlargement process by enlarging an image block on the encoding side. . FIG. 1 is a block diagram illustrating a configuration of a video signal encoding device according to Embodiment 1 of the present invention. 100
Is an encoding device that receives the video input 10 and encodes it. Reference numeral 11 denotes a rearranging unit that rearranges the video input 10 in the order in which the encoding process is performed, and uses the rearranged image as a target image. 1
Reference numeral 2 denotes a subtractor that calculates a difference between the compensated reference image and the target image. Reference numeral 13 denotes n × n DCT means for performing DCT conversion of the rearranged signal, and
Generate a T coefficient. Numeral 14 denotes a quantizing means, and
Quantizes the nDCT signal. Reference numeral 15 denotes a VLC unit that performs variable-length coding of the quantized n × n DCT coefficients and the motion vector, and outputs the result as a code output 16 to the outside of the coding apparatus 100. Reference numeral 17 denotes a decoding unit, which includes an inverse quantization unit 171, an inverse DCT unit 172, an adder 173, a frame memory 174, and a motion compensation unit 175, and is used for detecting a motion vector and generating the difference data. A reference image, which is an image, is generated. 18 is a motion detecting means,
A motion vector used in the motion compensation of the reference image and the variable length coding operation is detected. Reference numeral 171 denotes an inverse quantization unit that inversely quantizes the DCT coefficient quantized by the quantization unit 14 and decodes the DCT coefficient.
172 is an inverse DCT means for decoding difference data from the DCT coefficients. Reference numeral 173 denotes an adder, which generates a new uncompensated reference image from the decoded difference data and the compensated reference image input from the motion compensation unit 175. A frame memory 174 stores the uncompensated reference image. Reference numeral 175 denotes a motion compensating unit that performs motion compensation on the uncompensated reference image stored in the frame memory 174 using the motion vector to generate a new compensated reference image. The operation of the video signal encoding device according to the first embodiment configured as described above will be described below.

In the description of the first embodiment, for simplicity, an example in which n = 4 is taken as an example of an n × n pixel block. First, when receiving the input of the video input 10, the rearranging means 11 rearranges the video input 10 in the encoding order,
This is output to the subtractor 12 and the motion detecting means 18 as a target image. The motion detecting means 18 compares the uncompensated reference image stored in the frame memory 174 with the target image and detects the distance of motion between the two as a motion vector in the same manner as in the conventional example. Are output to the VLC means 15 and the motion compensation means 175. Subtractor 1
2 obtains the target image and the compensated reference image output from the motion compensating unit 175, calculates the difference between the two, and outputs the result to the n × n DCT unit 13 as differential data. When the n × n DCT means 13 obtains the difference data, it first divides the difference data into 4 × 4 pixel block units. Next, the n × n DCT means performs DCT processing on the divided 4 × 4 pixel block as a unit to obtain a 4 × 4 DCT.
It is converted to T coefficients, which are further converted to 8 × 8 DCT coefficients, and output to the quantization means 14.

Here, the correspondence between the 4 × 4 DCT coefficients and the 8 × 8 DCT coefficients generated therefrom will be described.
FIG. 2 is a conceptual diagram illustrating a process of generating an 8 × 8 pixel block from a 4 × 4 pixel block. As shown in the figure,
The configuration of the 8 × 8 DCT coefficients is such that the entirety of the 4 × 4 DCT coefficients is stored as a low-frequency component in the upper left portion obtained by equally dividing the 8 × 8 matrix forming the 8 × 8 DCT coefficients on the orthogonal coordinate axes. And a high-frequency component in which all elements are 0 is added to the portion. Although n = 4 is set in the above description, n is generally a low-frequency component of 8 × 8 DCT coefficients.
The × nDCT coefficient is always arranged at the upper left end of the 8 × 8 DCT coefficient.

When the 8 × 8 DCT coefficient configured as described above is input, the quantization means 14 performs quantization,
It outputs to the VLC means 15 and the decoding means 17.
The VLC means 15 calculates the quantized 8 × 8 DCT coefficients,
When the motion vector and the motion vector are input, the variable-length coding is performed and output to the outside of the device as a code output 16.

The operation of generating a reference image in the decoding means 17 is performed in the same manner as in the coding apparatus 400 according to the prior art. That is, in the decoding means 17, the inverse quantization means 171 converts the 8 × 8
When a DCT coefficient is input, it is inversely quantized to 8 × 8D
Decode the CT coefficients. The inverse DCT means 172 performs an inverse DCT process on the decoded 8 × 8 DCT coefficient, decodes the difference data, and outputs the result to the adder 173. Adder 173
Adds the compensated reference image output from the motion compensating unit 175 and the difference data, and stores the result in the frame memory 174 as an uncompensated reference image. The motion compensator 175 obtains the uncompensated reference image from the frame memory 174 and the motion vector from the motion detector 18, generates a new compensated reference image from both of them, And output to the adder 173.

The reference image generated by the above operation is an image decoded by processing the above 8 × 8 DCT coefficients, and is four times as large as the original image to be encoded. Size. Therefore, the motion vector detected by the motion detecting means 18 indicates the distance of the motion between the image obtained by enlarging the image to be coded to four times the size and the reference image.

Next, the operation of the decoding device for decoding the video output from the code input will be described. For this, the above-mentioned ordinary decoding device according to the prior art can be used. That is, the ordinary decoding apparatus 500 shown in FIG.
In the operation of (3), the inverse DCT means 53
An 8 × 8 pixel block is generated by performing an inverse DCT process on the T coefficient. In the first embodiment, the size of the pixel block before encoding is determined by the n × n DCT unit 13 as described above. 4 × 4 (because n = 4)
Has been reduced to. Therefore, the image size at the time of decoding is 8 2/4 ² = 4, which is 4 times that before encoding, by taking the ratio between the ^two , and therefore, the 4 times enlarging process of the moving image is achieved by the above-described series of operations. Will be.

The operation in the case where n = 4 has been described above. If n is an integer satisfying 0 <n <8, the input image is formed in a size of 8 ² / n ² times by the same operation. It becomes possible to encode.

As described above, according to the video signal encoding apparatus according to the first embodiment of the present invention, the reordering unit 11 which takes the video input as the target image, and calculates the difference between the uncompensated reference image and the target image. The subtracter 12 performs DCT transform, and 0 <n <
N × n DCT means 13 for generating n × n DCT coefficients by an integer n such as 8, quantization means 14, VLC means 15 for performing variable length coding, decoding for generating compensated and uncompensated reference images Means 17 and a motion detecting means 18 for detecting a motion vector, a pixel block, which is a unit of arithmetic processing, is reduced to 8 / n times in advance, so that a size of one side length can be input. Since it is possible to obtain a coded video having a size of 8 / n times that of an image, it is possible to use a normal decoding device that handles only 8 × 8 DCT coefficients, and to perform a moving image enlargement process with a reduced amount of arithmetic processing of a personal computer. It becomes possible.

In the above 8 × 8 DCT coefficient,
The high frequency component can be set to any number other than 0. Further, the quantizing means 14 can quantize only the n × n DCT coefficient which is a low-frequency component with respect to the 8 × 8 DCT coefficient, and can set the remaining portion to 0 without performing the operation. The operation in this case is as follows. n × nD
The CT means 13 outputs n × n DCT coefficients, and the quantization means 14 quantizes the n × n DCT coefficients. At this time, all high-frequency components formed by portions other than the n × n DCT coefficients are set to 0. For the 8 × 8 DCT coefficient, the VLC means 15 quantizes the n × nD
The variable-length coding is performed with the value of the high frequency component remaining as 0 as the CT coefficient, and the code output is obtained. On the other hand, the decoding means 17 calculates the quantized nx
When an nDCT coefficient is input, it is decoded by inverse DCT processing to obtain a reference image composed of n × n pixel blocks.

As described above, since the reference image generated by the above operation has the same size as the image of the video input 10,
The motion detecting means 18 can obtain a motion vector in units of n × n pixel blocks.

(Embodiment 2) FIG. 3 is a block diagram showing a configuration of a video signal encoding apparatus according to Embodiment 2 of the present invention. Reference numeral 33 denotes a thinning unit that reduces the size of image data. The configuration of the second embodiment is the same as that of the first embodiment except that a thinning unit 33 is added between the subtractor 32 and the n × n DCT unit 34. The operation of the video signal encoding device 300 according to Embodiment 2 configured as described above will be described below.

Here, as in the first embodiment, an example of an n × n pixel block in which n = 4 is used. When the video input 30 is input, the rearranging unit 31 outputs the video input 30 to the subtractor 32 and the motion detecting unit 39 as a target image. The motion detecting means 39 generates a motion vector in the same manner as in the first embodiment. Next, the subtractor 32 generates difference data from the target image and the compensated reference image input from the motion compensation unit 375. Upon obtaining the difference data, the thinning means 33 converts the length of one side of the difference data into a size 4/8 times larger, and obtains n ×
Output to nDCT means. The n × n DCT means 34 divides the thinned data into 4 × 4 pixel blocks, and further performs DCT processing to generate 4 × 4 DCT coefficients. Next, the quantization means 35 quantizes the 4 × 4 DCT coefficient. At this time, based on the quantized 4 × 4 DCT coefficient, the quantization means 35 holds this as it is as a low-frequency component, and A new quantized 8 × 8 DCT coefficient with all 0s is generated. The VLC means 36 calculates the quantized 8 × 8D
When the CT coefficient and the motion vector input from the motion detection means 39 are obtained, these are subjected to variable-length coding and output as a code output 37 to the outside of the device.

Next, the motion vector used in the above operation will be described. As in the first embodiment, the decoding means 38 decodes the uncompensated reference image from the quantized 8 × 8 DCT coefficients, and
84. Further, the motion compensating unit 385 generates a compensated reference image by performing motion compensation on the uncompensated reference image using the motion vector input from the motion detecting unit 39. When the adder 383 obtains the compensated reference image and the difference data, the adder 383 adds the two to generate a new uncompensated reference image and stores it in the frame memory 384.

The new uncompensated reference image whose one-side length is 4/8 = 1/2 times by the thinning means 33 is further multiplied by 8/4 = 2 times by the quantization means 35. It has become something. Therefore, the size of the new uncompensated reference image and the size of the image to be encoded by the reordering unit 31 are the same. Therefore, the motion detecting means 39 is, like the encoding device according to the prior art,
A motion vector is detected in units of 8 × 8 pixel blocks.

In the above description, it is assumed that n = 4. However, if the value of n is an integer of 0 <n <8, the same operation can be performed. It is possible to encode only the components.

As described above, according to the video signal encoding apparatus according to the second embodiment of the present invention, since the thinning means 33 is added to the first embodiment, the thinning processing and the low-order DCT processing are performed. In combination, the amount of calculation processing can be reduced, and in addition, it is possible to encode a moving image by changing the resolution.

In each of the embodiments, the description particularly relates to a video signal encoding device. However, the configuration and operation of the video signal encoding device are described as a video signal encoding method, for example, on a network. The video signal encoding method described above can be utilized in a form stored in a recording medium such as a CD-ROM or a floppy disk as a video signal encoding program.

[0046]

According to the video signal encoding apparatus of any one of the first to sixth aspects, in the video signal encoding apparatus for encoding a video signal, the video signals are rearranged in the order of encoding, and A reordering unit that is a target image that is a target image, and a first reference image that is a motion-compensated reference image obtained from a decoding unit that will be described later.
This is compared with the target image, and a second reference image, which is a motion-compensated reference image, is obtained from a motion detecting unit that detects the result as a motion vector and a decoding unit. 2 and a subtractor that calculates the difference between the reference image and the target image and outputs the difference data as difference data, and calculates the difference data as n × n (n is an integer of 0 <n <8).
N × n DCT means for generating n × n DCT coefficients from the n × n pixel block, and quantizing the n × n DCT coefficients to obtain a quantized n × n
Variable length coding (VLC) that obtains quantization means as DCT coefficients, the motion vector, and the quantized n × n DCT coefficients, performs variable length coding, and obtains coded signals
Means, the motion vector, and the quantized n × nD
By providing decoding means for decoding the first and second reference images using CT coefficients, the block size as an operation unit can be made smaller than a normal 8 × 8 pixel, Since the image can be enlarged in the area and encoded using DCT processing of a low order, it is possible to reduce the amount of arithmetic processing at the time of encoding. Further, since the image is enlarged without performing the interpolation processing on the input image, it is possible to provide an enlarged image without distortion.

According to the video signal encoding apparatus of the present invention, in the video signal encoding apparatus of any one of the first to sixth aspects, the motion detecting means converts the target image to 8 / n (8 / n). n is an integer of 0 <n <8), and further includes a size conversion unit for converting the image to be coded by the size conversion unit and the first reference image. By comparing with the above, it is determined that the motion vector is detected, so that the above effect can be obtained.

According to the video signal encoding apparatus of the present invention, in the video signal encoding apparatus of any one of the first to seventh aspects, the target image is multiplied by n / 8 (n is 0 <
(n <8 is an integer), and further includes a thinning-out unit that encodes the thinned-out image by changing the size of the pixel block according to the reduction ratio. Thus, it is possible to reduce the amount of calculation processing as compared with. In addition, the above-described thinning means and DCT of low order
By combining the means, it is possible to further reduce the amount of arithmetic processing and change the resolution of a moving image.

According to the video signal encoding method of the ninth aspect, in the video signal encoding method for encoding a video signal, the video images are rearranged in the order of encoding, and are images to be encoded. Rearrangement step as a target image;
A motion detecting step of obtaining a first reference image which is an output of a decoding step to be described later, which is a non-motion-compensated reference image, comparing the obtained first reference image with the target image, and detecting a result as a motion vector; And a second reference image, which is a motion-compensated reference image, which is an output of the decoding step, and obtains a difference between the second reference image and the target image. A subtraction step of outputting the difference data as n × n (n is 0
<N <8) The pixel block is divided into
Generate n × n DCT coefficients from × n pixel blocks
× nDCT step, a quantization step of quantizing the n × nDCT coefficient to obtain a quantized n × nDCT coefficient, and acquiring the motion vector and the quantized n × nDCT coefficient to obtain a variable length coding And a decoding step of decoding the first and second reference images using the motion vector and the quantized n × n DCT coefficients. With the provision of (1), the same effect as that by the video signal encoding device according to any one of claims 1 to 6 can be obtained.

According to the video signal encoding program recording medium of the present invention, in a video signal encoding program recording medium recording a video signal encoding program for encoding a video signal, the video signal encoding program is ,
The video image is rearranged in the order of encoding, a rearranging step as a target image which is an image to be encoded, and a first image which is a motion-compensated reference image which is an output of a decoding step described later. Obtaining a reference image, comparing the obtained image with the target image, and detecting the result as a motion vector; and a second reference, which is a motion-compensated reference image, which is an output of the decoding step. A subtraction step of obtaining a difference between the second reference image and the target image and outputting the difference as difference data; and n × n (n is 0 <n)
(Integer of <8) divided into pixel blocks and further divided into n × n
N × n for generating n × n DCT coefficients from a pixel block
A DCT step, and a quantization step of quantizing the n × n DCT coefficients to obtain quantized n × n DCT coefficients;
A variable length coding (VLC) step of obtaining the motion vector and the quantized n × n DCT coefficient, performing variable length coding to obtain a coded signal, the motion vector, and the quantized nx Using the nDCT coefficient, the first,
And a decoding step of decoding the second reference image, whereby the same effect as that by the video signal encoding apparatus according to any one of claims 1 to 6 is recorded. To bring.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a video signal encoding device according to a first embodiment of the present invention.

FIG. 2 is a conceptual diagram showing a method of enlarging a DCT coefficient according to the first embodiment.

FIG. 3 is a block diagram illustrating a configuration of a video signal encoding device according to a second embodiment of the present invention.

FIG. 4 is a block diagram illustrating a configuration of a video signal encoding device according to a conventional technique.

FIG. 5 is a block diagram showing a configuration of a normal decoding device according to a conventional technique.

FIG. 6 is a block diagram illustrating a configuration of a decoding apparatus that enlarges and decodes a moving image according to a conventional technique.

FIG. 7 is a conceptual diagram illustrating a method of expanding a DCT coefficient according to a conventional technique.

FIG. 8 is a conceptual diagram showing an enlargement method by cell repeat according to a conventional technique.

[Explanation of symbols]

 10, 30, 40 Video input 11, 31, 41 Rearrangement means 12, 32, 42 Subtractor 13, 34 n × n DCT means 14, 35, 44 Quantization means 15, 36, 45 VLC means 16, 37, 46 Code Output 17, 38, 47 Decoding means 18, 39, 48 Motion detecting means 33 Decimating means 43 DCT means 100, 300, 400 Encoding device 171, 381, 471 Inverse quantization means 172, 382, 472 Inverse DCT means 173, 383, 473 Adders 174, 384, 474 Frame memory 175, 385, 475 Motion compensating means 500, 600 Decoding device 51, 61 Code input 52, 62 VLD means 53, 63 Dequantizing means 54, 65 Inverse DCT means 55 , 66 Motion compensation means 56, 67 Video output 64 Enlarged block generation means

Claims (10)

[Claims] 1. A video signal encoding apparatus for encoding a video signal, comprising: a rearranging unit that rearranges the video signal in a coding order and sets the video signal as a target image which is an image to be encoded; Obtaining a first reference image which is a reference image not subjected to motion compensation, comparing the obtained first reference image with the target image, and detecting a result of the first reference image as a motion vector; A second reference image, which is the obtained reference image, a subtracter that calculates a difference between the second reference image and the target image, and outputs the difference as difference data, n × n DCT means for dividing into n × n (n is an integer of 0 <n <8) pixel blocks and further generating n × n DCT coefficients from the n × n pixel blocks, and quantizing the n × n DCT coefficients And quantum N × nDC
A quantization means for obtaining a T coefficient; a variable length coding (VLC) means for obtaining the motion vector and the quantized n × n DCT coefficient, performing variable length coding and obtaining a coded signal; A video signal encoding apparatus, comprising: decoding means for decoding the first and second reference images using a vector and the quantized n × n DCT coefficients. 2. The video signal encoding device according to claim 1, wherein the motion detecting means is configured to perform n × n (n is 0 <n <8) between the first reference image and the target image. A video signal encoding apparatus for detecting a motion vector in pixel block units. 3. The video signal encoding device according to claim 1, wherein said n × n DCT means newly generates an 8 × 8 DCT coefficient having said n × n DCT coefficient in its low-frequency component. A video signal encoding device, comprising: 4. The video signal encoding apparatus according to claim 3, wherein the quantizing means quantizes only the n × n DCT coefficient which is a low-frequency component when quantizing the 8 × 8 DCT coefficient. A video signal encoding apparatus for converting all remaining components to zero. 5. The video signal encoding apparatus according to claim 4, wherein said VLC means performs variable length encoding on said 8 × 8 DCT coefficients. . 6. The video signal encoding apparatus according to claim 5, wherein said decoding means operates using said quantized 8 × 8 DCT coefficient as a calculation unit. Encoding device. 7. The video signal encoding apparatus according to claim 1, wherein said motion detecting means has a size of 8 / n times (n is an integer of 0 <n <8) the target image. And a motion vector detecting unit that compares the target image converted by the size converting unit with the first reference image to detect a motion vector. Video signal encoding device. 8. The video signal encoding apparatus according to claim 1, further comprising a thinning unit that reduces the target image to n / 8 times (n is an integer of 0 <n <8). A video signal encoding device. 9. A video signal encoding method for encoding a video signal, wherein the video images are rearranged in the order of encoding and rearranged as a target image which is an image to be encoded. A motion detection step of acquiring a first reference image, which is an output of the step, which is a non-motion-compensated reference image, comparing the first reference image with the target image, and detecting a result as a motion vector; A second reference image, which is a motion-compensated reference image, which is an output of the step, is obtained.
Subtracting the difference between the reference image and the target image and outputting the difference as difference data; dividing the difference data into n × n (n is an integer of 0 <n <8) pixel blocks; An n × n DCT step for generating an n × n DCT coefficient from the n × n pixel block, and quantizing the n × n DCT coefficient to obtain a quantized n × nDCT
A quantization step for obtaining a T coefficient; a variable length coding (VLC) step for obtaining the motion vector and the quantized n × n DCT coefficient and performing variable length coding to obtain a coded signal; Using the motion vector and the quantized n × n DCT coefficients, the first,
And a decoding step of decoding the second reference image. 10. A video signal encoding program recording medium on which a video signal encoding program for encoding a video signal is recorded, wherein the video signal encoding program rearranges the video images in an encoding order and encodes A reordering step as a target image which is an image to be obtained, and a first reference image which is a motion-compensated reference image, which is an output of a decoding step to be described later, is compared with the target image. And obtaining a second reference image, which is a motion-compensated reference image, which is an output of the decoding step, and a motion detection step of detecting the result as a motion vector.
Subtracting the difference between the reference image and the target image and outputting the difference as difference data; dividing the difference data into n × n (n is an integer of 0 <n <8) pixel blocks; An n × n DCT step for generating an n × n DCT coefficient from the n × n pixel block, and quantizing the n × n DCT coefficient to obtain a quantized n × nDCT
A quantization step for obtaining a T coefficient; a variable length coding (VLC) step for obtaining the motion vector and the quantized n × n DCT coefficient and performing variable length coding to obtain a coded signal; Using the motion vector and the quantized n × n DCT coefficients, the first,
And a decoding step for decoding the second reference image. A video signal encoding program recording medium, comprising: