JP3907623B2 - Video signal encoding / decoding device and encoding / decoding method - Google Patents

Description

  The present invention relates to a video signal encoding / decoding device and encoding / decoding method for encoding / decoding a video signal using motion compensation prediction.

  As an encoding means in a video signal encoding / decoding device, a combination of motion compensated prediction and DCT (discrete cosine transform) is often used. The conventional example described below also uses this.

11 to 15 show, for example, a conventional video signal encoding apparatus and a decoding apparatus for reproducing a video signal encoded by the encoding apparatus, as shown in ISO / IEC 13818-2 Draft International Standard. It is a figure for demonstrating.
FIG. 11 is a block diagram showing a schematic configuration of the video signal encoding apparatus, FIG. 12 is a block diagram showing a schematic configuration of the video signal decoding apparatus, and FIGS. 13 and 14 show encoding of the video signal. FIG. 15 is a diagram showing a vector code for encoding (only the vector codes in some search ranges are shown in FIG. 15). is there.
FIG. 16 is a diagram showing a relationship between a difference value of a motion vector, which will be described later, and its appearance probability.

In general, in coding using motion compensation prediction and DCT, one piece of image information is divided into a plurality of grid-like small areas (hereinafter referred to as coding target areas), and coding is performed for each small area. Do.
What is referred to as motion compensated prediction is that several images (hereinafter referred to as reference images) encoded in the past with respect to the encoding target area of an image to be encoded (hereinafter referred to as an encoding target image). The area that is most similar and the same size as the encoding target area (hereinafter referred to as the prediction area) is detected, and only the difference signal between the prediction area and the encoding target area is encoded. To be transmitted.

At this time, it is necessary to simultaneously transmit information on which region is the prediction region to the decoding means, but this information is referred to as a motion vector and is different between an interlaced image and a non-interlaced image. In this conventional example, for the sake of explanation, it is assumed to be composed of two vectors, a horizontal motion vector and a vertical motion vector.
FIG. 14 conceptually shows this motion compensation prediction.

  On the other hand, the decoding means detects the prediction region from the transmitted motion vector and the reproduced reference image, and adds the difference signal transmitted to the video signal of the prediction region. Here, the original signal of the encoding target area can be reproduced.

  Further, as shown in FIG. 13, the prediction area is extended by ± h pixels in the horizontal direction and ± v lines in the vertical direction around the same horizontal / vertical position as the encoding target area of the encoding target image in the reference image. The region is selected from the region (hereinafter referred to as a search region). In general, it is necessary to widen the search area in order to increase the coding efficiency for fast moving images. Therefore, the conventional apparatus is configured so that the size of the search area can be appropriately selected.

Now, a specific configuration of the conventional video signal encoding apparatus will be described with reference to FIG.
In the figure, 1a is an input terminal for a video signal, 2a is an output terminal for an encoded video signal, 3a is a subtracting means, 4a is a DCT means for converting the video signal to a horizontal / vertical spatial frequency for information compression, 5a Is quantization means, 6a is inverse quantization means, 7a is IDCT (Inverse Discrete Cosine Transform) means for reconverting the frequency-converted video signal to the original video signal, 8a is addition means, 9a is memory means, and 12a is Switch (switching) means, 13a is variable length coding means, 14a is transmission buffer means, 15a is code amount control means, and 18a is motion detection means.

A part of the video signal 101 input from the input terminal 1a is input to the motion detection means 18a to generate a motion vector, and is also made a difference signal 102 from the prediction region signal in the subtraction means 3a.
The difference signal 102 is frequency-converted by the DCT means 4a and further quantized by the quantization means 5a.

Then, a part of the quantized difference signal 104 is retransformed through the inverse quantization means 6a and the inverse DCT means 7a to be the original difference signal, and the signal in the prediction region is added by the addition means 8a. And is stored as a reference image in the memory means 9a. On the other hand, the remaining difference signal 104 is encoded and multiplexed together with the motion vector 112 generated by the motion detector 18a in the variable length encoder 13a.
Here, variable-length coding is one of coding methods in which a short code word is assigned to a symbol with a high appearance probability and a long code word is assigned to a symbol with a low appearance probability.

The multiplexed signal 114 is transmitted from the output terminal 2a via the transmission buffer means 14a or recorded on a recording medium (not shown).
The code amount control means 15a adaptively changes the quantization step of the quantization means 5a in response to a signal such as the remaining memory capacity in the transmission buffer means 14a so that overflow does not occur.

  On the other hand, the reference image stored in the memory means 9a is input to the first terminal of the switch means 12a and also input to the motion detection means 18a. (A zero signal is input to the second terminal of the switch means 12a.)

The motion detection means 18a detects the motion vector 112 as described above for each encoding target area of the encoding target image from the input reference image 108 and the video signal (coding target image) 101 (FIG. 13). , FIG. 14).
The detected motion vector 112 is sent to the variable length coding means 13a, where a difference value from the motion vector of the adjacent encoding target region is calculated, and a vector code obtained by variable length coding this is a variable length code. The difference signal 104 is multiplexed.

  The output 110 of the motion detection means 18a is also used as a switching signal for the switch means 12a. Based on this signal, the video signal of the reference image is converted into the signal 109 of the prediction area and input to the subtraction means 3a and the addition means 8a. Is done. Further, the signal 110 is also used as a difference signal 104 and a motion vector coding switching signal in the variable length coding means 13a.

Next, a specific configuration of the video signal decoding apparatus that decodes the video signal encoded as described above will be described with reference to FIG.
In the figure, 1b is an input terminal for an encoded video signal, 2b is an output terminal for a decoded video signal, 14b is a reception buffer means, 13b is a variable length decoding means, 5b is an inverse quantization means, and 4b is an inverse quantization means. IDCT means, 8b is addition means, 9b is memory means, and 12b is switch means.

The encoded video signal 201 input from the input terminal 1b is input to the variable length decoding unit 13b via the reception buffer unit 14b. The variable length decoding means 13 b decodes the encoded video signal 202 and separates it into a motion vector 213 and a difference signal 203.
The separated difference signal 203 is inversely quantized by the inverse quantization means 5b and converted to the original difference signal 205 by the IDCT means 4b. Further, the difference signal 205 is added to the prediction area signal 208 in the adding means 8b and returned to the original encoding target area signal 206, and a part of the difference signal 205 is stored in the memory means 9b. The target image is output from the output terminal 2b.

  On the other hand, the memory unit 9b generates a prediction region signal 207 from the encoding target region signal 206 from the addition unit 8b and the motion vector 213 decoded by the variable length decoding unit, and the prediction region signal 207 is variable length. The adder 8b is configured to input the signal via the switch unit 12b that is switched based on the switching signal 209 for motion compensation prediction generated by the decoding unit. Note that a zero signal is input to one end of the switch means 12b. When this terminal is selected, a signal not subjected to motion compensation prediction is output from 8b.

  The conventional video signal encoding / decoding device is configured as described above, and is configured so that the size of the search area can be set as appropriate in order to increase the encoding efficiency for fast moving video. .

  However, as shown in FIG. 16, when the size of the search area is different, the appearance probability of the difference value of the motion vector is different accordingly. For this reason, when the vector code of the motion vector is created in the variable length coding means 13a, it is necessary to make the vector code different according to the size of the search area.

  Therefore, in the conventional apparatus, as shown in FIG. 15, two codes of motion code (a predetermined variable length code) and motion residual (a code length determined according to the search range) are combined. The vector code is generated according to the above, and by combining these, a different vector code corresponding to the size of the search area is generated.

  Here, as shown in FIG. 16, the difference value of the motion vector has a feature that the appearance probability of the vector having a small difference value is high regardless of the size of the search area. However, since the conventional vector code did not consider this point at all, as shown in FIG. 15, the code length of the vector code indicating a smaller vector difference value becomes longer as the search area becomes larger. have.

  This means that although the search area has been expanded to increase the encoding efficiency for fast moving images, the average word length of the vector code becomes longer due to the expansion. As a result, there has been a problem that the conversion efficiency deteriorates and the image quality deteriorates.

  Further, in the conventional apparatus, different vector codes are generated according to the size of the search area by combining the two codes of motion code and motion residual, so that the optimal vector for any search area In order to obtain a code, it is necessary to have a plurality of types of vector codes corresponding to the size of the search area in parallel. Therefore, there is a problem that the hardware / software scale is inevitably large and is not practical.

  The present invention has been made to solve the problems of the conventional apparatus as described above. Even if the search area is expanded, the encoding efficiency does not deteriorate, and the hardware / software scale is reduced. An object of the present invention is to obtain a video signal encoding / decoding device and encoding / decoding method.

The video signal encoding apparatus according to the reference invention is
The encoding target area is defined as an area obtained by dividing one piece of image information into a plurality of areas.
The upper encoding target area is defined as an area obtained by collecting a plurality of the encoding target areas.
When a reference area is defined as an arbitrary area having the same size as the upper encoding target area in a reference image that is an image encoded in the past,
A first motion detecting unit that detects a first motion vector indicating which of the plurality of reference regions in the reference image is shifted as the upper coding target region;
When a search area is defined as an area having a predetermined size centered on an area specified by the first motion vector in the reference image,
A second motion for detecting a second motion vector indicating which one of the arbitrary regions having the same size as the encoding target region in the search region is shifted. Detection means;
Variable length coding means for coding the first and second motion vectors output from the first and second motion detection means.

The video signal encoding apparatus according to the reference invention is characterized in that the first and second motion vectors are encoded by vector codes corresponding to the first and second motion vectors, respectively.

In the video signal encoding device according to the reference invention, the variable-length encoding unit includes a first motion vector detected for the upper encoding target region and upper encoding adjacent to the upper encoding target region. The difference value from the first motion vector detected for the target area, the second motion vector detected for the encoding target area, and the encoding target area adjacent to the encoding target area A difference value from the second motion vector is coded by a vector code.

The video signal encoding method according to the reference invention is:
The encoding target area is defined as an area obtained by dividing one piece of image information into a plurality of areas.
The upper encoding target area is defined as an area obtained by collecting a plurality of the encoding target areas.
When a reference area is defined as an arbitrary area having the same size as the upper encoding target area in a reference image that is an image encoded in the past,
A first motion detection step of detecting a first motion vector indicating which of the plurality of reference regions in the reference image is shifted as the upper encoding target region;
When a search area is defined as an area having a predetermined size centered on an area specified by the first motion vector in the reference image,
A second motion for detecting a second motion vector indicating which one of the arbitrary regions having the same size as the encoding target region in the search region is shifted. A detection process;
And a variable length encoding step for encoding the first and second motion vectors output from the first and second motion detection means.

The video signal encoding method according to the reference invention is characterized in that the first and second motion vectors are encoded by corresponding vector codes.

The video signal encoding method according to the reference invention includes a first motion vector detected for the upper encoding target region in the variable length encoding step and upper encoding adjacent to the upper encoding target region. The difference value from the first motion vector detected for the target area, the second motion vector detected for the encoding target area, and the encoding target area adjacent to the encoding target area A difference value from the second motion vector is coded by a vector code.

The video signal decoding apparatus according to the present invention is:
The encoding target area is defined as an area obtained by dividing one piece of image information into a plurality of areas.
The upper encoding objective area, defined as the encoding target area a plurality collected area,
A reference area is defined as an arbitrary area having the same size as the upper encoding target area in a reference image that is an image encoded in the past,
Defining a first motion vector as a vector indicating which of the plurality of reference regions in the reference image the upper encoding target region is shifted;
When the second motion vector is defined as a vector indicating which one of the arbitrary regions having the same size as the encoding target region in the reference image is shifted. In addition,
Variable length decoding means for separating and decoding the first and second motion vector codes from the encoded video signal including the first and second motion vector codes corresponding to the first and second motion vectors. When,
Motion vector reproduction means for outputting a motion vector based on the first and second motion vectors corresponding to the first and second motion vector codes output from the variable length decoding means;
Image information reproducing means for reproducing the one piece of image information based on the motion vector output from the motion vector reproducing means ;
It is characterized by providing.
The video signal decoding apparatus according to the present invention is
The encoding target area is defined as an area obtained by dividing one piece of image information into a plurality of areas.
An upper encoding target region is defined as a region in which a plurality of the encoding target regions are collected,
A reference area is defined as an arbitrary area having the same size as the upper encoding target area in a reference image that is an image encoded in the past,
Defining a first motion vector as a vector indicating which of the plurality of reference regions in the reference image the upper encoding target region is shifted;
When the second motion vector is defined as a vector indicating which one of the arbitrary regions having the same size as the encoding target region in the reference image is shifted. In addition,
A variable length decoding step of separating and decoding the first and second motion vector codes from the encoded video signal including the first and second motion vector codes corresponding to the first and second motion vectors. When,
A motion vector reproduction step of outputting a motion vector based on the first and second motion vectors corresponding to the first and second motion vector codes obtained by the variable length decoding step;
An image information reproduction step of reproducing the one piece of image information based on the motion vector obtained by the motion vector reproduction step;
A switch step of switching whether to cause the image information reproducing means to reproduce the image using the first motion vector;
It is characterized by providing.

Further, in the video signal decoding device according to the present invention, the motion vector reproducing means reproduces a motion vector by vector combining the first and second motion vectors ,
It is characterized by.

A video signal decoding method according to the present invention includes:
The encoding target area is defined as an area obtained by dividing one piece of image information into a plurality of areas.
The upper encoding objective area, defined as the encoding target area a plurality collected area,
A reference area is defined as an arbitrary area having the same size as the upper encoding target area in a reference image that is an image encoded in the past,
Defining a first motion vector as a vector indicating which of the plurality of reference regions in the reference image the upper encoding target region is shifted;
When the second motion vector is defined as a vector indicating which one of the arbitrary regions having the same size as the encoding target region in the reference image is shifted. In addition,
A variable length decoding step of separating and decoding the first and second motion vector codes from the encoded video signal including the first and second motion vector codes corresponding to the first and second motion vectors. When,
A motion vector reproduction step of outputting a motion vector based on the first and second motion vectors corresponding to the first and second motion vector codes obtained by the variable length decoding step;
An image information reproduction step of reproducing the one piece of image information based on the motion vector obtained by the motion vector reproduction step ;
It is characterized by including.
The video signal decoding method according to the present invention includes:
The encoding target area is defined as an area obtained by dividing one piece of image information into a plurality of areas.
An upper encoding target region is defined as a region in which a plurality of the encoding target regions are collected,
A reference area is defined as an arbitrary area having the same size as the upper encoding target area in a reference image that is an image encoded in the past,
Defining a first motion vector as a vector indicating which of the plurality of reference regions in the reference image the upper encoding target region is shifted;
When the second motion vector is defined as a vector indicating which one of the arbitrary regions having the same size as the encoding target region in the reference image is shifted. In addition,
A variable length decoding step of separating and decoding the first and second motion vector codes from the encoded video signal including the first and second motion vector codes corresponding to the first and second motion vectors. When,
A motion vector reproduction step of outputting a motion vector based on the first and second motion vectors corresponding to the first and second motion vector codes obtained by the variable length decoding step;
An image information reproduction step of reproducing the one piece of image information based on the motion vector obtained by the motion vector reproduction step;
A switch step of switching whether to cause the image information reproducing means to reproduce the image using the first motion vector;
It is characterized by including.

Further, in the video signal decoding method according to the present invention, in the motion vector reproduction step, the motion vector is reproduced by vector combining the first and second motion vectors ,
It is characterized by.

  According to the present invention, the motion vector of the encoding target area of the image information is represented by the combination of the first motion vector indicating the global motion of the image information and the second motion vector indicating the local motion. Therefore, the code amount of the first motion vector can be reduced, and it is not necessary to provide a plurality of search areas for detecting the second motion vector.

  Further, according to the present invention, even when the first motion vector is not detected, the first motion vector is used by using the other first motion vector based on the characteristics of normal image information such as panning and human visual characteristics. It is possible to create one motion vector or substitute the detected first motion vector.

  According to the present invention, two-stage motion compensation prediction is performed using the first motion vector indicating the global motion and the second motion vector indicating the local motion. Therefore, there is an effect that a video signal encoding / decoding device and encoding / decoding method having high encoding efficiency can be obtained even in a fast-moving image while reducing the hardware / software scale. .

  Further, according to the present invention, even if there is a higher coding target area where the first motion vector cannot be detected, the first motion vector can be easily obtained instead of the first target vector, and the first motion vector obtained by the present invention can be obtained. Since the first motion vector is obtained on the basis of the characteristics of image information such as panning and human visual characteristics, the video signal encoding / decoding device and the code that suppress the deterioration of the reproduction image quality to a slight extent There is an effect that an encoding / decoding method is obtained.

Example 1.
Embodiments of the present invention will be described below with reference to the drawings.
1 to 8 are diagrams for explaining a video signal encoding apparatus according to Embodiment 1 of the present invention and a decoding apparatus that reproduces a video signal encoded by the encoding apparatus.

  FIG. 1 is a block diagram showing a schematic configuration of the video signal encoding device, FIG. 2 is a block diagram showing a schematic configuration of the video signal decoding device, and FIG. 3 shows an example of the configuration of the first motion detection means. It is a block diagram. FIG. 4 is a conceptual diagram showing the concept of the encoding target region and the upper encoding target region in the present embodiment, and FIGS. 5 and 6 illustrate the concept of motion compensation prediction performed when the video signal is encoded in the present embodiment. FIG. 7 is a diagram illustrating a vector code for encoding, and FIG. 8 is a conceptual diagram illustrating a relationship between a first motion vector and an image according to the present embodiment.

As described above with reference to FIG. 16, the motion vector difference value in a normal image has a very high appearance probability of a small vector difference value regardless of the size of the search region. This means that the normal temporal movement of the image often moves in the same manner in a certain area as represented by camera panning and the like. Therefore, in the case of an image with fast motion, even if the motion vector itself takes a large value, the difference value of the motion vector mostly takes a small value.
In addition, considering human visual characteristics, human visual characteristics are relatively good for fast movements in a certain area such as panning, but on the contrary, the local characteristics of the screen For fast movement, human visual characteristics are extremely deteriorated.

The present invention utilizes the characteristics of such an image and human visual characteristics. In the present embodiment, first, an upper encoding target region as shown in FIG. 4 is defined. This is configured by collecting a plurality of conventional encoding target areas.
Then, as shown in FIG. 5, the first motion vector indicating the global motion is detected for each of the upper encoding target areas constituting the encoding target image. Thereafter, as shown in FIG. 6, a predetermined search centered on the starting point of the first motion vector for each of the encoding target regions included in the upper encoding target region in which the first motion vector is detected. A second motion vector in the region is detected, and each of these motion vectors is encoded.

  According to this embodiment, since the first motion vector is detected for a wide area, the code amount of the first motion vector with respect to the entire code amount is very small. In addition, since the second motion vector has already detected a motion of a wide area by the first motion vector, local motion detection within a limited range is performed even for a fast motion image. It is sufficient to set an appropriate search area in advance and detect the second motion vector in this search area. Therefore, there is no need to provide a plurality of vector codes in parallel as in the prior art.

The specific apparatus configuration of this embodiment will be described below. FIG. 1 is a block diagram showing the configuration of such a video signal encoding apparatus according to the first embodiment of the present invention.
In the figure, 1c is an input terminal for a video signal, 2c is an output terminal for an encoded video signal, 3c is a subtracting means, 4c is a DCT means for converting the video signal into a horizontal / vertical spatial frequency for information compression, 5c Is quantization means, 6c is inverse quantization means, 7c is IDCT (Inverse Discrete Cosine Transform) means for reconverting the frequency-converted video signal to the original video signal, 8c is addition means, 9c is memory means, 10c is First motion detection means, 12c is switch (switching) means, 13c is variable length coding means, 14c is transmission buffer means, 15c is code amount control means, and 18c is second motion detection means.

A part of the video signal 301 inputted from the input terminal 1c is inputted to the first motion detecting means 10c and the second motion detecting means 18c, and also inputted to the subtracting means 3c to obtain the signal 309 of the prediction region. The difference signal 302 is used.
The difference signal 302 is frequency-converted by the DCT means 4c and further quantized by the quantization means 5c.

Then, a part of the quantized difference signal 304 is retransformed through the inverse quantization means 6c and the inverse DCT means 7c to be the original difference signal, and the signal 309 in the prediction region is added by the addition means 8c. The original video signal is stored as a reference image in the memory means 9c.
On the other hand, the remaining difference signal 304 is encoded together with the first and second motion vectors 312 and 313 generated by the first motion detecting unit 10c and the second motion detecting unit 18c in the variable length encoding unit 13c. Is multiplexed.

The multiplexed signal 314 is transmitted from the output terminal 2c via the transmission buffer means 14c or recorded on a recording medium (not shown).
Note that the code amount control means 15c adaptively changes the quantization step of the quantization means 5c in response to a signal such as the remaining memory capacity in the transmission buffer means 14c so that overflow does not occur.

  On the other hand, the reference image stored in the memory means 9c is input to the first terminal of the switch means 12c and also input to the second motion detection means 18c. (A zero signal is input to the second terminal of the switch means 12c.)

In the second motion detection unit 18c, the code of the encoding target image is determined from the input reference image 308, the video signal (encoding target image) 301, and the first motion vector 313 generated by the first motion detection unit 10c. For each encoding target region, motion detection is performed within a predetermined search region centered on the region specified by the first motion vector of the higher-order encoding target region to detect a second motion vector (see FIG. 6).
The detected first motion vector 313 and second motion vector 312 are sent to the variable length coding means 13c, where the difference values between the motion vector in the adjacent higher coding target region and coding target region are respectively determined. This is calculated, converted into a vector code by a vector code as shown in FIG. 7, and multiplexed on a variable length encoded difference signal 304.

  The output 310 of the second motion detection means 18c is also used as a switching signal for the switch means 12c. Based on this signal, the video signal 308 of the reference image is converted into a signal 309 in the prediction area, and the subtraction means 3c and addition are performed. The signal 310 is inputted to the means 8c, and the signal 310 is also used as a coding switching signal between the difference signal 304 and the first and second motion vectors in the variable length coding means 13c.

As shown in FIG. 7, in this embodiment, the motion vector is coded by the vector code of the first motion vector and the vector code of the second motion vector.
In the present embodiment, an 8-bit fixed length code is shown as the vector code of the first motion vector. However, the present invention is not limited to this, and other bit lengths or variable length codes may be used.
Further, in the present embodiment, the vector code in the reference search range shown in the conventional example is shown as the vector code of the second motion vector, but the present invention is not limited to this, and may be a vector code in another search range.

Next, a video signal decoding apparatus for decoding the video signal encoded as described above will be described with reference to FIG.
In the figure, 1d is an input terminal for an encoded video signal, 2d is an output terminal for a decoded video signal, 14d is a reception buffer means, 13d is a variable length decoding means, 5d is an inverse quantization means, and 4d is an inverse quantization means. IDCT means, 8d is addition means, 9d is memory means, 12d is switch means, and 17d is motion vector reproduction means.

The encoded video signal 401 input from the input terminal 1d is input to the variable length decoding unit 13d via the reception buffer unit 14d. The variable length decoding means 13d decodes the encoded video signal 402 and separates it into a first motion vector 410, a second motion vector 411, and a difference signal 403.
The separated difference signal 403 is inversely quantized by the inverse quantization means 5d and converted to the original difference signal 405 by the IDCT means 4d. Further, this difference signal 405 is added to the prediction region signal 408 in the adding means 8d and returned to the original encoding target region signal 406, and a part of the difference signal 405 is stored in the memory means 9d and the original encoding is performed. The target image is output from the output terminal 2d.

  On the other hand, the memory unit 9d generates a prediction region signal 407 from the encoding target region signal 406 from the addition unit 8d and the motion vector 412 decoded by the variable length decoding unit and vector-combined by the motion vector reproduction unit 17d. The prediction area signal 407 is input to the adding means 8d via the switching means 12d that is switched based on the switching signal 409 for motion compensation prediction generated by the variable length decoding means. Note that a zero signal is input to one end of the switch means 12d. When this signal terminal is selected, a reproduction signal for which motion compensation prediction is not performed is output from the adder means 8d.

Next, the first motion vector detection method in the present embodiment will be described. FIG. 3 is a diagram showing an example of a specific configuration of the first motion detecting unit 10c shown in FIG.
In the figure, 19c is a low-pass filter (LPF) means, 20c is sub-sampling means, 21c is memory means, and 22c is representative vector detection means.

The video signal 301 input to the first motion detection unit 10c passes through the LPF unit 19c to remove high frequency components, and is subsampled by the subsampling unit 20c to reduce the hardware scale. At this time, since the LPF means 19c is applied as the pre-processing of the sub-sampling, the influence of aliasing distortion on the motion detection can be removed.
The subsampled video signal is stored as a reference image in the memory means 21c and is directly given to the representative vector detecting means 22c. In the representative vector detection unit 22c, the first motion vector as described with reference to FIG. 5 is based on the upper encoding target region of the encoding target image composed of the input video signal and the reference image from the memory unit 21c. Is detected.

FIG. 8 is a diagram showing the relationship between the first motion vector and the image in this embodiment.
In the figure, vertical lines indicate the respective images, horizontal short lines indicate the boundaries of the upper encoding target area, and arrows indicate the first motion vector.
As can be seen from the figure, the first motion vector is detected for all the upper coding target regions.

Example 2
Next, a second embodiment of the present invention will be described.
FIG. 9 is a diagram illustrating a first relationship between the first motion vector and the image.

  In the above-described first embodiment, the case where the first motion vector is detected for all the upper coding target regions of the coding target image has been described. However, the first motion vector is a fast motion of the image. In some cases, it may not be detected depending on the size of the upper encoding target area.

  The present embodiment relates to a method for creating a first motion vector in a higher-order encoding target region in which such a first motion vector is not detected, and is a constant defined between an encoding device and a decoding device. Based on the above rule, the first motion vector of the higher coding target region is created from the first motion vector of the other higher coding target region.

  FIG. 9 shows a case where the first motion vector is detected at a certain image interval. In the figure, the first motion vector is detected for every m images. At this time, in the image in which the first motion vector is not detected, the first motion vector is detected from the first motion vector in the closest future image in which the first motion vector is detected by the following method. Create a vector.

That is, when the first motion vector of the nth image and the n + m image is detected, if the first motion vector in the n + m image is v, the first motion vector vf for the past image in the n + y image. The first motion vector vb for the future image is created as follows.
vf = {y / m} × v
vb = {(my) / m} * (-v)

  The first vector generation method may be basically a method determined according to a common rule between the encoding device and the decoding device, and may be a method other than the above formula.

Example 3
FIG. 10 is a diagram illustrating a third embodiment of the present invention, and illustrates a second relationship between the first motion vector and the image.

Unlike the second embodiment, the present embodiment is a case where the first motion vector is detected only for one higher-order encoding target image among the encoding target images. In such a case, in the present embodiment, the detected first motion vector is substituted as the first motion vector of the other higher-order encoding target image.
Even if it does in this way, as above-mentioned, since a normal image often moves the whole screen as one lump, such as panning, it is not a big problem.

In the second and third embodiments, the first motion vector vector code is multiplexed on the second motion vector or the difference signal vector code in the upper coding target region where the first motion vector is not detected. Needless to say, it is not necessary to transmit the data.
In each of the above-described embodiments, one image, that is, a frame image on TV is used as a unit when detecting the first and second motion vectors. However, the first and second motion vectors are used as a unit of field image. You may make it detect.

It is a block diagram which shows schematic structure of the encoding apparatus in the video signal encoding / decoding apparatus of this invention. It is a block diagram which shows schematic structure of the decoding apparatus in the video signal encoding / decoding apparatus of this invention. It is a block diagram which shows the structure of the 1st motion detection means in the video signal encoding / decoding apparatus of this invention. It is a conceptual diagram which shows the concept of the encoding object area | region and the high-order encoding object area | region in the video signal encoding / decoding apparatus of this invention. It is a conceptual diagram which shows the concept of the 1st motion vector detection in the video signal encoding / decoding apparatus of this invention. It is a conceptual diagram which shows the concept of the 2nd motion vector detection in the video signal encoding / decoding apparatus of this invention, and the relationship between the 1st, 2nd motion vector. It is a figure which shows the vector code in the video signal encoding / decoding apparatus of this invention. It is a figure which shows the relationship between the 1st motion vector and each image in the video signal encoding / decoding apparatus of this invention. It is a figure which shows the method of producing the 1st motion vector which was not detected in the video signal encoding / decoding apparatus of this invention. It is a figure which shows the method of substituting the other detected 1st motion vector for the 1st motion vector which was not detected in the video signal encoding / decoding apparatus of this invention. It is a block diagram which shows schematic structure of the encoding apparatus in the conventional video signal encoding / decoding apparatus. It is a block diagram which shows schematic structure of the decoding apparatus in the conventional video signal encoding / decoding apparatus. It is a conceptual diagram which shows the concept of the search area | region for detecting a motion vector in the conventional video signal encoding / decoding apparatus. It is a conceptual diagram which shows the concept of the motion vector detection in the conventional video signal encoding / decoding apparatus. It is a figure which shows the vector code in the conventional video signal encoding / decoding apparatus. It is a figure which shows the relationship between the difference value of the motion vector in the conventional video signal encoding / decoding apparatus, and its appearance probability.

Explanation of symbols

1a, 1b, 1c, 1d: input terminals, 2a, 2b, 2c, 2d: output terminals, 3a, 3c: subtraction means, 4a, 4c: DCT (discrete cosine transform) means, 4b: IDCT (inverse discrete cosine transform) Means, 5a, 5c: quantization means, 5b, 5d: inverse quantization means, 6a, 6c: inverse quantization means, 7a, 7c: IDCT means, 8a, 8b, 8c, 8d: addition means, 9a, 9b, 9c, 9d: memory means, 10c: first motion detection means, 12a, 12b, 12c, 12d: switch (switching) means, 13a, 13c: variable length encoding means, 13b, 13d: variable length decoding means, 14a, 14c: transmission buffer means, 14b, 14d: reception buffer means, 15a, 15c: code amount control means, 17d: motion vector reproduction means, 18a, 18c: second motion detection means, 1 c: a low pass filter (LPF) means, 20c: sub-sampling means, 21c: memory means, 22c: the representative vector detecting means

Claims (6)

The encoding target area is defined as an area obtained by dividing one piece of image information into a plurality of areas.
The upper encoding objective area, defined as the encoding target area a plurality collected area,
A reference area is defined as an arbitrary area having the same size as the upper encoding target area in a reference image that is an image encoded in the past,
Defining a first motion vector as a vector indicating which of the plurality of reference regions in the reference image the upper encoding target region is shifted;
When the second motion vector is defined as a vector indicating which one of the arbitrary regions having the same size as the encoding target region in the reference image is shifted. In addition,
Variable length decoding means for separating and decoding the first and second motion vector codes from the encoded video signal including the first and second motion vector codes corresponding to the first and second motion vectors. When,
Motion vector reproduction means for outputting a motion vector based on the first and second motion vectors corresponding to the first and second motion vector codes output from the variable length decoding means;
Image information reproducing means for reproducing the one piece of image information based on the motion vector output from the motion vector reproducing means ;
A video signal decoding apparatus comprising:   The encoding target area is defined as an area obtained by dividing one piece of image information into a plurality of areas.
  An upper encoding target region is defined as a region in which a plurality of the encoding target regions are collected,
  A reference area is defined as an arbitrary area having the same size as the upper encoding target area in a reference image that is an image encoded in the past,
  Defining a first motion vector as a vector indicating which of the plurality of reference regions in the reference image the upper encoding target region is shifted;
  When the second motion vector is defined as a vector indicating which one of the arbitrary regions having the same size as the encoding target region in the reference image is shifted. In addition,
  Variable length decoding means for separating and decoding the first and second motion vector codes from the encoded video signal including the first and second motion vector codes corresponding to the first and second motion vectors. When,
  Motion vector reproduction means for outputting a motion vector based on the first and second motion vectors corresponding to the first and second motion vector codes output from the variable length decoding means;
  Image information reproducing means for reproducing the one piece of image information based on the motion vector output from the motion vector reproducing means;
  Switch means for switching whether to cause the image information reproducing means to reproduce the image using the first motion vector;
  A video signal decoding apparatus comprising: The motion vector reproducing means reproduces a motion vector by vector combining the first and second motion vectors ;
The video signal decoding apparatus according to claim 1 or 2 , wherein The encoding target area is defined as an area obtained by dividing one piece of image information into a plurality of areas.
The upper encoding objective area, defined as the encoding target area a plurality collected area,
A reference area is defined as an arbitrary area having the same size as the upper encoding target area in a reference image that is an image encoded in the past,
Defining a first motion vector as a vector indicating which of the plurality of reference regions in the reference image the upper encoding target region is shifted;
When the second motion vector is defined as a vector indicating which one of the arbitrary regions having the same size as the encoding target region in the reference image is shifted. In addition,
The first and the encoded video signal including the first and second motion vector code corresponding to the second motion vector, the variable length decoding step of decoding separating said first and second motion vector code When,
A motion vector reproduction step of outputting a motion vector based on the first and second motion vectors corresponding to the first and second motion vector codes obtained by the variable length decoding step;
An image information reproduction step of reproducing the one piece of image information based on the motion vector obtained by the motion vector reproduction step ;
A video signal decoding method including:   The encoding target area is defined as an area obtained by dividing one piece of image information into a plurality of areas.
  An upper encoding target region is defined as a region in which a plurality of the encoding target regions are collected,
  A reference area is defined as an arbitrary area having the same size as the upper encoding target area in a reference image that is an image encoded in the past,
  Defining a first motion vector as a vector indicating which of the plurality of reference regions in the reference image the upper encoding target region is shifted;
  When the second motion vector is defined as a vector indicating which one of the arbitrary regions having the same size as the encoding target region in the reference image is shifted. In addition,
  A variable length decoding step of separating and decoding the first and second motion vector codes from the encoded video signal including the first and second motion vector codes corresponding to the first and second motion vectors. When,
  A motion vector reproduction step of outputting a motion vector based on the first and second motion vectors corresponding to the first and second motion vector codes obtained by the variable length decoding step;
  An image information reproduction step of reproducing the one piece of image information based on the motion vector obtained by the motion vector reproduction step;
  A switch step of switching whether to cause the image information reproducing means to reproduce the image using the first motion vector;
  A video signal decoding method including: In the motion vector reproduction step, a motion vector is reproduced by vector synthesis of the first and second motion vectors ;
6. The video signal decoding method according to claim 4 or 5, wherein:

Images