Motion picture encoding method and motion picture encoding device

The present invention relates to a moving image coding method and a moving image coding device having an inter-field prediction mode.

Since the amount of moving image data is generally large, high-efficiency coding is performed when the moving image data is transmitted from the transmitting device to the receiving device, or when it is stored in the storage device. Here, "high-efficiency coding" is a coding process for converting a certain data string into another data string, and refers to a process for compressing the amount of the data.

There are two types of moving image data, one that is mainly composed of frames and the other that is composed of fields. Hereinafter, the prior art of the method of compressing the field image will be mainly described. As a highly efficient coding method for moving image data, a frame / field prediction coding method is known.

FIG. 1 shows a block diagram of this frame / field predictive coding. This coding method utilizes the fact that the moving image data has a high correlation in the time direction. To briefly explain the operation of FIG. 1, a difference image between the input original image and the predicted image is generated by the subtractor 39, and the difference image is generated by the orthogonal conversion means 31, the quantization means 32, and the coefficient entropy coding means. Encode at 40. Further, the output of the quantization means 32 is encoded by the inverse quantization means 33 and the inverse orthogonal conversion means 34 to restore the difference image, and the difference image restored by the decoded image generation means 35 and the predicted image used at the time of coding. Restore the image. The restored image is stored in the decoded image storage means 36, the motion vector calculation means 37 calculates the motion vector between the image and the next input image, and the motion vector predicts the motion vector in the prediction image generation means 38. Generate an image. The generated motion vector is encoded by the vector entropy coding means 41 and output via the MUX 42 together with the coefficient coding data encoded by the coefficient entropy coding means 40. That is, since moving image data generally has a high degree of similarity between frame / field data at one timing and frame / field data at the next timing, this property is used in the frame / field prediction coding method. To do. For example, in a data transmission system using a frame / field predictive coding method, motion vector data representing "movement" from an image of a previous frame / field to an image of a target frame / field in a transmitting device, and the previous frame thereof. Generates the difference data between the predicted image of the target frame / field created using the motion vector data from the image of the field and the actual image of the target frame / field, and sends the motion vector data and the difference data to the receiving device. To do. On the other hand, the receiving device reproduces the image of the target frame / field from the received motion vector data and the difference data.

This frame / field predictive coding in FIG. 1 has outlined the frame / field predictive coding, but the frame predictive coding and the field predictive coding will be further described below.

Fig. 2 and Fig. 3 are the above-mentioned ISO / IEC MPEG-2 / MPEG-4 (hereinafter MPEG-2, MPEG-4), and as of August 2002, jointly with ITU-T and ISO / IEC. ITU-T H.264 / ISO / IEC MPEG-4 Part 10 (Advanced Video Coding: AVC) under standardization Final Committee Draft ("Joint Final Committee Draft (JFCD) of Joint Video Specification (ITU-T REC, H. 264 | ISO / IEC 14496-10 AVC) ", JVT-D157, or ISO / IEC JTC1 / SO29 / WG11 MPEG02 / N492, July 2002, Klagenfurt, AT) (hereinafter abbreviated as AVC FCD) , A description of the format for encoding field images.

That is, each frame consists of two fields, namely the Top field and the Bottom field. FIG. 2 is a diagram illustrating the positions of each pixel of luminance and color difference and the fields to which they belong. As shown in FIG. 2, the odd-numbered lines such as the first luminance line (50a), the third luminance line (50b), the fifth luminance line (50c), the seventh luminance line (50d), and so on are Top. The even-th line, which belongs to the field, the luminance second line (51a), the luminance fourth line (51b), the luminance sixth line (51c), the luminance line 8 (51d), and so on, belongs to the Bottom field. Similarly, for the color difference component, the odd-numbered lines such as the color difference 1st line (52a), the color difference 3rd line (52b), etc. belong to the Top field, and the color difference 2nd line (53a), the color difference 4th line (53b) Even-numbered lines such as ... belong to the Bottom field.

The Top field and Bottom field represent images at different times. Next, the spatiotemporal arrangement of the Top field and the Bottom field will be described with reference to FIG. Since FIG. 3 and thereafter, the technique according to the present invention relates to the vertical component of the motion vector. Therefore, in the present specification, the pixels of the horizontal component are not shown, and the horizontal component of the motion vector is not shown. All are described as 0 for convenience. In addition, the positional relationship between the brightness and color difference pixels of each field is correctly illustrated.

In FIG. 3, the vertical axis represents the pixel position of the vertical component of each field, and the horizontal axis represents the passage of time. Since there is no displacement of the position due to the field in the horizontal component of the pixel of each image, the illustration and description of the pixel in the horizontal direction are omitted in this figure.

As shown in FIG. 3, the pixel position of the color difference component is shifted by 1/4 pixel from the pixel position in the luminance field. This is because when the frame is composed of both the Top and Bottom fields, the relationship of the pixel positions as shown in FIG. 2 is satisfied. It takes about 1/60 second between the adjacent fields of Top and Bottom (64a: 65a, 65a: 64b ...) based on the NTSC format. The time from the Top field to the Top field (64a: 64b ...) Or from the Bottom field to the Bottom field (65a: 65b ...) is about 1/30 second.

Hereinafter, the frame prediction coding mode and field prediction of the field image adopted in MPEG-2 and AVC FCD will be described. FIG. 4 illustrates a method of constructing a frame from two consecutive fields (adjacent Top and Bottom fields) in the frame prediction mode.

As shown in this figure, the frame is reconstructed from two fields (Top and Bottom fields) that are continuous in time. Figure 5 illustrates the frame prediction mode. In this figure, it is assumed that each frame 84a, 84b, 84c ... Has already been reconstructed from two consecutive fields (Top and Bottom fields) as described in FIG. In this frame prediction mode, coding is performed on the coded target frame composed of both the Top and Bottom fields. Then, as the reference image, one reference frame is constructed from the two fields (Top and Bottom fields) accumulated for continuous reference, and is used for the prediction of the pre-coded target frame. Then, these two frame images are encoded according to the block diagram shown in FIG. In the case of this frame prediction coding mode, regarding the method of expressing the motion vector, the zero vector, that is, (0,0) points to the pixels at the same spatial position. Specifically, Frame The motion vector pointing to the motion vector (0,0) points to the pixel position 81 of Frame # 1 (84a) with respect to the pixel 82 having the brightness belonging to # 2 (84b).

Next, the field prediction coding mode will be described. FIG. 6 is a diagram illustrating a prediction method in the inter-field prediction mode. In the field prediction mode, the coding target is a single Top field (94a, 94b ...) or Bottom field (95a, 95b ...) input as the original image. Then, as the reference image, the Top field or Bottom field accumulated in the past can be used. Here, it is generally defined that the original image field and the reference field have the same parity, that the original image field and the reference field are both the Top field or both the Bottom field. For example, the field prediction of 90 with the same parity in the figure is the Top field for both the original image (94b) and the reference (94a). Similarly, different parity between the original image field and the reference field is generally defined as one of the original image field and the reference field being the Top field and the other being the Bottom field. For example, in the field prediction of different parity shown in 91 in the figure, the original image is Bottom. The field (95a) and the reference are the Top field (94a). Then, the original field image and the reference field image are encoded according to the block diagram shown in FIG.

In the conventional technique, a motion vector is obtained based on the positions of pixels in each frame / field in both the frame mode and the field mode. A motion vector calculation method and a pixel associative method when a motion vector is given in the conventional method will be described.

FIG. 7 is a diagram defining the coordinates of the frame / field image, which is widely used in coding such as MPEG-2, MPEG-1, AVC FCD. In the figure, the white circle is the target frame / field, which is the pixel definition position (181). Here, regarding the coordinates in this frame / field image, the origin (0,0) is set at the upper left of the screen, and the pixel definition positions are 1, 2, 3 ... in the horizontal and vertical directions. A value is assigned. That is, the coordinates of the nth pixel in the horizontal direction and the mth pixel in the vertical direction are (n, m). According to this, the coordinates of the position where the pixels are interpolated between the pixels are also defined in the same manner. That is, the position (180) of the black circle in the figure is 1,5 pixels in the horizontal direction and 2 pixels in the vertical direction from the upper left pixel, so the coordinates of the position (180) are (1.5, 2.0). It is expressed as. In the field image, there are only half the pixels of the frame image in the vertical direction, but even in this case, the same treatment as in FIG. 7 is performed based on the positions of the pixels existing in each field.

The definition of the motion vector between fields will be described using the coordinate system of FIG. 7. FIG. 8 is a diagram illustrating a conventional method of calculating a motion vector between corresponding pixels between fields. To define a motion vector, you need the position of the reference source and the position of the reference destination. Then, the motion vector is defined between these two points. Here, the motion vector between the point where the coordinate 201 in the reference source field is (Xs, Ys) and the point where the coordinate 202 in the reference destination field is (Xd, Yd) is obtained. In the conventional method of calculating the motion vector between pixels corresponding to the fields, the motion vector is obtained by the same method described below regardless of whether the reference source and the reference destination are the Top field or the Bottom field. Was there. That is, the reference source field coordinates 201 (Xs, Ys) and the reference destination field coordinates 202 (Xd, Yd) are input to the motion vector calculation means 200, and (Xd-Xs, Yd) is used as the motion vector 203 between the two points. -Ys) is given.

Further, FIG. 9 is a diagram illustrating a method of calculating pixels indicated by motion vectors defined between fields in the prior art. Here, it is assumed that the motion vector is derived by the method of FIG. 8 described above. In order to obtain the coordinates of the reference destination, the position and motion vector of the reference source are required. In the case of this figure, the motion vector 211 (X, Y) is given to the point where the coordinates 212 in the reference source field are (Xs, Ys), and in the reference field obtained by using both of them. It is assumed that the coordinates of In the conventional method of calculating the motion vector between pixels corresponding to between fields, the position of the reference field is described below regardless of whether the reference source and the reference destination are the Top field or the Bottom field. Was required. That is, the motion vector 211 (X, Y) and the reference source field coordinates 212 (Xs, Ys) are input to the pixel mapping means 210, and the coordinates (Xs + X, Ys + Y) are given as the reference destination field coordinates 213. It is to be done.

The definition of the relationship between the vector and the pixel position in FIG. 9 above is the same for the luminance component and the color difference component. Here, in MPEG-1 / MPEG-2 / AVC FCD, which is a general moving image coding method, only the luminance component is encoded in the vector, and the vector of the color difference component is derived by scaling the luminance component. Will be done. Especially in AVC FCD, the number of vertical pixels and the number of horizontal pixels are half the number of pixels of the luminance component, so the motion vector for obtaining the predicted pixel of the luminance component accurately determines the motion vector of the luminance component. It is defined as being scaled in half.

FIG. 10 is a diagram illustrating a method of obtaining a color difference component motion vector from such a conventional luminance component motion vector. That is, when the brightness motion vector 221 is (MV_x, MV_y) and the color difference motion vector 222 is (MVC_x, MVC_y), the color difference component motion vector 222 is generated by the color difference component motion vector generation means 220 (MVC_x, MVC_y). ) = (MV_x / 2, MV_y / 2) It is calculated according to the formula (Equation 1). In this derivation method, in the conventional method, it does not matter whether the motion vector makes a prediction between the fields having the same parity or between the fields having different parity.

In AVC FCD, it is possible to take 1/4 pixel accuracy as the accuracy of the motion vector of the luminance component. From this, as a result of Equation 1, as the accuracy of the motion vector of the color difference component, a vector having an accuracy of 1/8 pixel after the decimal point can be taken.

A method of calculating the interpolated pixels of the color difference component defined by AVC FCD will be described with reference to FIG. In the figure, black circles indicate integer pixels, and dotted white circles indicate interpolated pixels. Here, the coordinates of the interpolation pixel G (256) are the horizontal coordinates of the points A (250) and the points C (252) divided internally into α: 1-α, and the coordinates in the vertical direction are. , It is assumed that the vertical coordinates of the point A (250) and the point B (251) are internally divided into β: 1-β. Here, α and β are values of 0 or more and less than 1. When calculating the interpolated pixel G (256) defined at the above position, use the surrounding integer pixels A (250), B (251), C (252), D (253) and α, β. Therefore, it is required as follows.

G = (1-α) (1-β) A + (1-α) β B + α (1-β) C + α β D (Equation 2) Color difference using Fig. 11 The method of interpolating the pixel of the component is an example for obtaining the interpolated pixel, and there is no problem even if another calculation method is used.

<p> In this field coding mode, in the prediction between the fields in which the original field and the reference field are different, that is, the fields having different parity, the zero vectors of the motion vectors of both the luminance component and the color difference component are not parallel in the definition of AVC FCD. That is, according to the conventional definition, when the prediction is made using the motion vector of the color difference component obtained from the motion vector of the luminance component, the pixels at positions spatially deviated from the luminance component are used. This will be described with reference to FIG. In the figure, it is assumed that the Top field 130, the Bottom field 131, and the Top field 132 are consecutive in time with time. Here, Bottom field 131, Top I am trying to code using field 130. At this time, in the inter-field coding, the motion vector between the same lines in each field is defined as zero in the vertical direction. Therefore, when a zero vector (0,0) is assigned to the second line pixel 133a of the brightness belonging to the Bottom field 131, this pixel is predicted from the pixel 135a of the second line of the brightness of the Top field 130. Will be done. Similarly, if a zero vector (0,0) is assigned to the pixel 134a of the color difference of the first line belonging to the Bottom field 131, this pixel is from the pixel 137a of the first line of the color difference of the Top field 130. is expected. Similarly, the third line pixel 133b of brightness and the second line pixel 134b of color difference belonging to the Top field 132 are the third line pixel 135b of brightness and the second line pixel of color difference on the Bottom field 131, respectively. Predicted from 137b. Originally, it is preferable that the motion vectors of the color difference and the luminance are parallel. Therefore, if the motion vector of the luminance is left as it is, the pixels of the original color difference 134a and 134b are predicted from the positions of 136a and 136b, respectively. It should be done.</p><p> As mentioned above, in the prediction between fields with different parity: -The zero vectors of luminance and color difference are not parallel. I explained that. This causes the following problems in AVC FCD for all vectors in the prediction between fields with different parity. 13 and 14 illustrate this problem. Show the problem according to the AVC FCD. Since the subject of the present invention relates only to the vertical component of the motion vector, in the following description, all the horizontal components of the motion vector are set to 0 for convenience.</p><p> FIG. 13 is a diagram illustrating a problem in obtaining a color difference component motion vector from a luminance component motion vector in the prior art when the reference destination is the Bottom field and the reference source is the Top field.</p><p> In AVC FCD, as shown in Equation 1, the number of vertical pixels and the number of horizontal pixels are half the number of pixels of the luminance component, so the motion vector for obtaining the predicted pixel of the luminance component is the motion of the luminance component. It is defined as a half-scaled vector. It does not matter whether the motion vector makes predictions between frames, between fields with the same parity, or between fields with different parity.</p><p> Now, it is shown that this definition becomes a problem when obtaining the motion vector of the color difference from the motion vector of the luminance defined between the fields of different parity. In FIG. 13, the pixel 140 of the first line of the reference source Top field luminance component has (0,1) as a prediction vector, and as a result, the pixel position 141 of the second line of the reference destination Bottom field luminance component is predicted. Point as a value.</p><p> In this case, the motion vector of the color difference pixels belonging to the same block is obtained as the motion vector (0, 1/2) according to Equation 1. Then, when the motion vector (0,1 / 2) is used as the predicted value of the pixel 142 of the first line of the reference source Top field color difference component, 1 from the pixel of the first line of the reference Bottom field color difference component is used. The pixel position 143, which is shifted downward by / 2 pixels, is used as the predicted value.</p><p> In this case, the luminance motion vector (0,1) and the color difference motion vector (0,1 / 2) are not parallel. Preferably, it is necessary to use the predicted pixel position 145 of the color difference component of the referenced Bottom field to which the motion vector of the color difference parallel to the motion vector of the luminance is applied.</p><p> FIG. 14 is a diagram illustrating a problem in obtaining a color difference component motion vector from a luminance component motion vector in the prior art when the reference destination is the Top field and the reference source is the Bottom field. Similar to the description of FIG. 13, in FIG. 14, the pixel 150 of the first line of the reference source Bottom field luminance component has (0,1) as a prediction vector, and as a result, the two lines of the reference Top field luminance component. The pixel position 151 of the eye is indicated as a predicted value.</p><p> In this case, the motion vector of the color difference pixels belonging to the same block is obtained as the motion vector (0, 1/2) according to Equation 1. Then, when the motion vector (0,1 / 2) is used as the predicted value of the pixel 152 of the first line of the reference source Bottom field color difference component, 1 from the pixel of the first line of the reference Top field color difference component. The pixel position 153 shifted downward by / 2 pixels will be used as the predicted value.</p><p> In this case, the luminance motion vector (0,1) and the color difference motion vector (0,1 / 2) are not parallel. Preferably, it is necessary to use the predicted pixel position 155 of the color difference component of the reference Top field to which the motion vector of the color difference parallel to the motion vector of the luminance is applied.</p><p> As described above, when the parity of the reference destination and the reference source field are different, the conventional prediction method refers to the pixel at the position deviated by the brightness and the color difference, and not only the zero vector but all the vectors. , In the predicted image, the predicted image is deviated due to the brightness and the color difference.</p><p> In the above description, the motion vector of the brightness and the motion vector of the color difference are parallel or not parallel when the direction in the time direction, that is, the time axis direction from the reference source field to the reference destination field is also included in the motion vector. I use it to mean. The same applies to the following description.</p><p> An object of the present invention is to provide a moving image coding method capable of improving the prediction efficiency of color difference components and improving the coding efficiency in coding between different field images, and a moving image coding apparatus. That is.</p>

<p> The present invention solves the above problems. In the motion image coding method of the present invention, motion compensation prediction between fields is performed for a motion image signal in which each frame is composed of two fields and the number of pixels of the vertical component of color difference and the number of pixels of the vertical component of brightness are different. In the moving image coding method in which the above is performed and the coding process is performed, the motion vector of the brightness component indicating the vertical movement of one pixel of the brightness component of the field image in units of 4 is MVy and the vector component. When the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image is MVCy, the reference field and the reference source field are the Top field or Bottom field, respectively. According to either, when both the referenced field and the referencing field are Top fields or Bottom fields, the referenced field is the Top field and the referencing field is the Bottom field, based on the calculation expressed by MVCy = MVy. Sometimes MVCy = MVy- Based on the calculation represented by 2, when the referenced field is the Bottom field and the referencing field is the Top field, the motion of the color difference component from the motion vector of the luminance component is based on the calculation represented by MVCy = MVy + 2. Generate a vector.</p><p> The motion image coding device of the present invention predicts motion compensation between fields for a motion image signal in which each frame is composed of two fields and the number of pixels of the vertical component of color difference and the number of pixels of the vertical component of brightness are different. In the motion image coding method that performs the coding process, the means for determining whether the reference field and the reference source field are either the Top field or the Bottom field, and the value of the vector component is in units of 4. Shows the vertical movement of one pixel of the brightness component of the field image. The motion vector of the brightness component is MVy, and the value of the vector component is 8 as a unit. When the motion vector of the color difference component is MVCy and the reference field and the reference source field are both Top fields or Bottom fields, the reference field is the Top field based on the calculation expressed by MVCy = MVy. When the referencing field is a Bottom field, MVCy = MVy-2 When the referenced field is the Bottom field and the referencing field is the Top field based on the calculation represented by, the motion vector of the luminance component to the motion vector of the color difference component is based on the calculation represented by MVCy = MVy + 2. Has a means to generate. According to the present invention, the color difference component motion vector generated by the method suitable for each is used according to the parity of the reference field and the reference source field. It is possible to solve the problem of the color difference component motion vector caused by the method of assigning to the Bottom field.</p>

<p> According to the present invention, it is possible to obtain the motion vector of the color difference component parallel to the motion vector of the luminance component even between fields having different parity, and the reference pixel position of the luminance component and the color component, which has been a problem in the conventional method. It is possible to solve problems such as deviation.</p>

First, an embodiment of the present invention in coding will be described. In the embodiment of the present invention, in a motion image coding method for predicting motion compensation between fields for a motion image signal composed of a plurality of fields, a plurality of motion vectors of color difference components are generated from motion vectors of brightness components. The color difference component motion vector generation means of the above is provided, and further, the selection means for selecting the color difference component motion vector generation means used for generating the color difference component motion vector by the parity of the reference destination field and the reference source field of the motion vector is further provided. It is characterized in that the prediction vector of the color difference component is generated from the motion vector information of the brightness information by the color difference component motion vector generation means selected in. Here, in the selection means, one that generates a motion vector of a color difference component parallel to the luminance component is selected.

If the color difference component motion vector from the reference field to the reference field is parallel to the brightness component motion vector from the reference field to the reference field, the brightness component motion vector and color difference from the reference field to the reference field Since the spatial displacements of the component motion vectors are the same, that is, the spatial positional relationship between the brightness component motion vector and the color difference component motion vector is maintained, there is no color shift between the fields.

Here, what is important is that in the prior art, even if the luminance component motion vector and the color difference component motion vector as mathematical expressions are parallel, the relationship between the luminance pixels and the color difference pixels constituting each field is mapped. Sometimes it means that they are not parallel.

Here, the above-mentioned plurality of color difference component motion vector generation means include the following three types. First, the first color difference component motion vector generation means is selected by the selection means when the reference destination field and the reference source field have the same parity. The second color difference component motion vector generation means is selected by the selection means when the reference destination field is the Top field and the reference source field is the Bottom field. The third color difference component motion vector generation means is selected by the selection means when the reference destination field is the Bottom field and the reference source field is the Top field.

The method of finding the motion vector of the color difference component parallel to the motion vector of the luminance component depends on the parity of the reference field and the reference field of the motion vector. If both fields have the same parity, the former is the Top field and the latter. Is a Bottom field, and the former is a Bottom field and the latter is a Top field. The calculation method differs depending on the three types. Therefore, in the embodiment of the present invention, an appropriate one is selected from the means for generating the motion vector of the three kinds of luminance components and the vector of the color difference component parallel to the motion vector of the three kinds of luminance components according to the reference source and the reference destination fields. , Generates a motion vector of color difference components.

Specifically, when the reference destination field and the reference source field have the same parity, in the first color difference component motion vector generation means, the value of the vector component is in units of 1, and the vertical of one pixel of the brightness component of the field image. When the movement vector of the brightness component indicating the movement in the direction is MVy and the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image is MVCy in units of 1 as the vector component value, MVCy = Calculate as MVy ÷ 2.

Then, when the reference destination field is the Top field and the reference source field is the Bottom field, the second color difference component motion vector generation means has a vector component value of 1 as a unit for one pixel of the brightness component of the field image. MVCy when the motion vector of the brightness component indicating vertical movement is MVy and the motion vector of the color difference component indicating vertical movement of one pixel of the color difference component of the field image is MVCy with the vector component value in units of 1. Calculate as = MVy ÷ 2 + 0.25.

When the reference destination field is the Bottom field and the reference source field is the Top field, the third color difference component motion vector generation means has a vector component value of 1 as a unit for one pixel of the brightness component of the field image. MVCy when the motion vector of the brightness component indicating vertical movement is MVy and the motion vector of the color difference component indicating vertical movement of one pixel of the color difference component of the field image is MVCy with the vector component value in units of 1. = MVy ÷ 2-0.25.

Further, depending on the definition, the unit indicating the motion of one pixel of the luminance component motion vector and the color difference component motion vector may be different. Here, it is assumed that when the definition of the brightness component motion vector changes by 4, it represents the movement in the brightness image for one pixel, and when the definition of the color difference component motion vector changes by 8, it represents the movement of one pixel. When the motion in the color difference image is represented, when the reference field and the reference source field have the same parity, the first color difference component motion vector generation means sets the motion vector of the brightness component to MVy and the motion vector of the color difference component to MVy. When MVCy is set, it is calculated as MVCy = MVy.

In the case of the same vector definition, when the reference field is the Top field and the reference source field is the Bottom field, the second color difference component motion vector generation means sets the luminance component motion vector to MVy and the color difference component. When the motion vector is MVCy, it is calculated as MVCy = MVy + 2.

Further, in the case of the same vector definition, when the reference field is the Bottom field and the reference source field is the Top field, the third color difference component motion vector generation means sets the luminance component motion vector to MVy and the color difference component. When the motion vector is MVCy, it is calculated as MVCy = MVy 2.

Further, since the coding method of the embodiment of the present invention can also be used as a decoding method , the decoding device basically has the same function as the coding device and operates in the same manner. In the following embodiments, a coding device will be mainly described. Since the present invention relates to the vertical component of the motion vector, all the horizontal components of the motion vector are set to 0 for convenience. Further, the embodiment relating to the decoding device also has the same configuration as the embodiment of the coding device.

Hereinafter, embodiments will be described assuming a case where the present invention is applied to AVC FCD. FIG. 15 is a diagram illustrating a method of calculating a color difference component motion vector from a luminance component motion vector according to the embodiment of the present invention. In the embodiment of the generation means for generating the motion vector of the color difference component from the motion vector of the luminance component in the field prediction in the present embodiment, the generation means is derived from three types of motion vector generation means for the color difference component and one selection means. It is composed.

The operation of the embodiment of the present invention in FIG. 15 will be described below. First, let the motion vector 231 of the given luminance component be (MV_x, MV_y). Then, the vector of the luminance component is given as an input of the first color difference component motion vector generation means 233, the second color difference component motion vector generation means 234, and the third color difference component motion vector generation means 235. Then, each output is input to the selection means 230. Then, in the selection means 230, based on the information of the parity 237 of the reference source field of the input motion vector and the parity 238 of the reference destination of the motion vector, the first, second, and third color difference component motion vector generation means Select one of the outputs and output it as the vector component (MVC_x, MVC_y) of the motion vector 232 of the color difference component.

FIG. 16 is a diagram illustrating a first color difference component motion vector generation means. In the present embodiment, the motion vector 261 of the brightness having the vector value of (MV_x, MV_y) is input to the first color difference component motion vector generating means 260, and the motion vector 261 having the vector value of (MVC1_x, MVC1_y) is obtained. It represents that the motion vector candidate 262 of one color difference is output. Then, the first color difference motion vector candidate 262 is calculated by the color difference component motion vector generating means 260 from the luminance motion vector 261 by the following equation: (MVC1_x, MVC1_y) = (MV_x / 2, MV_y / 2) ... It is calculated according to (Equation 3). Then, the obtained motion vector candidate 262 of the first color difference is output to the selection means.

FIG. 17 is a diagram illustrating a second color difference component motion vector generation means. In the present embodiment, the motion vector 271 of the brightness having the vector value of (MV_x, MV_y) is input to the second color difference component motion vector generating means 270, and the second color difference component motion vector generating means 270 has the vector value of (MVC2_x, MVC2_y). It represents that the motion vector candidate 272 of the second color difference is output. Then, the second color difference motion vector candidate 272 is calculated by the color difference component motion vector generation means 270 from the luminance motion vector 271 by the following equation, (MVC2_x, MVC2_y) = (MV_x / 2, MV_y / 2 + 1 /. 4) Calculated according to (Equation 4). Then, the obtained motion vector candidate 272 with the second color difference is output to the selection means.

FIG. 18 is a diagram illustrating a third color difference component motion vector generation means. In the present embodiment, a motion vector 281 having a brightness having a vector value of (MV_x, MV_y) is input to the third color difference component motion vector generating means 280, and a motion vector 281 having a vector value of (MVC3_x, MVC3_y) is input. It shows that the motion vector candidate 282 of the three color differences is output. Then, the third color difference motion vector candidate 282 is calculated by the color difference component motion vector generation means 280 from the luminance motion vector 281 by the following equation: (MVC3_x, MVC3_y) = (MV_x / 2, MV_y / 2 - 1 / 4) Calculated according to (Equation 5). Then, the obtained motion vector candidate 282 of the third color difference is output to the selection means.

FIG. 19 is a diagram illustrating an embodiment of the selection means 240 in the present invention. First, in the present embodiment, the parity 247 of the motion vector reference source field and the parity 248 of the motion vector reference destination field are each determined by the condition determination table 241, and the selection information 249 of the color difference component motion vector generation means to be selected. Is output. In the present embodiment, when this condition determination table 241 is used, selection information for selecting the first color difference component motion vector candidate 244 is output when both the reference destination field and the reference source field are equal. When the reference destination field is the Top field and the reference source field is the Bottom field, selection information for selecting the second color difference component motion vector candidate 245 is output. Then, when the reference destination field is the Bottom field and the reference source field is the Top field, selection information for selecting the third color difference component motion vector candidate 246 is output.

Here, the first color difference component motion vector candidate 244 is in 262 in FIG. 16, the second color difference component motion vector candidate 245 is in 272 in FIG. 17, and the third color difference component motion vector candidate 246 is in 282 in FIG. , Each is connected. Then, the selector 243 selects one of the first color difference component motion vector candidate 244, the second color difference component motion vector candidate 245, and the third color difference component motion vector candidate 246 according to the above-mentioned selection information 249. Output (MVC_x, MVC_y) as the motion vector 242 of the color difference component.

FIG. 20 is a diagram showing an example of calculating a color difference component vector from a luminance component vector when the reference destination is the Bottom field and the reference source is the Top field according to the embodiment of the present invention.

In the example of this figure, the luminance motion vector (MV_x, MV_y) that predicts the pixel 160 of the reference source Top field luminance component is (0, 1). In this case, the pixel position 161 of the referenced Bottom field luminance component is selected for predicting the luminance pixel 160. With respect to such a vector, a process of obtaining a color difference component motion vector for use in predicting the pixel 162 of the reference source Top field color difference component according to the configuration of FIG. 15 of the present embodiment will be described below.

First, in the case of FIG. 20, the referenced field is the Bottom field and the referenced field is the Top field. Therefore, the third color difference component motion vector candidate is selected as the selection information 249 by the condition determination table 241 of FIG. Here, according to Equation 5, the third color difference component motion vector candidate is (MVC3_x, MVC3_y) = (MV_x / 2, MV_y / 2-1 / 4) = (0/2, 1/2 --1 / 4) = (0, 1/4) (Equation 6). Then, this value is output as the motion vector 242 of the color difference component in FIG. When this vector (0, 1/4) is applied to the pixel 162 of the reference source Top field color difference component, the pixel position 163 of the reference destination Bottom field color difference component is used as the predicted value. In FIG. 20, the vertical positional relationship of each pixel corresponds to the actual case. As can be seen in this figure, the luminance component motion vector (0,1) and the color difference component motion vector (0,1 / 4) are parallel. As a result, the present invention eliminates the color shift between the luminance component and the color difference component, which has been a problem in the prior art.

Similarly, FIG. 21 is a diagram showing an example of calculating a color difference component vector from a luminance component vector when the reference destination is the Top field and the reference source is the Bottom field according to the embodiment of the present invention.

In the example of this figure, the luminance motion vector (MV_x, MV_y) that predicts the pixel 170 of the reference source Bottom field luminance component is (0, 1). In this case, the pixel position 171 of the reference destination Top field luminance component is selected for predicting the pixel 170 of the reference source Bottom field luminance component. With respect to such a vector, a process of obtaining a color difference component motion vector to be used for predicting the pixel 172 of the reference source Bottom field color difference component will be described below according to the configuration of FIG. 14 of the present embodiment.

First, in the case of FIG. 21, the referenced field is the Top field and the referenced field is the Bottom field. Therefore, the second color difference component motion vector candidate is selected for the selection information 249 by the condition determination table 241 of FIG. Here, according to Equation 4, the second color difference component motion vector candidate is (MVC2_x, MVC2_y) = (MV_x / 2, MV_y / 2 + 1/4) = (0/2, 1/2 + 1 / 4) = (0, 3/4) (Equation 7). Then, this value is output as the motion vector 242 of the color difference component in FIG. When this vector (0, 3/4) is applied to the pixel 172 of the reference source Bottom field color difference component, the pixel position 173 of the reference destination Top field color difference is used as the prediction value as the position used for prediction. In FIG. 21, the vertical positional relationship of each pixel corresponds to the actual case. As can be seen in this figure, the luminance component motion vector (0,1) and the color difference component motion vector (0,3 / 4) are parallel. As a result, the present invention eliminates the color shift between the luminance component and the color difference component, which has been a problem in the prior art.

Here, in the examples of FIGS. 20 and 21, the case of a specific vector has been described, but by applying the present embodiment to the prediction between other different parity fields, the prediction without the difference between the brightness and the color difference Is possible.

When the parity of both the reference destination and the reference source fields are equal, the above-mentioned color shift does not occur. Therefore, the same configuration as the color difference component motion vector generation means 220 from the conventional luminance component motion vector of FIG. 10 is used. The result of the first color difference component motion vector means 233 of the present invention is selected and used as the color difference component motion vector 232. In this case, since the color difference component motion vector obtained by the present invention is equivalent to the result of the prior art, the description in the present embodiment is omitted.

Further, in another embodiment of the present invention, the equations (3), (4) and (5) are different depending on how the units of the luminance component motion vector and the color difference component motion vector are taken. 22 to 24 are diagrams illustrating another embodiment of the first color difference component motion vector generating means, the second color difference component motion vector generating means, and the third color difference component motion vector generating means in the present invention. ..

Here, it is assumed that when the definition of the luminance component motion vector changes by 4, it represents the motion in the luminance image for one pixel, and when the definition of the color difference component motion vector changes by 8, it represents the motion of one pixel. Assuming that the motion in the color difference image is represented, the motion vector candidate 262a of the first color difference is calculated by the color difference component motion vector generation means 260a from the luminance motion vector 261a by the following equation, (MVC1_x, MVC1_y) = (MV_x). , MV_y) Calculated according to (Equation 8). Then, the obtained motion vector candidate 262a having the first color difference is output to the selection means.

The second color difference motion vector candidate 272a is calculated by the color difference component motion vector generation means 270a from the luminance motion vector 271a by the following equation: (MVC2_x, MVC2_y) = (MV_x, MV_y + 2) (Equation 9) ) Is calculated according to. Then, the obtained motion vector candidate 272a of the second color difference is output to the selection means.

The third color difference motion vector candidate 282a is obtained from the luminance motion vector 281a by the color difference component motion vector generation means 280a, using the following equation: (MVC3_x, MVC3_y) = (MV_x, MV_y-2) (Equation 10) ) Is calculated according to. Then, the obtained motion vector candidate 282a having the second color difference is output to the selection means.

Although the present embodiment has been described by taking AVC FCD as an example, the description here is only one embodiment and does not limit other embodiments. (Appendix 1) In a motion image coding method that predicts motion compensation between fields for a motion image signal composed of a plurality of fields, a plurality of color difference components that generate a motion vector of a color difference component from a motion vector of a brightness component. It has a motion vector generation means and a selection means for selecting one of the color difference component motion vector generation means used for generating the color difference component motion vector by inputting the parity of the reference destination field and the reference source field of the motion vector. A moving image coding device characterized in that a prediction vector of a color difference component is generated from a motion vector information of brightness information by a color difference component motion vector generation means selected in. (Appendix 2) In Appendix 1, the color difference component motion vector generation means for generating the color difference component motion vector from the luminance component motion vector is selected by the selection means when the reference destination field and the reference source field have the same parity. The first color difference component motion vector generation means, the reference destination field is the Top field, and the reference source field is Bottom. A second color difference component motion vector generation means selected by the selection means at the time of the field, and a third color difference selected by the selection means when the reference field is the Bottom field and the reference source field is the Top field. A moving image coding device comprising a component motion vector generating means. (Appendix 3) In Appendix 2, the first color difference component motion vector generation means generates a motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 1 of the vector component value. MVy, the value of the vector component is 1 as a unit, and when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image is MVCy, it is calculated as MVCy = MVy ÷ 2. Video encoding device. (Appendix 4) In Appendix 2, the second color difference component motion vector generation means generates a motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 1 of the vector component value. MVy, when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image in units of 1 is MVCy. A moving image encoding device characterized in that it is obtained as MVCy = MVy ÷ 2 + 0.25. (Appendix 5) In Appendix 2, the third color difference component motion vector generation means generates a motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 1 of the vector component value. MVy, when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image in units of 1 is MVCy, it is characterized by being calculated as MVCy = MVy ÷ 2-0.25. Video encoding device. (Appendix 6) In Appendix 2, the first color difference component motion vector generation means generates a motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 4 as the vector component value. MVy, a moving image characterized by being obtained as MVCy = MVy when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image is MVCy with the value of the vector component in units of 8. Encoding device. (Appendix 7) In Appendix 2, the second color difference component motion vector generation means uses MVy as the motion vector of the brightness component, which indicates the vertical motion of one pixel of the brightness component of the field image in units of 4, and the vector component. The moving image coding is characterized by being obtained as MVCy = MVy + 2 when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image in units of 8 is MVCy. apparatus. (Appendix 8) In Appendix 2, the third color difference component motion vector generation means generates a motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 4 as the vector component value. MVy, when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image in units of 8 is MVCy, it is characterized by being obtained as MVCy = MVy 2. Video encoding device. (Appendix 9) In a moving image decoding method that predicts motion compensation between fields for a moving image signal composed of multiple fields. A color difference component used to generate a motion vector by inputting a plurality of color difference component motion vector generation means for generating a motion vector of a color difference component from a motion vector of a brightness component and the parity of a reference field and a reference source field of the motion vector. It has a selection means for selecting one of the motion vector generation means, and is characterized in that a prediction vector of the color difference component is generated from the motion vector information of the brightness information by the color difference component motion vector generation means selected by the selection means. Motion vector decoding device. (Appendix 10) In Appendix 9, the color difference component motion vector generation means for generating the color difference component motion vector from the brightness component motion vector is selected by the selection means when the reference destination field and the reference source field have the same parity. When the first color difference component motion vector generation means and the reference field is the Top field and the reference source field is the Bottom field, the second color difference component motion vector generation means selected by the selection means and the reference field are Bottom field and referrer field is Top A moving image decoding device comprising a third color difference component motion vector generating means selected by a selection means in the field. (Appendix 11) In Appendix 10, the first color difference component motion vector generation means generates a motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 1 of the vector component value. MVy, the value of the vector component is 1 as a unit, and when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image is MVCy, it is calculated as MVCy = MVy ÷ 2. Video decoding device. (Appendix 12) In Appendix 10, the second color difference component motion vector generation means generates a motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 1 of the vector component value. MVy, when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image in units of 1 is MVCy, it is characterized by being calculated as MVCy = MVy ÷ 2 + 0.25. Video decoding device. (Appendix 13) In Appendix 10, the third color difference component motion vector generation means uses MVy as the motion vector of the brightness component, which indicates the vertical motion of one pixel of the brightness component of the field image in units of 1, and the vector component. A moving image characterized in that the value of MVCy = MVy ÷ 2-0.25 when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image in units of 1 is MVCy. Decryptor. (Appendix 14) In Appendix 10, the first color difference component motion vector generation means generates a motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 4 as the vector component value. MVy, a moving image characterized in that it is obtained as MVCy = MVy when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image is MVCy with the value of the vector component in units of 8. Decryptor. (Appendix 15) In Appendix 10, the second color difference component motion vector generation means uses MVy as the motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 4 and the vector component. A moving image decoding device characterized in that MVCy = MVy + 2 is obtained when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image in units of 8 is MVCy. (Appendix 16) In Appendix 10, the third color difference component motion vector generation means generates a motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 4 as the vector component value. MVy, the value of the vector component is 8 as a unit, and when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image is MVCy, it is obtained as MVCy = MVy-2. Video decoding device. (Appendix 17) In a moving image coding / decoding method for predicting motion compensation between fields for a moving image signal composed of a plurality of fields. To generate a color difference component motion vector by inputting the steps of providing a plurality of color difference component motion vector generation means for generating a color difference component motion vector from a brightness component motion vector and the parity of the motion vector reference destination field and the reference source field. It has a selection step to select one of the color difference component motion vector generation means to be used, and the color difference component motion vector generation means selected in the selection step generates a prediction vector of the color difference component from the motion vector information of the brightness information. A program that allows a computer to realize the characteristic video coding / decoding method. (Appendix 18) In Appendix 17, the reference field and the reference source field are selected by the selection step as the means for generating the motion vector of the color difference component from the motion vector of the brightness component when the reference destination field and the reference source field have the same parity. When the first color difference component motion vector generation means and the reference field is the Top field and the reference source field is the Bottom field, the second color difference component motion vector generation means selected by the selection step and the reference field are Bottom field and referrer field is Top A program characterized by having a third color difference component motion vector generating means selected by a selection step in the field. (Appendix 19) In Appendix 18, the first color difference component motion vector generating means generates a motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 1 of the vector component value. MVy, the value of the vector component is 1 as a unit, and when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image is MVCy, it is calculated as MVCy = MVy ÷ 2. program. (Appendix 20) In Appendix 18, the second color difference component motion vector generation means generates a motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 1 of the vector component value. MVy, when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image in units of 1 is MVCy, it is characterized by being calculated as MVCy = MVy ÷ 2 + 0.25. Program to be. (Appendix 21) In Appendix 18, the third color difference component motion vector generation means uses MVy as the motion vector of the brightness component, which indicates the vertical motion of one pixel of the brightness component of the field image in units of 1, and the vector component. A program characterized in that the value of is calculated as MVCy = MVy ÷ 2-0.25 when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image is MVCy. (Appendix 22) In Appendix 18, the first color difference component motion vector generation means generates a motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 4 as the vector component value. MVy, a program characterized in that MVCy = MVy is obtained when the motion vector of the color difference component, which indicates the vertical movement of one pixel of the color difference component of the field image in units of 8 as the vector component value, is MVCy. (Appendix 23) In Appendix 18, the second color difference component motion vector generation means uses MVy as the motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 4 and the vector component. A program characterized in that MVCy = MVy + 2 when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image is MVCy. (Appendix 24) In Appendix 2, the third color difference component motion vector generation means generates a motion vector of the brightness component indicating the vertical motion of one pixel of the brightness component of the field image in units of 4 as the vector component value. MVy, when the motion vector of the color difference component indicating the vertical movement of one pixel of the color difference component of the field image in units of 8 is MVCy, it is characterized by being obtained as MVCy = MVy 2. program. (Appendix 25) In a moving image coding / decoding method for predicting motion compensation between fields for a moving image signal composed of a plurality of fields. To generate a color difference component motion vector by using the steps of providing a plurality of color difference component motion vector generation means for generating a color difference component motion vector from a brightness component motion vector and the parity of the motion vector reference destination field and the reference source field as inputs. It has a selection step to select one of the color difference component motion vector generation means to be used, and the color difference component motion vector generation means selected in the selection step generates a prediction vector of the color difference component from the motion vector information of the brightness information. A featured video coding / decoding method. (Appendix 26) In Appendix 25, as a means for generating a motion vector of a color difference component for generating a motion vector of a color difference component from a motion vector of a brightness component, when the reference destination field and the reference source field have the same parity, they are selected by a selection step. When the first color difference component motion vector generation means and the reference field is the Top field and the reference source field is the Bottom field, the second color difference component motion vector generation means selected by the selection step and the reference field are Bottom field and referrer field is Top A moving image coding / decoding method comprising a third color difference component motion vector generating means selected by a selection step in the field.

<figref num="1">Block diagram of inter-frame predictive signable device</figref><figref num="2">The figure explaining the position of each pixel of luminance and color difference, and the field to which they belong</figref><figref num="3">The figure explaining the spatiotemporal position in the vertical direction of each pixel of brightness and color difference in a field image.</figref><figref num="4">The figure explaining the relationship between a field and a frame in a frame coding mode</figref><figref num="5">The figure explaining the prediction method in the inter-frame prediction coding mode</figref><figref num="6">The figure explaining the prediction method in the inter-field prediction mode</figref><figref num="7">Diagram explaining the coordinates of the field image</figref><figref num="8">The figure explaining the calculation method of the motion vector between the corresponding pixels between the fields of the conventional method.</figref><figref num="9">The figure explaining the calculation method of the pixel pointed by the motion vector of the conventional method.</figref><figref num="10">The figure explaining the method of obtaining the color difference component motion vector from the conventional luminance component motion vector.</figref><figref num="11">The figure explaining the calculation method of the interpolated pixel of a color difference component.</figref><figref num="12">Figure to explain the zero vector between fields with different parity in the prior art Figure to explain the principle of Direct Mode</figref><figref num="13">The figure explaining the problem of the prior art when obtaining the color difference component motion vector from the luminance component motion vector when the reference destination is a Bottom field and the reference source is a Top field.</figref><figref num="14">A diagram illustrating problems in the prior art when obtaining a color difference component motion vector from a luminance component motion vector when the reference destination is the Top field and the reference source is the Bottom field.</figref><figref num="15">The figure explaining the method of generating the color difference component motion vector from the luminance component motion vector in this invention.</figref><figref num="16">The figure explaining the embodiment of the 1st color difference component motion vector generation means in this invention.</figref><figref num="17">The figure explaining the embodiment of the 2nd color difference component motion vector generation means in this invention.</figref><figref num="18">The figure explaining the embodiment of the 3rd color difference component motion vector generation means in this invention.</figref><figref num="19">The figure explaining the embodiment of the selection means in this invention.</figref><figref num="20">An example of deriving the color difference component motion vector from the luminance component motion vector when the reference destination is the Bottom field and the reference source is the Top field according to the present invention will be described.</figref><figref num="21">An example of deriving the color difference component motion vector from the luminance component motion vector when the reference destination is the Top field and the reference source is the Bottom field according to the present invention will be described.</figref><figref num="22">It is a figure explaining another embodiment of the 1st color difference component motion vector generation means in this invention.</figref><figref num="23">It is a figure explaining another embodiment of the 2nd color difference component motion vector generation means in this invention.</figref><figref num="24">It is a figure explaining another embodiment of the 3rd color difference component motion vector generation means in this invention.</figref>

Code description

31 Orthogonal conversion means 32 Quantization means 33 Inverse quantization means 34 Inverse orthogonal conversion means 35 Decoded image generation means 36 Decoded image storage means 37 Motion vector calculation means 38 Prediction image generation means 39 Prediction error signal generation means 40 Coefficient entropy coding means 41 Motion Vector Entropy Coding Means 42 Multiplying Means 50a-50d Top Field Brightness No. 1,3,5,7 Line 51a-51d Bottom Field Brightness No. 2,4,6,8 Line 52a-52b Top Field Color Difference No. 1, 3 lines 53a-53b Bottom Field Color difference 2nd and 4th lines 64a-64c Top Field 65a-65c Bottom Field 81 Frame # 1 Luminance component 82 Frame # 2 Luminance component 84a-84c Frame # 1 ~ # 390 Same parity field prediction 91 Different parity field prediction 94a-94b Top Field95a-95b Bottom Field130 Top Field131 Bottom Field132 Top Field133a-133b Coding Target Luminance Component 134a-134b Coding Target Luminance Component 135a-135b Reference Field Luminance Component 136a-136b Preferred Color Difference Component for Prediction 137a-137b Reference Field Luminance Component 140 Reference Source Top Field Luminance Component Pixel 141 Used as Predicted Value Reference Bottom Bottom Field Luminance component pixel position 142 Reference source Top Pixel of field color difference component 143 Reference destination used as predicted value Bottom Field color difference component pixel position 145 Preferred color difference component predicted pixel position 150 Reference source Bottom field Brightness component pixel 151 Referenced destination Top field Brightness component used as predicted value Pixel position 152 Reference source Bottom field Pixel of color difference component 153 Reference destination used as predicted value Top field Pixel position of color difference formation 155 Predicted pixel position of preferred color difference component 160 Reference source Top field Pixel of brightness component 161 Reference used as predicted value Pixel position of the destination Bottom field brightness component 162 Reference source Top Pixel position of the field color difference component 163 Reference destination Bottom field Pixel position of the color difference component used as a predicted value 170 Reference source Bottom field Brightness component pixel 171 Reference destination used as predicted value Top field Brightness component pixel position 172 Reference source Bottom field Color difference component pixel Reference destination used as predicted value Top field Color difference pixel position 180 Coordinates Position to be calculated 181 Pixel definition position 200 Motion vector calculation means 201 Reference source field coordinates 202 Reference field coordinates 203 Motion vector 210 Pixel mapping means 211 Motion vector 212 Reference field coordinates 213 Reference field coordinates 220 Color difference component Motion vector generation Means 221 Brightness component motion vector 222 Color difference component motion vector 230 Selection means 231 Brightness component motion vector 232 Color difference component motion vector 233 First color difference component motion vector generation means 234 Second color difference component motion vector generation means 235 Third color difference component motion vector generation means 237 Motion vector reference source field parity 238 Motion vector reference field parity 240 Selection means 241 Condition judgment table 242 Color difference component motion vector 243 Selector 244 First color difference component Motion vector candidate 245 Second color difference component Motion vector candidate 246 Third color difference component Motion vector candidates 247 Motion vector reference source field parity 248 Motion vector reference field parity 249 Selection information 250 to 255 Integer pixel 256 Interpolated pixel 260,260a First color difference component motion vector generating means 261,261a Brightness component motion vector 262,262a First color difference component motion vector candidate 270,270a Second color difference component motion vector generating means 271,271a Brightness component motion vector 272,272 a Second color difference component motion vector candidate 280,280a Third color difference component motion vector generation means 281,281a Brightness component motion vector 282,282a Third color difference component motion vector candidate