JP3152765B2 - Image coding device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image coding apparatus for a still image or a moving image. About the method.

[0002]

2. Description of the Related Art Multi-tone image signals such as TVs and photographs have an enormous amount of information. Therefore, when digitized and recorded or transmitted on a recording medium such as a tape or a disk, the recording capacity and transmission capacity of the medium are increased. It is necessary to reduce the data rate due to the problem of the bandwidth of the road. For this purpose, various studies have been made on high-efficiency coding for reducing the code amount by using the redundancy of image data.

[0003] An orthogonal transform encoding method is often used as such an image encoding method. In this method, input digital image data is divided into blocks of m × n pixels, and orthogonal transform such as two-dimensional discrete cosine transform (DCT) is performed for each block, and each coefficient (transform coefficient) obtained by this transform is performed. ) Is performed with a predetermined quantization width. afterwards,
Sort each quantized transform coefficient into a one-dimensional data sequence,
The code amount is reduced by using run-length coding and variable-length coding.

[0004] In addition to the above basic configuration, the quantization width is adaptively switched for each block, and the quantization width is reduced by reducing the quantization width for blocks in which the quantization distortion during coding is conspicuous. By suppressing the occurrence, visual image quality characteristics may be improved. At this time, in order to suppress an increase in the code amount, a quantization width is set to be coarse for a block in which quantization distortion is not conspicuous.

In order to adaptively set the quantization width for each block as described above, it is necessary to classify the blocks into blocks in which image quality deterioration is easily noticeable and blocks in which the image quality deterioration is not so noticeable. In the related art, attention has been paid to only a block to be quantized, and a quantization mode is set using a parameter calculated from a pixel value or a transform coefficient value in the block. For example, in the 1990 Television Society Annual Meeting 7-1, “Study of High Efficiency Coding for Digital VTR,” the block to be quantized is further divided into sub-blocks, and each sub-block is It is performed to detect whether or not there is a drastic change, and determine a quantization characteristic according to the result. This is a block in which the image quality deterioration is conspicuous is a block in which a flat portion and a rapidly changing portion are mixed in one block.
This is based on the idea that in a block having almost no flat part, the image quality degradation is not conspicuous even when coarsely quantized.

According to this method, when an edge exists in a block, ringing caused by coarsely quantizing the orthogonal transform coefficient obtained in the block has a significant adverse effect on a flat portion in the block. This is a relatively effective way to reduce

On the other hand, devices such as a TV phone, a TV conference, an optical disk device, a VTR, and a CATV require a technology for efficiently encoding a moving image signal. As such a moving image encoding method, a pixel value of an input image to be encoded (encoding target image) is predicted using a pixel value of an encoded reference image specified by a motion vector, and So-called motion-compensated prediction coding, which codes a prediction error and a motion vector, is known.

[0008] In this motion compensation predictive coding system,
A small area (block) consisting of multiple pixels for the image to be encoded
In order to perform motion compensation by dividing the reference image, the shift of the encoding target image with respect to the encoded reference image, in other words, the motion which is information indicating from which block of the reference image the block of the encoding target image has moved. A vector is obtained for each block. When obtaining a motion vector, a block is cut out from the reference image by changing the position little by little within the search range of the predetermined motion vector in order, and the block of the cut-out reference image and the corresponding block of the target block for motion vector detection are extracted. Find the sum of the absolute values of the differences between the pixel values at the position,
The displacement of the position at which the sum of absolute values is minimum relative to the target block is defined as the motion vector of the target block. In addition to obtaining the motion vector using the pixel size as a unit, the coding efficiency may be further increased by creating a reference block by interpolating from the surrounding pixel values and obtaining the motion accuracy smaller than the pixel size. is there.

In the moving picture coding method, for example, in the case of special reproduction in a VTR, not only an image one frame before, that is, a temporally immediately preceding image, but also an image several frames earlier or several frames later as a reference image. May be used as a reference image. In this case, since the encoding target image and the reference image are temporally separated from each other, the motion is generally larger than in the case of referring to the adjacent image, and the search range for the motion vector needs to be widened. A method of expanding the search range according to the temporal distance between the image to be encoded and the reference image and calculating the sum of absolute values for all the candidates in the region to obtain a motion vector is called a full search.

Further, when obtaining a motion vector from a reference image which is temporally distant, a motion vector is obtained within a certain search range for an image temporally adjacent to the image to be encoded, and the motion vector is searched for. As a center, a method called a telescopic search for obtaining a motion vector from a reference image may be used by sequentially obtaining a motion vector for an adjacent image.

Further, if a search for a motion vector is performed for the entire encoded image, the amount of calculation becomes enormous.
When calculating a motion vector from temporally adjacent images,
The search range of the motion vector is limited to the vicinity of a block located at the same position on the screen as the encoding target block of the input image data of the encoded image, and the center of the block is set as the search center, so that the search can be performed efficiently. There are also ways to do it.

Further, the input image and the coded image are formed into a pyramid structure as shown in FIG. 19, a block which is considered to be optimal as a prediction signal in an upper layer having a small number of pixels is detected, and the position of the detected block is determined by the number of pixels. A method is also known in which a search pixel unit is hierarchically refined using the center of a search region in a lower layer, which is often used. However, even in this method, the search range in the upper layer is limited to some extent as shown in FIG. 20, so that a block having a large motion and obtained in the upper layer is the peripheral portion A of the search range, Even when detection close to a block has not been performed, a block centered on A, whose correlation is not so high, is searched for in fine pixel units. Therefore, the arithmetic capability of the device cannot be used effectively, and an optimum block centered on B is not detected in the lower layer.

For this reason, when the motion is large, the block having the highest correlation with the encoding target block of the input image exists in the encoded image, but the block cannot be detected. , The difference from the block having a lower correlation or the pixel value itself of the partial area of the input image is encoded. As a result, when the code amount becomes very large or when the code amount is limited to a predetermined value or less, it is necessary to increase the quantization width, and the reproduction image quality is significantly deteriorated.

[0014]

In the conventional motion-compensated predictive coding method, a large area in an input image having a large amount of motion is searched for a narrow area in which a motion vector search is not effective. In addition, since the motion vector is searched with high search accuracy, there is a problem that the processing capability of the apparatus cannot be fully utilized and an optimal prediction signal cannot be obtained.

Accordingly, it is an object of the present invention to provide an image coding apparatus capable of performing effective motion vector detection so that an optimal prediction signal can be obtained.

[0016]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention provides a signal processing means for performing a process for lowering the resolution of image data, and an input image data which has been processed by the signal processing means. First motion vector detection for detecting a primary motion vector over a first search range with reference to a reference image block in encoded image data processed by the signal processing means for a current block to be encoded Means, for the encoding target block in the input image data not processed by the signal processing means,
A second motion vector for obtaining a secondary motion vector with a predetermined search accuracy over a second search range with reference to a reference image block in encoded image data not processed by the signal processing means.
And the primary motion vector detected by the first motion vector detecting means is any one of a central portion and an outer peripheral portion within the first search range.
Depending on whether detected from minute, our size and the center position of the second search range in the second motion vector detection means
And control means for controlling at least one of the search accuracy ,
Encoding means for encoding a difference between the reference image block specified by the secondary motion vector detected by the second motion vector detection means and the encoding target block.

Further, the signal processing means reduces the resolution by performing processing including sub-sampling processing on the image data.

According to one aspect, when the primary motion vector is obtained from a central portion of the first search range, the control means can further refine the search accuracy of the second motion vector detection means. And the size of the second search range
When the primary motion vector is obtained from the outer peripheral portion of the first search range, the search accuracy of the second motion vector detecting means is further coarsened to reduce the second motion vector.
Is characterized by making the size of the search range wider.

According to another aspect, when the primary motion vector is obtained from an outer peripheral portion of the first search range, the control means sets the center position of the second search range to the first search range. It is characterized by being further away from the center position of the search range.

According to still another aspect, when the primary motion vector is obtained from a central portion of the first search range, the control means sets the center position of the second search range to the first search range. And the size of the second search range is further narrowed by making the search accuracy of the second motion vector detection means finer, and the primary motion vector is set to the position indicated by the first search vector. When the position is obtained from the outer peripheral portion of the range, the center position of the second search range is made farther away from the center of the first search range, and the search accuracy of the second motion vector detecting means is made coarser. Then, the size of the second search range is made wider.

[0021]

[0022]

[0023]

[0024]

[0025]

[0026]

[0027]

[0028]

[0029]

[0030]

[0031]

[0032]

[0033]

[0034]

[0035]

As described above, according to the present invention, the second motion vector detecting means determines whether the primary motion vector detected by the first motion vector detecting means has been detected from which region in the motion vector search range. At least one of the search range and search accuracy of the next motion vector is controlled.

[0036]

[0037]

[0038]

[0039]

[0040]

[0041]

More specifically, when the primary motion vector obtained by the first motion vector detecting means is obtained from the central part of the search area, the search accuracy of the second motion vector detecting means (the search pixel (Unit), and a secondary motion vector optimal for a small motion block is searched from a narrow search range. If a primary motion vector is obtained from an outer peripheral portion of the search range, a second motion vector is detected. The search accuracy of the means is coarsened, and a secondary motion vector optimal for a block with large motion is searched from a wide search range.

When the primary motion vector obtained by the first motion vector detecting means is obtained from the outer peripheral portion of the search range, the search range of the second motion vector detecting means is changed to the first motion vector. The center position of the search range of the second motion vector detecting means is controlled to be further away from the center of the search range of the first motion vector detecting means so that the portion overlapping the search range of the vector detecting means is reduced. Then, a secondary motion vector optimal for a block having a large motion is obtained.

Further, when the primary motion vector obtained by the first motion vector detecting means is obtained from the center of the search range, the center position of the search range of the second motion vector detecting means is set to the first position. If the search accuracy is refined as the position corresponding to the motion vector, the optimal secondary motion vector for the small motion block is searched from the narrow search range, and if the primary motion vector is obtained from the outer peripheral portion of the search range, The center position of the search range in the second motion vector detecting means is set so as to reduce the portion where the search range in the second motion vector detecting means overlaps the search range in the first motion vector. The second motion vector detecting means is controlled so as to be further away from the center of the range, the search accuracy of the second motion vector detecting means is reduced, and the optimal secondary motion vector Search for.

As described above, according to the present invention, by controlling at least one of the search range and the search accuracy of the secondary motion vector according to the region within the search range from which the primary motion vector has been detected, the optimum Since a suitable motion vector can be detected as a secondary motion vector and an optimal prediction signal can always be obtained regardless of the magnitude of the motion of the image, the coding efficiency is improved.

In this case, the first motion vector detecting means should pay attention to the area of the search range from which the primary motion vector has been detected, and focus on the image data whose resolution has been reduced. Therefore, the amount of processing and the amount of hardware required for detecting the primary motion vector can be reduced.

[0047]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows an example of the configuration of an image encoding apparatus in which the present invention is applied to image compression by discrete orthogonal transformation as a first embodiment.

In FIG. 1, digital image data as an input signal is, for example, 8 × 8
After being cut out into blocks of the size of a pixel, they are input to the two-dimensional discrete orthogonal transform unit 102, where two-dimensional discrete orthogonal transform, for example, two-dimensional DCT, is performed for each block. The output of the two-dimensional discrete orthogonal transform unit 102 is data of a transform coefficient, and is quantized by the quantizer 103 using a quantization width calculated by a quantization width setting method described later, thereby reducing the amount of information. . The output of the quantizer 103 is input to a variable-length coding unit 104 and subjected to variable-length coding such as Huffman coding, so that the code amount is reduced. The encoded data output from the variable length encoding unit 104 is stored in the buffer 105.

The multipliers 131 and 132 determine the quantization width of the quantizer 103 by using the basic quantization table 107.
The reference value of the quantization scale calculated by the quantization width control parameter calculation unit 109 for the value that determines the ratio of the quantization width to each transform coefficient contained in A value multiplied by the amount of change from the reference value of the quantization scale calculated by the provided adaptive quantization control unit 108 is set.

The quantization width control parameter calculation unit 109 monitors the occupation state of the code generated by the variable length coding unit 104 in the buffer 105 so that the code amount output from the buffer 105 falls within the set code rate. ,
Control the quantization scale accordingly. That is, if the generated code amount is larger than a certain set value, the quantization scale is increased by the quantization width control parameter calculation unit 109, thereby coarsening the quantization width and suppressing the generated code amount. Conversely, when the generated code amount is smaller than the set value, the quantization scale is reduced to reduce the quantization width to improve the image quality, and increase the generated code amount up to the set code rate.

As described above, the quantization width control parameter calculation unit 1
According to the result of the calculation in step 09, the quantization width is changed in accordance with the property of the input image data, thereby controlling the code rate to be appropriate. However, in such quantization width control, since the quantization width is changed without considering the visual sense of image quality deterioration, there is waste in code amount distribution in the image. In other words, for image portions where deterioration is not so noticeable, the quantization width is further reduced to reduce the code amount, and for image portions where deterioration is conspicuous, the quantization width is reduced. Therefore, it is desired to improve subjective image quality. The adaptive quantization control unit 108 performs this control.

Hereinafter, the control method by the adaptive quantization control unit 108 will be described with reference to FIG. Consider that the quantization width of the central block is determined by the nine blocks shown in FIG. First, for each block, the block activity calculation unit 106 calculates a parameter indicating whether the block is a flat image block with little change in image or a rapidly changing image block (this is called block activity) in the block activity calculation unit 106. Calculate each.
In other words, the block activity is a level of difficulty of encoding, and the larger the value, the more difficult it is to encode. As the block activity, in this embodiment, (1) the average value of the pixel values in the block and the absolute value sum of the difference between the pixel values are used, but (2) the average value of the pixel values in the block is used. The sum of squares of the difference between the value and the value of each pixel, (3) the sum of the absolute values of the difference values between adjacent pixels, or (4)
It is also possible to use a value such as the sum of absolute values of coefficients excluding the DC component of the DCT transform coefficient.

In this embodiment, adaptive quantization for changing the quantization size in four stages is performed. Therefore, the adaptive quantization control unit 108 sets the change width of the quantization by the following steps.

First, as a first step, the classifying unit 114 classifies each block into four classes according to the size of the block activity. The four classes are A,
B, C, and D. Then, the class information of each block is stored in the memory 111.

Next, as a second step, the central block for which the quantization width is to be determined is shown in FIG.
(B) When classified into the D class as in (c), look at the classes of the surrounding eight blocks and find out if the number of blocks classified into the D class is less than a certain specified value (for example, three). Find out if. If the number of blocks in the D class of the eight peripheral blocks exceeds the specified value as shown in FIG. 2B, the number of blocks is kept as it is, and if the number of blocks is smaller than the specified value as shown in FIG. Is performed for all blocks in the image.

Thereafter, as a third step, a change in the quantization width for each class is determined as shown in FIG. FIG.
Indicates the amount of change from the reference value of the quantization width corresponding to each of the classes A to D of the block activity.
In this example, the quantization width corresponding to the C class is made to match the reference quantization width obtained by the quantization width control parameter calculation unit 109.

That is, an image area where blocks with high activity are concentrated is determined to be a part where visual image quality deterioration is relatively inconspicuous, and quantization is made coarse, and conversely, an image area where activity is relatively small. Since the deterioration is conspicuous, the image quality is improved by reducing the quantization width. In addition, when a high activity block is in a relatively flat image area in a state where the block is high, the deterioration of the block becomes more conspicuous. Is to be reduced.

The blocks classified into the A class from the beginning in the first step correspond to flat areas of the image. In such a region, if the quantization width is made coarse, deterioration due to block distortion occurs, so the quantization size is made small.However, since the amount of information in this block is originally small, the code amount by making the quantization width small is used. The increase is not so great. Therefore, in the first step, it is possible to increase the amount of code distribution to the blocks in which the distortion is categorized into the D class, in which the distortion is conspicuous, and the image quality improving effect is enhanced.

In FIG. 3, the amount of change from the reference value of the quantization width of blocks belonging to each of the classes A to D is determined according to the number of blocks classified into each class in one screen.
That is, the code amount generated when a certain quantization width is set is predicted, and the quantization widths of the blocks belonging to the A and B classes are set to such an extent that the quantization width ΔQ of the blocks belonging to the D class is visually permissible. When the value is made coarser than the reference value, the quantization width is set to be lower than the quantization width so as to increase the code amount of one screen, which is smaller than the case where the reference value is used.

At this time, the visually permissible quantization width of the block belonging to class D changes according to the activity of the entire screen, and the higher the activity of the entire screen, the smaller the amount of change of the permissible quantization width. . FIG. 4 shows this state, in which the horizontal axis represents the activity of the entire screen (all activities), and the vertical axis represents the quantization width ΔQ of the block belonging to the D class.

The quantization parameters obtained in the quantization width control parameter calculation section 109 and the adaptive quantization control section 108 are multiplexed with the coded image data by the multiplexing section 120 and output.

In the above-described first embodiment, good results can be obtained by setting the quantization width up to the third step in the adaptive quantization control unit 108, but it is possible to classify into more classes. In order to reduce the unnaturalness of an image caused by a discontinuous change in image quality caused by a sudden change in the quantization width in a screen, further improvements can be made.

FIG. 6 shows the improved second embodiment. In addition to the configuration of the first embodiment shown in FIG. 1, a two-dimensional low-pass filter 113 is newly provided in the adaptive quantization control unit 108. Has been added to In this embodiment, a two-dimensional array consisting of the variation width of quantization of each block obtained in the third step in the first embodiment is considered.
By operating the dimensional low-pass filter 113,
A sharp change in the quantization width in each region of the image is avoided.

The present invention can be applied to an image coding apparatus using inter-frame prediction. In the inter-frame prediction, as shown in FIG. 6, since images of a plurality of frames adjacent in the time direction are used, when performing adaptive quantization, information from blocks adjacent in the time direction can also be used. Thus, it is possible to prevent the image quality from suddenly changing over time.

FIG. 7 shows a third embodiment in which the present invention is applied to image coding using inter-frame prediction. In this embodiment, the output of the quantizer 103 is
5, a difference between a signal obtained by motion-compensating an image signal of one frame before locally decoded via a two-dimensional discrete inverse orthogonal transformation unit 116, an adder 133 and a frame memory 117 by a motion compensation unit 118 and an input image signal; The orthogonal transform such as DCT is performed by the two-dimensional discrete orthogonal transform unit 102 on the prediction residual signal, which is the difference, by the subtractor 134, and the obtained transform coefficient is quantized by the quantizer 103 to obtain a variable-length code. The variable length coding is performed by the transformer 104.

The setting of the quantization width in this embodiment is performed in the same manner as in the first embodiment.
9, the coefficient obtained according to the code amount of the variable length code stored in the buffer 105 is corrected by adaptive quantization according to the present invention, and the value is changed to a value corresponding to each quantization scale of the basic quantization table 107. Set by multiplying by.

In this embodiment, the adaptive quantization controller 1
In step 08, the class information classified by the classifying unit 114 for the frame to be currently encoded is stored in the first memory 111, and the finally obtained class information of the previous frame is stored in the second memory 121. It is remembered. The correction value of the quantization width is obtained using each value and the output of the scene change detection unit 119.

Here, when determining the change in the quantization width in the center block in FIG. 6, the block which is spatially adjacent to this block and which is spatially at the same position as the preceding eight blocks Information is used.

First, the operations of the first step and the second step in the first embodiment are performed in the same manner using the information of the blocks adjacent in the space, and the classification into the class obtained by the classifying unit 114 is corrected. Perform

Next, if they differ from the class of the block of the previous frame by more than a preset value, for example, if a difference of more than two classes occurs,
The class is corrected so as to reduce the difference by one step. For example, as a result of the second step, if the block is classified into the D class, and if the block at the same position in the previous frame is the A class or the B class, the class of the block is further corrected to the C class. Do what you want. The class information finally obtained in this manner is written into the memory 121, and the class information is prepared for determining the correction value of the quantization width of the next frame. However, if the scene change detection unit 119 determines that a scene change has occurred, the correction based on information from temporally adjacent blocks is not performed.

Finally, the change value of the quantization width shown in FIG. 3 is determined according to the class of each block finally determined.

In each of the embodiments described above, each block is first classified into four classes based on the block activity, and the classified classes are classified according to the classes of the surrounding blocks in order to facilitate the calculation. Although the correction method is used, a method of directly determining the quantization width of the center block by directly using the activity of the surrounding blocks is also possible.

In each of the above embodiments, the adaptive quantization control unit 108 and the quantization width control parameter calculation unit 109 are separately described for setting the quantization width for the sake of explanation.
In practice, it may be advantageous to perform the adaptive quantization control by integrating the functions of the adaptive quantization control unit 108 and the quantization width control parameter calculation unit 109.

Next, a description will be given of an embodiment in which the present invention is applied to a moving picture coding apparatus using a motion compensation predictive coding method.

First, the principle of the present embodiment will be described.
The search for the motion vector is performed independently for each of the divided blocks, and despite the fact that the direction and magnitude of the motion are different for each region in the image, in the conventional technology, a uniform motion vector is used for each block. The search range was set.

On the other hand, in the present embodiment, an appropriate motion vector search range is set for each block, and the search range is set small for a block expected to have a small motion. Find the optimal motion vector. Conversely, by establishing a sufficient search range for a block that is expected to have a large motion, an optimal motion vector that does not cause a decrease in coding efficiency is obtained.

Further, in an image area having a pattern with high resolution and little deformation due to motion, a motion vector is obtained with higher accuracy to generate a reference image block, thereby greatly reducing a prediction error to be encoded. be able to. Conversely, for a figure that is greatly deformed due to motion or a flat image area having a very low resolution, obtaining a motion vector with high accuracy does not improve coding efficiency. Therefore, by setting the motion vector search area for each block and appropriately controlling the accuracy in obtaining the motion vector, it is possible to reduce the waste of calculation and shorten the encoding time.

Further, in the case of a moving image, the pattern of each image that is temporally continuous is determined when there is no scene change.
It can be considered that they are moving continuously in space. Therefore, when attention is paid to a specific pattern in the image, it is considered that the temporal change of the position is somewhat continuous.
In other words, when obtaining the motion vector of a block, knowing the past motion (or the motion when looking back from the future) that occurred in the block allows the block to determine what motion after (or before) It is possible to roughly predict what to do.

By utilizing this fact, in the present embodiment, when a motion vector from a reference image of a certain encoding target block is obtained, a reference image block located at a position corresponding to the encoding target block is encoded. Is set as the search center at the position indicated by the optimum motion vector used for the search. FIG. 10 shows this. When obtaining a motion vector for the block 405 in the encoding target image 403, the same vector 40 as the motion vector 406 obtained for the block at the same position of the reference image 402 which has already been encoded.
7, the position 408 defined by the vector is set as the search center.

In this method, the motion vector is not tracked following the motion of a certain object, but the motion of the object in a block at a predetermined position is predicted.
That is, in FIG. 10, for example, the relationship that the object existing at the position of 410 on the screen 403 becomes the expected position 411 on the screen 403 is not directly used. The motion in block 405 is predicted based on the prediction that there will be. Therefore, at the border of the two different moving areas,
For example, in the boundary region between the background and the foreground, if the size of the normal search range is set by such a method of setting the search center, an accurate motion vector may not be obtained.

Therefore, in the present embodiment, the motion vector of each block obtained in the image immediately before the input image to be coded is examined, and an area having a motion vector that is significantly different from the motion vectors of the surrounding blocks is examined. The center of the search range as given by the vector 407 in FIG. 10 is not moved for the block located in, and an area wider than the search range for the normal block is set as the search range.

Further, in the present embodiment, a more effective search for a motion vector is performed by focusing on the behavior of a motion vector candidate and an index indicating the magnitude of a prediction error for the motion vector candidate. The index indicating the magnitude of the prediction error is, for example, the sum of the absolute value of the difference value of the reference block candidate designated by the motion vector candidate and the difference value of the pixel at the corresponding position in the block of the encoding target block or the difference value. The sum of squares can be obtained, and it can be obtained by judging that the motion vector that minimizes the sum is the optimal motion vector.

Here, when the behavior of the motion vector candidate and the index of the magnitude of the prediction error for the motion vector candidate are examined, a typical behavior as shown in FIG. 11 is often obtained. In FIG. 11, the horizontal axis is a list of motion vector candidates located on a straight line in a certain direction in the order of spatial positional relationship, and the vertical axis is an index indicating the magnitude of the prediction error. FIG. 11A is a case of an image in which a relatively complicated pattern moves in parallel without much deformation due to movement, and the prediction error can be reduced by estimating an appropriate motion vector. FIG. 11B shows a case of a flat image, which corresponds to a region in which the coding efficiency is less reduced even if less accurate motion compensation is performed. FIG. 11C is an image in which the deformation due to motion is large and the coding efficiency is not significantly improved by motion compensation. FIG. 11D shows the same behavior as FIG. 11A, but in a range indicated by the horizontal axis in the figure, an appropriate motion vector has not yet been obtained.

The order of generating motion vector candidates is as follows.
As shown in FIG. 12, starting from the center of the motion vector search range, the motion vector candidates are set so as to be located sequentially from the inner periphery to the outer periphery, and the set point interval of the motion vector candidates becomes wider as moving to the outer periphery. To The index values are calculated for the motion vector candidates set in this way, and when the calculation is completed for a predetermined number of candidates, the vertical, horizontal, and diagonal straight lines shown in FIG. It is checked for each direction whether or not the index value for the motion vector candidate located above has the tendency as shown in FIG.

As a result, if there is a tendency as shown in FIG. 11A in a certain direction, it is determined that an optimum motion vector is included in the range already searched for that direction, and a motion existing in a wider area is determined. Stop calculating the index value of the vector candidate and calculate the index value by further increasing the search accuracy in the vicinity of the motion vector candidate at the position that takes the lowest value among the index values that have already been calculated. continue.

If there is such a tendency as shown in FIG. 11B or FIG. 11C, even if estimation of the motion vector is performed with high accuracy, the coding efficiency cannot be expected to be much improved. Stop the calculation.

When the tendency shown in FIG. 11D is shown, the calculation in the direction in which the index value increases is stopped. By performing such processing, it is possible to reduce the useless calculation of the index value, and it is possible to reduce the time of the entire encoding processing.

After the calculation is completed for all the motion vector candidates within the preset search range, the same tendency is checked. If the tendency as shown in FIG. 11D can still be detected, FIG. As shown, similar processing is performed by further expanding the search range in the direction in which the index value decreases. As a result, it is possible to avoid a decrease in coding efficiency due to an insufficient search range.

The tendency obtained from a set of such a position of a motion vector candidate and an index value corresponding to the position is shown in FIGS.
And is stored for one screen for each block, and is used to set the initial value of the search center of the motion vector search range and the search accuracy when obtaining the motion vector of the image to be subsequently processed. .

The amount of calculation required for initial setting and resetting of parameters such as the search range, the search center, and the search accuracy when searching for a motion vector as described above is smaller than that for calculating an index value for a motion vector candidate. Since the calculation can be performed with a much smaller calculation amount, the processing time can be greatly reduced as a result.

Next, the configuration of the moving picture coding apparatus according to the present embodiment will be described with reference to FIG.

In FIG. 8, the input image data 313 is divided into a plurality of blocks composed of a plurality of pixels by the block cutout unit 201, and cut out into 8 × 8 pixel blocks to be encoded. The extracted image signal of the encoding target block is input to the encoding unit 202 together with an image signal 301 of a reference image block motion-compensated by a motion compensation unit in the motion vector search unit 204, which will be described later. Encoding is performed on a prediction error signal that is a difference for each pixel of the reference image block.

The encoding unit 202 performs a two-dimensional discrete cosine transform on the prediction error signal as a typical process of the encoding process, and performs a quantization process on each transform coefficient.
For the quantization result, a high-efficiency encoding is performed on a set of a zero run and a value of a transform coefficient following the zero run using a Huffman code or the like. The encoded output 315 from the encoding unit 202 is
The multiplexed signal is multiplexed by the multiplexing unit 207 together with the motion vector 302 used to generate the reference image block, and is output as a final encoded output 314.

Further, the encoded output 315 is decoded by a local decoding unit 206 which performs the reverse operation of the encoding unit 202 to become encoded image data, and is used as a reference image for encoding the next input image data. Image buffer 20 as data
5 is stored.

The motion vector search unit 204 determines the order determined for a plurality of motion vector candidates according to parameters such as the search center, search range, and search fineness of the motion vector initially set by the search parameter setting unit 203. Is used to calculate the index value. In the process of calculating index values for the plurality of motion vector candidates, data such as a set of the position of the motion vector candidate and the index value is returned to the search parameter setting unit 203, and the search range and the fineness of the search are returned. Set again.

The motion vector search section 204 finally obtains a motion vector 302 having the minimum index value from a plurality of motion vector candidates, and the motion vector 302 is indicated by the motion vector 302 from the current block together with the motion vector 302. Reference image block 301 shifted by the position
From the image buffer 205 and output.

FIG. 9 is a block diagram showing a detailed configuration example of the motion vector search section 204 and search parameter setting section 203. First, the motion vector search unit 204 will be described.

In order to obtain a reference image block optimally motion-compensated for the encoding target block 311 in the input image data from the block extracting unit 201 in FIG.
In accordance with the motion vector candidates 310 sequentially output by the vector candidate generation unit 210, the reference candidate block extraction unit 209 outputs the reference candidate block 315 from the reference image data 312 stored in the image buffer 205.

The reference candidate block 315 and the encoding target block 311 are input to the index calculation unit 216, where the absolute value of the difference value of each pixel at the corresponding position between the two blocks 315 and 311 is calculated. The sum of all the pixels in the value block is input to the optimal vector detecting unit 208 as an index value 306 for the given motion vector candidate 310.

Since the motion vector search section 204 initially performs a coarse search, the optimum motion vector is obtained by the optimum vector detection section 208, and then a fine search is performed on a small area around the optimum motion vector. Find the final motion vector.

In the optimum vector detecting section 208, the motion vector candidates 310 sequentially inputted and the index value 3
06, when an index value smaller than the already input index value is input, it is held as a candidate for an optimal motion vector, and when the calculation of index values for all motion vector candidates is completed, The reference image block 301 corresponding to the vector 302 is output.

In FIG. 9, the motion vector search section 204
The reference image block 301 is output, and the difference (prediction error signal) from the current block is generated again in the coding unit 202 in FIG. 8. The difference generated in the index calculation unit 216 is used as the optimum vector detection unit. 208 to the encoding unit 2
02 may be output.

Next, the search parameter setting section 203 will be described.

As a first step of obtaining a motion vector for a certain block to be coded, the search parameter initial setting unit 211 determines a search center, a search range, and an initial value of search accuracy of a motion vector as search parameters. When determining these initial values, a motion vector for each block in the immediately preceding encoded image and a determination mode described later are used. The information on the motion vector and the determination mode is stored in the vector memory 212 and the determination mode memory 215, respectively.

As the search center of the motion vector, the value specified by the vector for the block at the same position as the block to be encoded in the immediately preceding encoded image stored in the vector memory 212 is normally used as it is. However, if the sum of the magnitudes of the difference vectors between the vector of the calculation target block stored in the vector memory 212 and the vectors for the eight surrounding blocks is larger than a preset value, the region performs a different motion. It is determined that it is a boundary between two objects that are moving or a part where the movement is rapid, and the shift of the search center by the vector is not performed exceptionally.

Search range 5 of motion vector in FIG.
The initial setting range of 02 is normally set to a predetermined range, but is set wider than usual when an exception described in the above-described setting of the search center is made. Also, when the determination mode 305 shows a tendency as shown in FIG. 11C, it is determined that the region is a region where the movement is sharp, and the search range is set wider than usual.

The initial setting of the fineness of the motion vector search is as follows.
As shown in FIG. 12, the vicinity of the search center 501 is set finely, and the setting is made such that the candidate points of the motion vector gradually become coarser as the distance from the center increases. However, when the determination mode is as shown in FIG. 11B or FIG.
The initial setting is made so that it is slightly coarser than in the case of the mode.

Next, resetting of the motion vector search range will be described.

In this embodiment, motion vector candidates are generated from candidate points close to the search center in accordance with the above-described initial setting of the search parameters, and index values for the motion vector candidates are calculated. The index value of the candidate point is calculated. Before resetting the motion vector search range, all index values for the candidate points in the area defined by the initial setting in the search range are calculated before the index values for a certain number of candidates, that is, within a certain range from the search center. When the index value for the candidate existing in is calculated, the initial setting of the search parameter is reviewed, and if necessary, the search parameter is reset.

With respect to the set of candidate points and their evaluation values obtained at this stage, those on four straight lines of a vertical direction 602, a horizontal direction 604, and oblique directions 603 and 605 shown in FIG. 13 are extracted. Then, the positional relationship between the candidate points is stored and arranged on the horizontal axis, and the tendency of the graph shape when the index value for the candidate point is plotted on the vertical axis is examined. Actually, the average value and the variance are obtained, and after performing the smoothing process, the evaluation values of the four points are examined to determine the inclination of the graph. According to the result, A, B, C, D according to FIG.
And the other five categories.

If the category is A in a certain direction, it is determined that the value of the optimum motion vector for the direction is included in the range already searched, and the motion vector existing in a wider area is determined. The calculation of the index values of the candidates is stopped, and the calculation is continued with the search accuracy further increased in the vicinity of the motion vector candidate at the position having the lowest value among the already calculated index values.

When the category is B or C, even if estimation of the motion vector is performed with high accuracy, the coding efficiency cannot be expected to be much improved. Therefore, the interval between the candidate points should be increased or the index value should be calculated. Stop. When the category is D, the calculation in the direction in which the index value increases is stopped.

Further, when the same classification is performed after all the candidates within the search range have been calculated and classification into the category D is possible, the search is further performed in the direction in which the index value decreases as shown in FIG. The same processing is performed by expanding the range. FIG. 13 shows an example in which the search range is expanded in the oblique direction 603 and the horizontal direction 604. 6 shaded
The area 06 is the expanded search range.

The classified result is stored in the determination mode memory 215 for each block, and is used as a reference for initial setting of the search parameters as described above when encoding the next image.

Next, another embodiment in which the present invention is applied to a moving picture coding apparatus using a motion compensation prediction method will be described with reference to FIG.
This will be described with reference to FIG.

In FIG. 14, a low-pass filter processing unit 1010 performs low-pass filter processing for suppressing spatially high-frequency components of image data sequentially input from an input terminal 1001, and outputs the processed image data. Sub-sample processing unit 1020
Performs a sub-sampling process of extracting some pixel signals at predetermined pixel intervals from the image data that has been subjected to the filtering process by the low-pass filtering unit 1010, and outputs the result. As described above, the processing for reducing the resolution of the input image data is performed by the low-pass filter processing unit 1010 and the sub-sampling processing unit 1020.

The reference image memory unit 1030 stores an upper layer image as shown in FIG. 19 created by the sub-sample processing unit 1020, and outputs an upper layer image of the coded image.

First block scan conversion processing unit 104
0 is created by the sub-sample processing unit 1020.
An upper layer image of the image to be encoded is output in units of blocks composed of a plurality of pixels.

The first motion vector detection unit 1050 determines a point specified by a motion vector stored in the motion vector storage unit 1060 from the upper layer image of the encoded image stored in the reference image memory unit 1030. A primary detection block having a correlation as high as possible with the block of the upper layer image of the input image data is searched for a centered predetermined search range, and a primary motion vector indicating the position of the primary block is output.

Second block scan conversion processing section 107
A value of 0 outputs image data sequentially input from the input terminal 1001 in units of blocks composed of a plurality of pixels.

The second motion vector detection / motion compensation error calculation section 1080 calculates the type of the primary motion vector output from the first motion vector detection section 1050 and input via the motion vector storage section 1060, that is, the primary motion vector. Depending on where in the search range the vector was found,
Either or both of the center position of the search range and the search accuracy (pixel unit to be searched) are controlled. Then, the second motion vector detection / motion compensation error calculation unit 1080 searches the reference local decoded image stored in the reproduction reference image memory unit 1090 for a secondary detection block having the highest correlation with the block of the input screen. I do. The pixel data of this secondary detection block is input to the local decoding addition unit 1140, and the difference data between the block of the input screen and the secondary detection block is sent to the DCT processing unit 1100, and the secondary detection block The secondary motion vector indicating the position is output to the encoding processing unit 1150 and the motion vector storage unit 1060, respectively.

The motion vector storage unit 1060 stores the primary motion vector from the first motion vector detection unit 1050, and stores the primary motion vector in the second motion vector detection / motion compensation error calculation unit 1
In addition to controlling the search range and search accuracy in 080,
Second motion vector detection / motion compensation error calculator 1080
Is converted into data in the same pixel unit as the search accuracy of the primary motion vector detection and stored, and used for the primary motion vector search of the next image to be encoded.

The DCT processing section 1100 performs a discrete cosine transform process on the block-based difference data output from the second motion vector detection / motion compensation error calculation section 1080 and outputs the result. The quantization processing unit 1110 has a DC
The data output from the T processing unit 1100 is requantized by a division process using a predetermined quantization width. The encoding processing unit 1150 encodes and outputs the motion vector output from the motion vector detection / motion compensation error calculation unit 1080 and the quantized DCT data output from the quantization processing unit 1110.

[0125] The inverse quantization processing unit 1120 applies, to the data output from the quantization processing unit 1110 when encoding an image to be used as a reference image in the future, the quantization used in encoding. Inverse quantization is performed by multiplication processing by the width. The inverse DCT processing unit 1130 includes an inverse quantization processing unit 112
An inverse discrete cosine transform is performed on the data in block units output from 0 to restore the difference data.

The local decoding adder 1140 outputs the pixel data of the secondary detection block output from the second vector detection / motion compensation error calculator 1080 and the inverse DCT processor 113.
0 and the restored difference data output from
The data is output to the reproduction reference image memory unit 1090.

Data 100 encoded in this way
2 is input to the local restoration processing unit 3010 via the communication device or the recording device 2000, and the motion vector information and the quantized DCT data are decoded. Then, the quantized DCT data is output to the inverse quantization processing unit 3020, and the pixel data of the secondary detection block is read from the reference reproduction image memory unit 3050 using the motion vector information.

The inverse quantization processing unit 3020 performs inverse quantization on the quantized DCT data by a multiplication process using the quantization width used at the time of encoding. Inverse DCT
The processing unit 3030 performs an inverse discrete cosine transform on the data in block units output from the inverse quantization processing unit 3020 to restore the difference data. The local decoding addition unit 3040 adds the pixel data of the secondary detection block output from the reference reproduction image memory unit 3050 and the restored difference data output from the inverse DCT processing unit 3030. Output. The reproduction reference image memory unit 3050 includes:
When the playback screen is a screen to be used as a reference screen for the future, the playback pixel data output from the local decoding addition unit 3040 is stored. Line scan conversion processing unit 3060
Outputs the reproduced pixel data output from the local decoding adder 3040 in block units as pixel data in line units for screen display.

FIG. 15 shows a first example of the motion vector detecting method according to the present embodiment.

Here, the upper layer screen R of the screen to be encoded
The description will be made assuming that the block sizes of u and the lower layer screen Rl are 4 pixels and 8 pixels, respectively, and the search range of the upper layer screen Su and the lower layer screen Sl of the reference screen is both ± 2 pixels. As shown in FIGS. 15 (b-1) and (b-5), the first
When the primary detection motion vector obtained by the motion vector detection unit 1050 is the outer periphery of the search range, the search area in the second motion vector detection / motion compensation error calculation unit 1080 is changed to the first motion vector detection unit 1050. The center position of the search range in the second motion vector detection / motion compensation error calculation unit 1080 is further shifted from the center of the search range in the first motion vector detection unit 1050 so that the portion overlapping with the area searched in step 1 is reduced. It is controlled so as to be separated, so that an optimal prediction signal can be obtained for a partial area having a large motion. As a result, the optimum search position B, which was missed in the conventional motion vector detection method shown in FIG.

FIG. 16 shows a second example of the motion vector detecting method according to the present embodiment. Also in this example, the block sizes of the upper layer screen Ru and the lower layer screen Rl of the screen to be encoded are set to 4 pixels and 8 pixels, respectively, and the search ranges of the upper layer screen Su and the lower layer screen Sl of the reference screen are both ±. The description will be made with two pixels. (B-2) of FIG.
As shown in (b-4), when the motion vector obtained by the first motion vector detection unit 1050 is the inner circumference of the search range, the second motion vector detection / motion compensation error calculation unit The search accuracy in 1080 is made fine in half-pixel units, and control is performed so as to search from a narrow range, so that an optimal prediction signal can be obtained for a partial area with small motion. As a result, a delicate search position that could not be detected by the conventional motion vector detection method shown in FIG. 20 can be detected with the same processing amount as in the related art.

FIG. 17 shows a third example of the motion vector detecting method according to the present embodiment. Also in this example, the block sizes of the upper layer screen Ru and the lower layer screen Rl of the screen to be encoded are set to 4 pixels and 8 pixels, respectively, and the search ranges of the upper layer screen Su and the lower layer screen Sl of the reference screen are both ±. The description will be made with two pixels. (B-1) of FIG.
As shown in (b-5) or (b-5), when the motion vector obtained by the first motion vector detection unit 1050 is the outer periphery of the search range, the second motion vector detection / motion compensation error calculation unit 1080 The center position of the search range in the second motion vector detection / motion compensation error calculator 1080 is set to the first motion so that the searched area has less overlap with the area searched by the first motion vector detector 1050. Control is performed so as to be further away from the center of the search range in the vector detection unit 1050 so that an optimal prediction signal can be obtained for a partial area having a large motion.

As shown in (b-2) and (b-4) of FIG. 17, when the motion vector obtained by the first motion vector detection unit 1050 is the inner circumference of the search range, the The motion vector detection / motion compensation error calculation unit 1080 controls the center position of the search area to be the position indicated by the primary detection motion vector, so that an optimal prediction signal can be obtained for a normal motion partial area. To

Further, as shown in FIG. 17 (b-3),
If the motion vector obtained by the first motion vector detection unit 1050 is located at the center of the search range, the second
Is controlled so that the center position of the search range in the motion vector detection / motion compensation error calculation unit 1080 becomes the position indicated by the primary detection motion vector, so that an optimal prediction signal can be obtained for a partial region of normal motion. I do. As a result, FIG.
The subtle search position that could not be detected by the conventional motion vector detection method shown in (1) can be detected with the same processing amount as before.

Further, the second motion vector detection / motion compensation error calculation section 1080 controls the search so that the search is performed in a narrow range by making the search in a pixel unit and a half pixel unit, so that the optimum prediction is performed in a small motion partial area. Get a signal. As a result, an optimal search position that has been missed by the conventional motion vector detection method shown in FIG. 20 or a delicate search position that cannot be detected can be detected with the same processing amount as before.

FIG. 18 shows a fourth example of the motion vector detecting method according to the present embodiment. In this example, the upper layer screen Ru and the lower layer screen Rl of the screen to be encoded have block sizes of 4 pixels and 8 pixels, respectively, and the search range of the upper layer screen Su and the lower layer screen Sl of the reference screen is an even number. Description will be given assuming that there are certain eight pixels.

As shown in FIG. 18A, the positions of the sample points on the upper layer screen of the screen to be coded and the positions of the sample points on the upper layer screen of the reference screen are in opposite phases to each other. Thus, the search in the upper layer can be symmetrical. Therefore, even if the possessed search range is an even number, the search range of the lower layer screen can be symmetrical by using the first and third examples, and the possessed processing capacity can be effectively reduced. Can be used.

[0138]

According to the present invention, it is possible to provide an image coding apparatus which is more efficient and has less visual deterioration in image quality.

Further, according to the present invention, coding can be performed by performing motion compensation prediction with a smaller amount of operation, and at the same time, the coding efficiency can be improved even when the motion of the image is small and the search range of the motion vector is not sufficient. It is possible to provide an image encoding device for a moving image in which the image quality does not decrease.

Furthermore, according to the present invention, by controlling at least one of the search range and the search accuracy of the final secondary motion vector in accordance with the region in the search range from which the primary motion vector is detected, An optimal prediction signal can always be obtained irrespective of the magnitude of image motion, and an image encoding device for a moving image with high encoding efficiency can be provided.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an embodiment of an image encoding apparatus according to the present invention.

FIG. 2 is an explanatory diagram of adaptive quantization according to the present invention.

FIG. 3 is a diagram showing an example of a change in quantization width due to adaptive quantization.

FIG. 4 is a diagram showing characteristics of a change in a quantization width with respect to an activity;

FIG. 5 is an explanatory diagram of an adaptive quantization method according to the present invention.

FIG. 6 is a block diagram showing another embodiment of the image encoding apparatus of the present invention.

FIG. 7 is a block diagram showing an embodiment in which the present invention is applied to an image coding apparatus using inter-frame prediction.

FIG. 8 is a block diagram showing an embodiment in which the present invention is applied to a moving picture coding apparatus using motion compensated prediction coding.

9 is a block diagram showing a configuration of a motion vector search unit and a search range setting unit in FIG.

FIG. 10 is an explanatory diagram of setting of a motion vector search center in the embodiment.

FIG. 11 is a view showing characteristics of motion vector candidates and their index values in the embodiment.

FIG. 12 is a diagram illustrating a setting range of a motion vector search range and a motion vector candidate in the embodiment.

FIG. 13 is an explanatory diagram of resetting of search parameters in the embodiment.

FIG. 14 is a block diagram showing another embodiment in which the present invention is applied to a moving picture coding apparatus using motion compensated predictive coding.

FIG. 15 is a diagram showing a first example of a motion vector detection method in the embodiment.

FIG. 16 is a diagram showing a first example of a motion vector detection method in the embodiment.

FIG. 17 is a diagram showing a first example of a motion vector detection method in the embodiment.

FIG. 18 is a diagram showing a first example of a motion vector detection method in the embodiment.

FIG. 19 shows a pyramid structure.

FIG. 20 is an explanatory diagram of a conventional motion vector detection method.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 101 ... Block extraction part 102 ... Two-dimensional discrete orthogonal transformation part 103 ... Quantizer 104 ... Variable length coding part 105 ... Buffer 106 ... Block activity calculation part 107 ... Basic quantization table 108 ... Adaptive quantization control part 109 ... Quantization width control parameter calculation unit 110: total activity calculation unit 111: memory 112: quantization change width calculation unit 113: two-dimensional low-pass filter 114: classifying unit 115: inverse quantizer 116: two-dimensional discrete orthogonal inverse transformation Unit 117 Frame memory 118 Motion compensation unit 119 Scene change detection unit 120 Multiplexing unit 121 Memory 201 Block cutout unit 202 Encoding unit 203 Search parameter setting unit 204 Motion vector search unit 205 Image buffer 206: Local decoding unit 207: Multiplexing unit 208 ... Optimal vector detection unit 209 ... Reference candidate block cutout unit 210 ... Vector candidate generation unit 211 ... Search parameter initial setting unit 212 ... Vector memory 214 ... Search parameter resetting unit 215 ... Judgment mode memory 216 ... Index Calculation unit 1020: Sub-sample processing unit 1030: Reference image memory unit 1040, 1070: Block scan conversion processing unit 1050: First motion vector detection unit 1080: Second motion vector detection / motion compensation error calculation unit 1090: For reference Playback image memory

──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Tomoo Yamakage 1 Toshiba-cho, Komukai-shi, Kawasaki-shi, Kanagawa Prefecture Within Toshiba Research Institute, Inc. (72) Inventor Tadahiro Oku 1 Toshiba-cho, Komukai-shi, Kawasaki-shi, Kanagawa Address Toshiba Research Laboratory Co., Ltd. (56) References JP-A-3-117991 (JP, A) JP-A-2-33286 (JP, A) JP-A-1-2333893 (JP, A) JP-A-3-233 4686 (JP, A) JP-A-2-145079 (JP, A) JP-A-60-177782 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) H04N ^7/ 24-7 / 68

Claims (5)

(57) [Claims] 1. A signal processing means for performing processing for lowering resolution on image data, and a block to be coded in input image data processed by the signal processing means is processed by the signal processing means. First motion vector detecting means for detecting a primary motion vector over a first search range with reference to a reference image block in the encoded image data after the decoding, and input image data not processed by the signal processing means. Second motion vector detection for obtaining a secondary motion vector with a predetermined search accuracy over a second search range with reference to a reference image block in the coded image data which is not processed by the signal processing means, for the block to be coded. means and said primary motion vector detected by the first motion vector detecting means in within the first search range Portion and the outer
At least one of the size, the center position, and the search accuracy of the second search range in the second motion vector detection means depends on which part of the peripheral portion is detected.
Control means for controlling one ; coding means for coding a difference between the reference image block specified by the secondary motion vector detected by the second motion vector detection means and the coding target block; An image encoding device comprising: 2. The image encoding apparatus according to claim 1, wherein said signal processing means performs processing including sub-sampling processing on the image data to reduce the resolution. 3. The control means according to claim 1, wherein said primary motion vector is obtained from a central portion of said first search range.
When the search accuracy of the second motion vector detecting means is made finer and the size of the second search range is made narrower, and the primary motion vector is obtained from the outer peripheral portion of the first search range, 2. The image encoding apparatus according to claim 1, wherein the search accuracy of the second motion vector detecting means is made coarser and the size of the second search range is made wider. 4. The control means according to claim 1, wherein said primary motion vector is obtained from an outer peripheral portion of said first search range.
2. The image coding apparatus according to claim 1, wherein a center position of the second search range is further away from a center position of the first search range. 5. The control means according to claim 1, wherein said primary motion vector is obtained from a central portion of said first search range.
The center position of the second search range is set as the position indicated by the first motion vector, and the search accuracy of the second motion vector detecting means is made finer to obtain the second search range.
More narrowly the size of, when the primary motion vector obtained from the outer peripheral portion of said first search range, from the center position of the second search range from the center of the first search range So that the search accuracy of the second motion vector detection means is coarser and the size of the second search range is larger.
2. The image coding apparatus according to claim 1, wherein the image coding apparatus further increases the width.