JP2000102017A - Image encoder and image encoding system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a highly efficient encoding system by freely setting a GOP(group of picture) structure and switching 'frame structure encoding' and 'field structure encoding'. SOLUTION: A pre-processing part 1 obtains an absolute value difference sum 11 in one field and the absolute value difference sum 12 between two fields by using image signals 102-104 inputted from an image memory 3 and a control part 4 compares the values of both and selects the encoding of the 'field structure encoding' when the absolute value difference 131 in one field is smaller and the 'frame structure encoding' when the absolute value difference sum 132 between the two fields is smaller. The image memory 3 outputs the image signal 151 by a field unit when the 'field structure encoding' is selected and by a frame unit when the 'frame structure encoding' is selected and a video encoding part 2 encodes the image signal 151 in a selected encoding mode.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a method for selecting a coding mode for performing more efficient coding when coding a moving image.

[0002]

2. Description of the Related Art MPEG-2 (Moving Picture E), which was internationally standardized in ISO / IEC in 1994 as a digital image encoding / decoding technique, has been developed.
xparts Group, Phase-2) is a typical example. In this recommendation, only the details of the decoding method are determined, and a certain degree of freedom is provided for the coding method. For example, when one of a plurality of modes can be selected in performing encoding, which mode is to be selected is left to the encoding apparatus, and therefore, even if the same MPEG-2 encoding method is used, the output is The coded data and image quality to be performed may differ depending on the device.

For example, not only MPEG-2 but also MPEG
The method of configuring “GOP structure coding”, which is also used in −1, is exemplified. In MPEG, three encoding types of “I picture”, “P picture”, and “B picture” are defined as typical encoding modes in units of pictures. The "I picture" is intra-frame coded and is usually inserted once into N frames (N is a natural number and need not be a fixed value). As a key frame for random access or when an error-producing transmission path is used, if an inter-frame prediction is performed, an error propagates to a succeeding picture. It is used for performing. The “P picture” is one-way motion compensated prediction interframe encoded, and is inserted once into M frames (M is a natural number and need not be a fixed value). An image that can be used for reference of prediction is either an “I picture” or an already encoded “P picture”. “B picture” is motion-compensated prediction inter-frame coding from both directions, and is performed using “I picture” or “P picture” as a prediction reference image. Since “B picture” can use two pictures as reference images, the overall coding efficiency can be made extremely high when the prediction efficiency of motion compensation prediction is high.

FIG. 11 shows the relationship between I / P / B pictures when N = 9 and M = 3 are fixed values. The figure shows the order displayed on the screen, and rearrangement is performed when encoding is actually performed. In the coding order, the first to "I picture" to the next "I picture" are referred to as "GOP (Grou
pOfPicture), and a GOP structure is often expressed using two parameters, N and M, in general. Regarding how to set the values of N and M, see section 4.1.1 of the introduction of the above recommendation, "These three image types can be organized very flexibly, and the choice is coded. It depends on the requirements of the device development and the application. "

[0005] The MPEG-2 video encoding system is
In order to increase the coding efficiency for interlaced signals, which is the current television signal scanning method, several modes for coding interlaced signals are provided. Here, the interlace signal is a signal in which one frame is formed by alternately scanning the scanning lines of the two fields of the top field and the bottom field, and 1/1 is provided between the top field and the bottom field. A time difference of two frame times occurs.
One mode for the encoding is "frame structure encoding" and "field structure encoding" for each frame.
There is one that can choose either one of. These two modes are described in the introduction of the above recommendation.
As described in section 1.2, it is general that a frame structure is generally suitable when movement is limited and a fine pattern is included, and a field structure is suitable when there is severe movement. Is said to.

[0006]

SUMMARY OF THE INVENTION As described above, MPEG
The -2 recommendation states that the GOP structure can be set freely and that "frame structure coding" and "field structure coding" can be switched for each frame. There is no description on the switching method, and efficient coding may not be possible.

[0007] As a result of the investigation, Japanese Patent Application Laid-Open No. 5-95545
And Japanese Patent Application Laid-Open No. 6-78289, both of which describe a method relating to a selection method of “frame structure coding” and “field structure coding”. An invention is described in which the sum of absolute value differences of each pixel between a field (top field) and a second field (bottom field) is calculated, and a comparison is made between a value of the absolute value difference story and a threshold to select. But
According to the present invention, it is only possible to determine the presence or absence of correlation between fields, and there is no formation of the presence or absence of correlation between frames (between two fields), so that there is a problem that sufficient efficiency cannot be improved. In the latter case, the absolute value difference sum for the prediction difference signal after the inter-frame motion compensation prediction obtained for each macroblock and the absolute value difference sum for the prediction difference signal after the inter-field motion compensation prediction are encoded. Integrate for the frame to be integrated and select the one with the smaller integrated data, find the median value of the motion vector, select the "field structure coding" when the value is larger, select one line per macroblock And the correlation between two lines are determined to determine which macroblock is larger in the entire frame, and selection is made. For each macroblock, the difference square sum between one line and the The invention has been described in which the sum of squared differences is calculated for the macroblocks of the entire frame, and the smaller value is selected. Since the selection of which structure is to be performed for the first time after the coding of the macroblock is completed, both coding results must be accumulated for one frame. There is a problem that the configuration of the apparatus becomes large and a delay of one frame period occurs before the output of the encoding result starts.

Therefore, according to the present invention, the GOP structure can be freely set, or "frame structure coding" can be performed for each frame.
An object of the present invention is to provide an image coding apparatus and an image coding method capable of obtaining a highly efficient coding method by clarifying a method for switching between the field coding and the “field structure coding”.

[0009]

According to the first aspect of the present invention, there is provided an image encoding apparatus for encoding a digitized moving image signal. A memory, an encoding unit that encodes an image signal output from the image memory in accordance with an encoding mode, and reads an image signal from the image memory to obtain a statistic between one field and a statistic between two fields. A pre-processing unit for controlling a structure of an image signal output from the image memory based on a statistic between one field and a statistic between two fields obtained in the pre-processing unit; And a control unit for controlling an encoding mode in the unit.

[0010] According to the second aspect of the present invention, in the first aspect,
In the image encoding device described in the description, the preprocessing unit calculates a sum of absolute value differences between one field as a statistic between one field,
It is characterized in that a sum of absolute difference between two fields is obtained as a statistic between two fields.

According to the third aspect of the present invention, there is provided the second aspect.
In the image encoding apparatus described above, when the input image signal is an interlace signal, the control unit compares the value of the absolute value difference sum between one field and the absolute value difference sum between two fields obtained by the preprocessing unit. When the sum of absolute value differences between one field is smaller, encoding is performed in units of one field,
When the sum of absolute value differences between two fields is smaller, the structure of the image signal output from the image memory is controlled so that the encoding is performed in units of one frame, and the encoding mode in the encoding unit is controlled. Features.

Further, in the invention according to claim 4, the encoding unit includes intra-frame encoding, one-way predictive encoding, and bidirectional predictive encoding as encoding modes, and the control unit includes a pre-processing unit. Is calculated from the sum of absolute differences between one field and the sum of absolute differences between two fields, and when the smaller value is smaller than a predetermined threshold, the number of frames of bidirectional predictive coding is increased. Encoded, and when the smaller value is larger than a predetermined threshold, control the structure of the image signal output from the image memory so that the encoding is performed by reducing the number of frames of bidirectional prediction encoding. The encoding mode is controlled in the encoding unit.

According to a fifth aspect of the present invention, there is provided an image encoding apparatus for encoding a digitized moving image signal, comprising: an image memory for storing a plurality of frames of the image signal;
An encoding unit that encodes the image signal output from the image memory in accordance with an encoding mode, and that outputs an encoding parameter used in the encoding; and an encoding unit that outputs the encoding parameter output from the encoding unit. Based on the threshold of
A control unit that controls a structure of an image signal output from the image memory and controls a coding mode in the coding unit.

Further, according to the invention described in claim 6, according to claim 5,
In the image encoding apparatus according to the aspect, when the input image signal is an interlaced signal, the encoding unit outputs a motion vector as an encoding parameter, and the control unit receives the motion vector from the encoding unit, When the average value of the magnitude of the motion vector is larger than a predetermined threshold, the image is encoded in units of one field, and when the average value of the motion vector is smaller than the predetermined threshold, the image is encoded in units of one frame. It is characterized by controlling the structure of the image signal output from the memory and controlling the coding mode in the coding unit.

[0015] According to the invention described in claim 7, according to claim 5,
In the image encoding device described, the encoding unit, as the encoding mode, intra-frame encoding, unidirectional predictive encoding,
Bidirectional predictive coding, and outputs a motion vector as a coding parameter, the control unit receives the motion vector from the coding unit, and when the average of the magnitude of the motion vector is smaller than a predetermined threshold, the control unit The encoding is performed by increasing the number of frames of the predictive encoding, and when the average of the magnitude of the motion vector is larger than a predetermined threshold, the frame is output from the image memory so that the number of frames of the bidirectional predictive encoding is reduced. Control the image signal structure
The encoding mode is controlled in the encoding unit.

According to the eighth aspect of the present invention, in the fifth aspect,
In the image encoding device described, the encoding unit, as the encoding mode, intra-frame encoding, unidirectional predictive encoding,
Bidirectional predictive encoding, and outputs a motion vector as an encoding parameter, the control unit inputs the sum of squares of the prediction difference signal from the encoding unit, the sum of squares of the prediction difference signal is greater than a predetermined threshold When the size is smaller, the number of frames of the bidirectional prediction coding is increased to perform coding, and when the average of the magnitude of the motion vector is larger than a predetermined threshold, the number of frames of the bidirectional prediction coding is reduced and the image memory is coded. And the encoding mode in the encoding unit is controlled.

According to the ninth aspect of the present invention, in the fifth aspect,
In the image encoding device described above, when the input image signal is an interlaced signal, the encoding unit outputs a prediction type of motion compensation prediction as an encoding parameter, and the control unit outputs a prediction of motion compensation prediction from the encoding unit. When the number of prediction types that are field predictions is larger than a predetermined threshold, the encoding is performed in units of one field. When the number of prediction types that are field predictions is smaller than a predetermined threshold, the encoding is performed in units of one frame. It is characterized by controlling the structure of the image signal output from the image memory so as to be encoded, and controlling the encoding mode in the encoding unit.

According to a tenth aspect of the present invention, in the image encoding method for encoding a digitized moving image signal, the statistic between one field and the statistic between two fields of the image signal is determined. A pre-processing step for obtaining, and a control step of controlling a coding mode while controlling a structure of an image signal based on the statistics between one field and the statistics between two fields obtained in the pre-processing step. And an encoding step of encoding the image signal, the structure of which is controlled in the control step, according to the controlled encoding mode.

According to the eleventh aspect of the present invention, in an image encoding method for encoding a digitized moving image signal, the image signal is encoded in accordance with an encoding mode and used for encoding. Encoding step of outputting an encoding parameter, based on the encoding parameter and a predetermined threshold output in the video encoding step, based on the structure of the image signal and the encoding mode in the video encoding step A control step to control;
It is characterized by having.

[0020]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiment 1 Hereinafter, an image encoding apparatus according to a first embodiment of the present invention will be described. FIG.
1 shows a configuration of an image encoding apparatus according to Embodiment 1 of the present invention. In FIG. 1, 1 is a pre-processing unit, 2 is a video encoding unit that encodes an image signal output from the image memory 3 in an encoding mode specified by the control of a control unit 4 described later, and 3 is a plurality of frames. And an image memory 4 for holding an image signal and outputting an image signal 151 having a structure in accordance with the control of the control unit 4 described later, and an image output from the image memory 3 based on the signal output from the preprocessing unit 1. Signal 15
1 and a control unit for controlling the encoding mode of the video encoding unit 2;
Reference numerals 3 and 14 denote difference processing units. Note that the video encoding unit 2
Is "field structure encoding" as the encoding mode.
And “frame structure coding”.

Reference numeral 101 denotes a digitized interlaced image signal input to the image memory;
Reference numerals 3 and 104 denote image signals for one field input from the image memory 3, 121 and 122 are difference signals, 131 is a sum of absolute value differences between one field, 132 is a sum of absolute value differences between two fields, and 141 is a coding mode selection. A signal, 151 is an image signal, and 161 is encoded data.

FIG. 2 shows image signals 102, 103 and 104.
Is shown in the context. In FIG. 2, the image signal 103 precedes the image signal 102 by one field time. The image signal 104 is 2
Field time, that is, one frame time ahead. Here, the description has been made with the image signal 102 as the top field, but the image signal 102 may be in the bottom field.

Next, the operation will be described. The preprocessing unit 1 includes image signals 102, 103, and 104 input from the image memory 3.
Is used to obtain a sum of absolute differences 11 between one field and a sum of absolute differences 12 between two fields. Specifically, first, the difference processing unit 13 calculates a difference between the image signal 102 and the image signal 103 to obtain a difference signal 121, and the addition processing unit 11 performs an addition process on the absolute value of the difference signal 121 for one field signal. To obtain a sum 131 of absolute value differences between one field. Similarly, the difference processing unit 14 obtains a difference between the image signal 102 and the image signal 104 to obtain a difference signal 122, and the addition processing unit 12 performs an addition process on the absolute value of the difference signal 122 for one field signal. An inter-field absolute value difference sum 132 is obtained. Each time the addition processing for one field is performed in this manner, the addition processing unit 11 and the addition processing unit 1
2 and the sum of absolute value differences 131 between one field respectively
And the sum of absolute value differences 132 between the two fields are output,
Output to the control unit 4.

The control unit 4 compares the two values, and when the one-field absolute value difference sum 131 is smaller, selects “field structure encoding” as the encoding mode, and selects the two-field absolute value difference sum. When the value of 132 is smaller, “frame structure coding” is selected as the coding mode. The selected result is output to the image memory 3 and the video encoding unit 2 as the encoding mode selection signal 141.

The image memory 3 selects "frame structure encoding" for each field when "field structure encoding" is selected based on the encoding mode selection signal 141 from the control unit 4. Output, the image signal 151 is output for each frame.

The video encoding unit 2 receives the image signal 151, and selects the encoding mode selected by the encoding mode selection signal 141 from the control unit 4, "field structure encoding" or "frame structure encoding". ", And outputs encoded data 161.

Therefore, according to the first embodiment, “field structure coding” or “frame structure coding” is performed according to the magnitude of the one-field absolute value difference sum 131 and the two-field absolute value difference sum 132. Is selected, it is possible to select the encoding mode having a higher correlation between the image signals, and it is possible to execute highly efficient encoding.

Embodiment 2 Next, a second embodiment of the image coding apparatus according to the present invention will be described. FIG. 3 shows the configuration of an image encoding apparatus according to Embodiment 2 of the present invention.
3, reference numeral 31 denotes a threshold memory, 133 denotes a threshold previously stored in the threshold memory 31, 171 denotes a GOP structure control signal, and 172 denotes an encoding mode control signal. Since the other configuration is the same as that of the first embodiment shown in FIG. 1, the same reference numerals are given and the description is omitted. However,
The video encoding unit 2 according to the second embodiment performs “field structure encoding” that performs “intra-frame encoding”, “unidirectional prediction encoding”, and “bidirectional prediction encoding” as encoding modes, "Frame structure coding".

Next, the operation will be described. Embodiment 1
Therefore, only the operation unique to the second embodiment will be described. Pre-processing unit 1 sums the absolute value difference between one field 1
31 and the sum of absolute value difference 132 between two fields are calculated and output to the control unit 4, the control unit 4 calculates the sum of absolute value difference 131 between one field input from the preprocessing unit 1 and the absolute value difference between two fields 131 Comparison with the value difference sum 132 is performed, and as a result of the comparison, the smaller value (hereinafter referred to as α) is compared with the value of the threshold value 133 set in the threshold value memory 31.

Here, when α is larger than the threshold value 133,
The control unit 4 outputs a GOP structure control signal 171 having a GOP structure of M = 1 and N = 9 to the image memory 3 so that the encoding is performed as shown in FIG. Control is performed so that the image signal 151 having the GOP structure shown in FIG. In FIG. 4, N
= 9 indicates an image signal in which “I picture” is inserted once in nine frames. Here, since M = 1 is set, “P picture” is inserted once in one frame, ie, every eight frames. I have.

At the same time, the control unit 4 determines that the video encoding unit 2 inputs from the image memory 3 but M = 1, N =
An encoding mode control signal 172 corresponding to the GOP structure control signal 171 is output to the video encoding unit 2 so that the image signal 151 having the GOP structure of No. 9 can be encoded. In addition,
In this case, since the GOP structure is M = 1 and N = 9,
In accordance with the output of the image signal 151 from the image memory 151 according to the encoding mode control signal 172, an encoding mode control signal for intra-frame encoding for encoding "I picture" is sent once, and then "P picture" The transmission of the encoding mode control signal of the unidirectional predictive encoding for encoding is repeated eight times in succession.

Conversely, when α is smaller than the threshold value 133, the control unit 4 controls the GOP structure control signal 171 to have a GOP structure of M = 3, N = 9 so that the encoding is performed as shown in FIG. Is output to the image memory 3, and control is performed such that the image signal 151 having the GOP structure shown in FIG. 5 is output from the image memory 3 to the video encoding unit 2. In FIG. 5, similarly to FIG. 4, “I picture” is inserted once in nine frames at N = 9, but “M” is set to M, so that “P picture” is inserted once in three frames, ie, “I picture”. "B picture" is inserted twice in three frames.

At the same time, the control unit 4 determines that M = 3 and N =
An encoding mode control signal 172 corresponding to the GOP structure control signal 171 is output to the video encoding unit 2 so that the image signal 151 having the GOP structure of No. 9 can be encoded. In addition,
In this case, since the GOP structure is M = 3 and N = 9,
In accordance with the output of the image signal 151 from the image memory 151 by the encoding mode control signal 172, "IBBPBB"
Control signals are sent in the order of "PBB". .

As a result, the image data 161 output from the video encoding unit 2 has the inter-field absolute value difference sum 1
This is data encoded in a GOP structure based on the smaller value α of the 31 and the sum of absolute value differences 132 between two fields 132 and the threshold value 133.

Therefore, according to the second embodiment, if the value of the absolute value difference sum α is smaller than the threshold value, it means that the correlation between the images is high, and the video encoding unit 2 encodes. Even if the M value of the GOP structure of the image signal is increased and a B picture is inserted between the I picture and the P picture, the prediction efficiency of the motion compensation is sufficiently obtained. If the value is larger, it means that the correlation between the images is low. Therefore, if the M value of the GOP structure of the image signal is increased, the efficiency of the motion compensation prediction of the P picture decreases. When the immediately preceding frame is used as the prediction reference image of the P picture, a decrease in encoding efficiency can be prevented, and highly efficient encoding can be performed.

Embodiment 3 In the second embodiment, in order to simplify the description, the GOP structure of the image signal encoded by the video encoding unit 2 is limited to two, that is, M = 1 and M = 3. Although described above, it is of course possible to take other M values. For example, the absolute value difference sum α
Is small, M = 4, 5, not M = 3.
.., Or when the value of α is large, M = 2, etc. instead of M = 1. Also, the value of N is 9
However, it is needless to say that a value other than N = 9 may be used.

Embodiment 4 FIG. In the second and third embodiments, the preprocessing unit 1 determines whether the inter-field absolute value difference sum 131, 1
32, the control unit 4 controls the GOP structure of the image signal 151 output from the image memory 3 to the video encoding unit 151. In the fourth embodiment, however, the control unit 4 outputs the image signal 151 from the video encoding unit 2. GO according to encoding parameter 162
The encoding is performed by controlling the P structure.

FIG. 6 shows the configuration of an image coding apparatus according to the fourth embodiment. In the figure, reference numeral 162 denotes an encoding parameter used for encoding transmitted from the video encoding unit 2 to the control unit 4. In the fourth embodiment, the pre-processing unit 1 provided in the second embodiment and the like is unnecessary, and the other configuration is the same as that of the second embodiment shown in FIG. The same reference numerals are given and the description is omitted.

Next, the operation will be described. First, when the digitized image signal 101 is input to the image memory 3 and output from the image memory 3 to the video encoding unit 2 as an image signal 151, the video encoding unit 2, although not shown,
Usually, motion compensation prediction and discrete cosine (DCT) transform are performed for each picture.

At this time, in the fourth embodiment, the video encoding unit 2 outputs the motion vector used for the motion compensation prediction as the encoding parameter 162 and sends it to the control unit 4. Since a motion vector is obtained for each small area of 16 pixels × 16 lines (called a macro block), a plurality of motion vectors are output from one screen.

The control unit 4 inputs a motion vector as an encoding parameter 162 from the video encoding unit 2, and when the average value of the magnitude of the motion vector is larger than the threshold value 31 stored in the threshold value memory 31, In the case where there is a variation in the direction of each motion vector, the motion between the images is large and the efficiency of motion compensation prediction is reduced, so the GOP structure control signal 171 for reducing the value of M in the GOP structure is transmitted to the image memory 3. And the GOP structure control signal 17
The coding mode control signal 172 of “intra-frame coding”, “one-way predictive coding”, or “bidirectional predictive coding” corresponding to 1 is sent.

Conversely, when the average value of the magnitudes of the motion vectors is smaller than the threshold value, the control unit 4 increases the efficiency of the motion compensation prediction because the motion between the images is small.
The value of M of the OP structure is increased to send a GOP structure control signal 171 to the image memory 3 and the GOP structure control signal 1
An encoding mode control signal 172 corresponding to 71 is sent.

As a result, the image data 161 output from the video encoding unit 2 is based on the GO based on the motion vector.
It becomes data encoded in the P structure.

Therefore, according to the fourth embodiment, it is possible to perform encoding by changing to the image signal 151 having the GOP structure according to the correlation between the image signals based on the motion vector. Becomes possible.

Embodiment 5 FIG. In the fourth embodiment, the method of changing the GOP structure according to the value of the motion vector has been described. However, in the fifth embodiment, “frame structure coding” is performed using motion parameters that are the same coding parameters. And a coding mode of “field structure coding”.

FIG. 7 shows a configuration of an image coding apparatus according to the fifth embodiment. In the figure, reference numeral 162 denotes an encoding parameter of an encoding condition used for encoding transmitted from the video encoding unit 2 to the control unit 4, and in this embodiment 5, it is a motion vector. In the fifth embodiment,
The pre-processing unit 1 provided in the first embodiment and the like is unnecessary, and other configurations are the same as those of the first and second embodiments shown in FIGS.
Since the configuration is the same as that of the second configuration, the same reference numerals are given and the description thereof is omitted.

Next, the main operation will be described. In the fifth embodiment, the control unit 4 inputs a motion vector as the coding parameter 162 from the video coding unit 2 and outputs the average value of the magnitude of the motion vector. Is larger than the threshold value 133 stored in the threshold value memory 31, the motion between images is large, and the coding mode of “field structure coding” is selected.
1 is output to the image memory 3 and the video encoding unit 2.

Conversely, when the average of the magnitudes of the motion vectors is smaller than the threshold value 133, the control unit 4 determines that the signal correlation between the top field and the bottom field is high because the motion between the images is small. , The encoding mode of “frame structure encoding” is selected, and the selected result is output to the image memory 3 and the video encoding unit 2 as an encoding mode selection signal 141.

Therefore, according to the fifth embodiment, even if the preprocessing unit 1 in the first embodiment is not provided, the “field structure encoding” is performed based on the motion vector at the time of encoding in the video encoding unit 2. By selecting the coding mode of “frame structure coding” and “frame structure coding”, it is possible to select the coding mode having a higher correlation between the image signals, and to perform highly efficient coding. .

Embodiment 6 FIG. In the fourth embodiment, the motion vector value is used as the encoding parameter 162 output from the video encoding unit 2 to the control unit 4. However, there are some other parameters that can be used. The sixth embodiment is characterized in that the video encoding unit 2 outputs the sum of squares of the prediction difference signal after the motion compensation prediction as the encoding parameter 162 instead of the motion vector. The configuration of the sixth embodiment is the same as the configuration of the fourth embodiment shown in FIG. 6, and the operation will be described with reference to FIG.

That is, in the sixth embodiment, as shown in FIG.
The controller 4 outputs the sum of squares of the prediction difference signal after motion compensation prediction, and the controller 4 inputs the sum of squares of the prediction difference signal after motion compensation prediction, and the value of the sum is larger than the threshold value 31 stored in the threshold value memory 13. If the value is large, the efficiency of the motion compensation prediction is reduced. Therefore, the GOP structure control signal 171 and the coding mode control signal 172 are stored in the image memory so that the value of M in the GOP structure is reduced and coding is performed. 3 and the video encoding unit 2 to control the encoding. Conversely, when the sum of squares of the prediction difference signal after the motion compensation prediction is smaller than the threshold value 31, the value of M of the GOP structure is changed. The GOP structure control signal 171 and the coding mode control signal 172 are output to the image memory 3 and the video coding unit 2 to control the coding so that the coding is performed at a larger size.

Therefore, according to the sixth embodiment, the image signal 1 is calculated based on the sum of squares of the prediction difference signal after the motion compensation prediction.
Since encoding is performed by changing the GOP structure of 51, similarly to the case of the motion vector, highly efficient encoding corresponding to the correlation between image signals is executed without a preprocessing unit as in the fourth embodiment. It is possible to do.

Embodiment 7 FIG. In the seventh embodiment, encoding parameter 1 to be output from video encoding unit 2 to control unit 4
62, the number of macroblocks for intra-frame coding and the number of macroblocks for inter-frame coding are output instead of the sum of squares of the motion vector and the prediction difference signal after motion compensated prediction. For this reason, the configuration of the seventh embodiment is the same as the configuration of the fourth embodiment shown in FIG. 6, and therefore the operation will be described based on FIG.

In other words, in the seventh embodiment, the video coding unit 2 determines whether the macroblock, which is the coding unit of the image signal 151, uses either the intra-frame coding mode or the inter-frame coding mode. It counts whether or not encoding has been performed, and counts each count value into the encoding parameter 162.
Is output to the control unit 4.

The control unit 4 accumulates the number of coding modes of each macroblock input as the coding parameter 162 to determine the number of intra-frame coded macroblocks and the number of inter-frame coded macroblocks. In each case, if the number of inter-coded macroblocks is smaller than the threshold value 133 stored in the threshold value memory 31,
Since the efficiency of the motion compensation prediction is reduced, the GOP structure control signal 171 and the coding mode control signal 172 are transmitted to the image memory 3 and the video coding system so that the coding is performed with the value of M of the GOP structure reduced. On the other hand, when the number of inter-coded macroblocks is larger than the threshold value 133, it indicates that the efficiency of motion compensation prediction is high. The GOP structure control signal 171 and the encoding mode control signal 172 are sent to the image memory 3 and the video encoding unit 2 so as to be executed.

Therefore, according to the seventh embodiment, the image signal 1 is set based on the number of coding modes of each macroblock.
Since the encoding is performed by changing the GOP structure of 51, it is possible to execute highly efficient encoding according to the correlation between the image signals without the preprocessing unit as in the fourth embodiment.

Embodiment 8 FIG. Embodiment 8 is characterized in that the number of prediction types of motion compensation prediction for each macroblock is accumulated, and coding selection between "field structure coding" and "frame structure coding" is performed. I do. For this reason, the configuration of the eighth embodiment is the same as the configuration of the fifth embodiment shown in FIG. 7, and the operation will be described based on FIG.

That is, for example, in the case of "frame structure coding" in MPEG-2, there are two types of motion compensation prediction, "field prediction" and "frame prediction" for each macroblock. , Video encoding unit 2
Outputs the number of prediction types when performing motion compensation prediction on each macroblock as an encoding parameter 162.

The control unit 4 accumulates the prediction type for each macroblock inputted as the encoding parameter 162 for one picture, and when "field prediction" is more, the correlation of the signal between two fields is calculated. The encoding mode selection signal 141 to select the "field structure encoding" because of its low performance.
On the other hand, if there are many “frame predictions”,
The encoding mode selection signal 141 is sent to the image memory 3 and the video encoding unit 2 so as to select “frame structure encoding” because the correlation of the image signal between the two fields is high.

Therefore, according to the eighth embodiment, the number of motion-compensated prediction types of each macroblock output from the video encoding unit 2 as the encoding parameter 162 is accumulated, and the “field structure code Pre-processing unit 2 for selecting and encoding “encoding” and “frame structure encoding”.
Without this, it is possible to execute highly efficient encoding according to the correlation of the image signals.

Embodiment 9 FIG. In addition, the first to eighth embodiments are described.
Is implemented not only individually, but also in each of Embodiments 1 to
Of course, it is also possible to carry out the combination of the eight.
Therefore, in the ninth embodiment, as an example of a combination of each of the first to eighth embodiments, an example of an image encoding device in which the first embodiment and the fourth embodiment are combined will be described. FIG.
The configuration of an image coding apparatus according to the ninth embodiment in which the first embodiment and the fourth embodiment are combined is shown below. 8, the same components as those in the configuration of the first embodiment shown in FIG. 1 or those in the fourth embodiment shown in FIG. 6 are denoted by the same reference numerals, and the description thereof will be omitted. The operation will be described.

That is, the control unit 4 of the eighth embodiment
In order to combine the functions of the control unit 4 of the first and second embodiments, the one-field absolute value difference sum 131 and the two-field absolute value difference sum 132 from the preprocessing unit 1 are input, and the Depending on the size, “field structure coding” or “frame structure coding” is selected, and the image memory 3 and the video coding unit 2 are used as the coding mode selection signal 141.
And a motion vector is input as an encoding parameter 162 from the video encoding unit 2, and by comparing the magnitude with a threshold value 31, a GOP structure control signal 171 is sent to the image memory 3, and the GOP structure control signal The coding mode control signal 17 of “intra-frame coding”, “one-way predictive coding” or “bidirectional predictive coding” according to 171
2 and the image signal 15 output from the image memory 3
1 and the encoding mode in the video encoding unit 2 are controlled.

Therefore, according to the ninth embodiment, by combining the first embodiment and the fourth embodiment,
“Field structure encoding” or “frame structure encoding” can be selected according to the magnitude of the one-field absolute value difference sum 131 and the two-field absolute value difference sum 132,
GO according to correlation between image signals based on motion vector
Since the encoding can be performed instead of the image signal 151 having the P structure, it is possible to execute more efficient encoding.

Although not described any further, a combination of the first and second embodiments as shown in FIG. 9 and a combination of the fourth and fifth embodiments as shown in FIG. 10 are also possible.
In addition, a combination of the two embodiments such as a combination of the first and fourth embodiments, a combination of the second and fifth embodiments, a combination of the first and fifth embodiments, a combination of the second and fourth embodiments, Combinations of the three embodiments, such as the first, second, and third embodiments, are also possible.

The video encoding unit 2 according to the first embodiment and the like has been described using the MPEG-2 video encoding method described in the section of the prior art as an example. Any other video coding system that can switch between “frame structure coding” and “field structure coding” can be applied.

Similarly, the video encoding unit 2 according to the second embodiment or the like can be applied to any encoding method having a structure similar to the GOP structure defined by MPEG.

In the first embodiment and the like, three signal lines 102, 103, and 1 are used as a method for inputting an image signal supplied from the image memory 3 to the preprocessing unit 1 in FIG.
Although the description has been made with reference to FIG. 04, a method of transmitting three signals by time-sharing one signal line is also possible.

Similarly, in Embodiment 1 and the like, FIG.
In the above description, the sum of the absolute value difference 131 between one field and the sum of the absolute value difference 132 between the two fields are described. However, instead of the sum of the absolute value differences, the sum of squares of the absolute value differences may be obtained. Also, the description has been made assuming that the signal is obtained by using all the signals for one field, but it is also possible to obtain the signal obtained by extracting a part of the image or performing the thinning processing at an arbitrary interval. By doing so, the amount of calculation of the preprocessing unit 1 can be reduced.

[0069]

As described above, according to the present invention, "field structure coding" or "frame structure coding" is performed based on the magnitude of the sum of absolute differences between one field and the sum of absolute differences between two fields, a motion vector, and the like. Encoding, or by controlling the GOP structure of the image signal to perform encoding, it is possible to select an encoding mode having a higher correlation between the image signals, and perform highly efficient encoding. It is possible to do.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating a configuration of an image encoding device according to a first embodiment.

FIG. 2 is a diagram illustrating a front-rear relationship of image signals 102, 103, and 104;

FIG. 3 is a diagram illustrating a configuration of an image encoding device according to a second embodiment.

FIG. 4 is a diagram showing a GOP structure when α is larger than a threshold value 133.

FIG. 5 is a diagram showing a GOP structure when α is smaller than a threshold value 133.

FIG. 6 is a diagram illustrating a configuration of an image encoding device according to a fourth embodiment.

FIG. 7 is a diagram illustrating a configuration of an image encoding device according to a fifth embodiment.

FIG. 8 is a diagram illustrating a configuration of an image encoding device according to a ninth embodiment in which the first and fourth embodiments are combined.

FIG. 9 is a diagram illustrating a configuration of an image encoding device in which Embodiments 1 and 2 are combined.

FIG. 10 is a diagram illustrating a configuration of an image encoding device in which Embodiments 4 and 5 are combined.

FIG. 11 shows a graph of I / when N = 9 and M = 3.
It is a figure which shows the relationship of a P / B picture.

[Explanation of symbols]

1 pre-processing unit, 2 video encoding unit, 3 image memory,
4 control unit, 11, 12 addition processing unit, 13, 14 difference processing unit, 101 image signal, 102, 103, 104
Image signal for one field, 121, 122 difference signal, 131 Sum of absolute value difference between one field, 132 2
Sum of absolute differences between fields, 141 encoding mode selection signal, 151 image signal, 161 encoded data.

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5C059 KK11 LA05 MA00 MA04 MA05 MA23 NN01 NN28 PP05 PP06 PP07 TA17 TC12 TC36 TD12 UA02 UA31 UA33 UA38

Claims (11)

[Claims] An image encoding apparatus for encoding a digitized moving image signal, comprising: an image memory for storing a plurality of frames of the image signal; and encoding the image signal output from the image memory according to an encoding mode. A preprocessing unit that reads an image signal from the image memory to obtain a statistic between one field and a statistic between two fields; and a statistic between one field obtained by the preprocessing unit. And controlling the structure of the image signal output from the image memory based on
An image encoding device, comprising: a control unit that controls an encoding mode in the encoding unit. 2. The method according to claim 1, wherein the preprocessing unit obtains a sum of absolute differences between one field as a statistic between one field and a sum of absolute differences between two fields as a statistic between two fields. The image encoding device according to claim 1. 3. When the input image signal is an interlaced signal, the control unit compares the value of the absolute value difference between one field and the sum of absolute value differences between two fields obtained by the pre-processing unit. When the sum of the absolute differences between the two fields is smaller, the encoding is performed in units of one field, and when the sum of the absolute differences between the two fields is smaller, the encoding is performed in the unit of one frame. The image coding apparatus according to claim 2, wherein the structure is controlled and a coding mode in the coding unit is controlled. 4. The encoding unit includes intra-frame encoding, unidirectional prediction encoding, and bidirectional prediction encoding as encoding modes, and the control unit controls the absolute value of one-field absolute value obtained by the preprocessing unit. The smaller of the values is obtained from the sum of the value difference and the sum of the absolute differences between the two fields, and when the smaller value is smaller than a predetermined threshold, the frame is encoded by increasing the number of frames of the bidirectional prediction encoding. When the value is larger than the predetermined threshold, the structure of the image signal output from the image memory is controlled so that the number of frames of the bidirectional prediction coding is reduced and the coding mode in the coding unit is changed. 3. The image coding apparatus according to claim 2, wherein the control is performed. 5. An image encoding apparatus for encoding a digitized moving image signal, comprising: an image memory for storing a plurality of frames of the image signal; and encoding the image signal output from the image memory according to an encoding mode. And an encoding unit that outputs an encoding parameter used at the time of encoding, and an image signal output from the image memory based on the encoding parameter output from the encoding unit and a predetermined threshold. A control unit that controls the structure of the encoding unit and controls an encoding mode in the encoding unit. 6. When the input image signal is an interlaced signal, the encoding unit outputs a motion vector as an encoding parameter, the control unit receives the motion vector from the encoding unit,
When the average value of the magnitude of the motion vector is larger than a predetermined threshold, encoding is performed in units of one field, and when the average value of the magnitude of the motion vector is smaller than the predetermined threshold, encoding is performed in units of one frame. 6. The image coding apparatus according to claim 5, wherein the structure of the image signal output from the image memory is controlled and the coding mode in the coding unit is controlled. 7. The encoding unit includes intra-frame encoding, unidirectional prediction encoding, and bidirectional prediction encoding as encoding modes, and outputs a motion vector as an encoding parameter. Input a motion vector from the encoding unit,
When the average of the magnitude of the motion vector is smaller than a predetermined threshold, the frame is encoded by increasing the number of frames of the bidirectional prediction encoding. When the average of the magnitude of the motion vector is larger than a predetermined threshold, the frame of the bidirectional prediction encoding is 6. The image encoding apparatus according to claim 5, wherein a structure of an image signal output from an image memory is controlled and an encoding mode in said encoding unit is controlled so that encoding is performed with a reduced number. apparatus. 8. The encoding unit includes intra-frame encoding, unidirectional predictive encoding, and bidirectional predictive encoding as encoding modes, and outputs a motion vector as an encoding parameter. The sum of squares of the prediction difference signal is input from the encoding unit, and when the sum of squares of the prediction difference signal is smaller than a predetermined threshold, the frame is encoded by increasing the number of frames of bidirectional prediction coding, and the average of the magnitude of the motion vector is calculated. When is larger than a predetermined threshold, the structure of the image signal output from the image memory is controlled so that the number of frames of the bidirectional prediction coding is reduced and the coding mode in the coding unit is controlled. The image encoding apparatus according to claim 5, wherein 9. When the input image signal is an interlaced signal, the coding unit outputs a prediction type of motion compensation prediction as a coding parameter, and the control unit inputs the prediction type of motion compensation prediction from the coding unit. When the number of prediction types of field prediction is larger than a predetermined threshold, coding is performed in units of one field, and when the number of prediction types of field prediction is smaller than the predetermined threshold, coding is performed in units of one frame. 6. The image coding apparatus according to claim 5, wherein the structure of the image signal output from the image memory is controlled and the coding mode in the coding unit is controlled. 10. An image coding method for coding a digitized moving image signal, comprising: a preprocessing step of obtaining a statistic between one field and a statistic between two fields of the image signal; Controlling the structure of the image signal and controlling the coding mode based on the statistics between one field and the statistics between two fields obtained in the processing step; and controlling the structure in the control step Encoding the encoded image signal in the controlled encoding mode. 11. An image encoding method for encoding a digitized moving image signal, which encodes the image signal according to an encoding mode and outputs an encoding parameter used in the encoding. And a control step of controlling a structure of the image signal and the encoding mode in the video encoding step based on an encoding parameter output in the video encoding step and a predetermined threshold. An image coding method characterized in that: