JP2000236539A - Image compression coder - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a synthesized image by editing an image in terms of a stream. SOLUTION: The position and size of an added image are termined on the basis of a processing control signal from a terminal 22, and a type determination circuit 14 instructs a coding type on the basis of a base coding image. An image signal is coded on the basis of the coding type and an output of a quantizer 6 is fed to a grammar generating circuit 33. The grammar generating circuit 23 generates data that follow a slice layer on the basis of a quantized output and outputs the data to a grammar synthesis circuit 24. The grammar synthesis circuit 24 synthesizes an output of the quantizer 6 with the base coded image to obtain the coded image of the synthesis image. Thus, the image is synthesized in terms of a stream.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image compression encoding apparatus for synthesizing images.

[0002]

2. Description of the Related Art In recent years, as a method of compressing an image signal,
A method for encoding a moving image signal such as MPEG-1 or MPEG-2 has been developed.

FIG. 7 is a block diagram showing a configuration of a conventional image compression encoding apparatus employing MPEG. FIG.
FIG. 3 is an explanatory diagram for describing a compression-encoded signal.

In FIG. 7, an input image input to a terminal 1 is usually a line-sequential image signal or an interlaced image signal input to an image monitor. This input image is input to the scanning line conversion circuit 2. The scanning line conversion circuit 2 converts the input line sequential or interlaced image signal into horizontal 16
It is converted into a macro block composed of 16 pixels × vertical lines and output. A macro block is a unit of encoding.

[0005] An image signal in units of macroblocks is supplied to an image rearranging circuit 15 and a motion detecting circuit 3. The image rearranging circuit 15 changes the order of the image signals according to an instruction of the type determining circuit 14 described later.

In MPEG, an I (Intra-coded picture) picture is compressed by using only information in an image, and a P (Predictive) picture is compressed by referring to information of a past image signal.
B (Bidirectionally pred) that performs compression using information of -coded picture) pictures and past and future image signals.
There are three types of pictures (ictive-coded picture). The type determination circuit 14 determines an encoding type at the time of encoding, and uses the determined encoding type as an image rearranging circuit.
15. Instruct the grammar generation circuit 11 and the motion compensation image memory 10. Since the B picture requires future image information, the image rearranging circuit 15 rearranges the images in the compression encoding order.

When generating an encoded output of an I picture, the image signal from the image rearranging circuit 15 is supplied to the DCT transform circuit 5 as it is via the subtractor 4. The DCT transform circuit 5 performs a DCT (Discrete Cosine Transform) process on the input image signal in a predetermined block unit, converts the input image signal into a spatial frequency component, and outputs the spatial frequency component to the quantizer 6. The quantizer 6 quantizes the exchange frequency component to reduce the code amount, and outputs the code amount to the grammar generation circuit 11. The grammar generation circuit 11 forms a hierarchical stream according to the grammar of the MPEG standard and outputs it as an encoded output.

On the other hand, the quantized output from the quantizer 6 is
It is supplied to the inverse quantizer 7 in order to obtain a reference image when the B and P pictures are created. The inverse quantizer 7 inversely quantizes the quantized output from the quantizer 6 and outputs the result to the inverse DCT transform circuit 8. The inverse DCT transform circuit 8 performs an inverse DCT process on the inverse quantized output, and returns the data before the DCT process by the DCT transform circuit 5.

When the output of the quantizer 6 is an I-picture quantized output, the output of the inverse DCT transform circuit 8 is a restored image of the original image. In this case, the output of the inverse DCT transform circuit 8 is directly supplied to the motion compensation image memory 10 via the adder 9. The motion compensation image memory 10 stores the restored image as a reference image.

Next, it is assumed that an encoded output of a P or B picture is generated. The output of the scanning line conversion circuit 2 is also supplied to the motion detection circuit 3, and the motion detection circuit 3 searches for the motion with the reference image for each macroblock and outputs it as a motion vector. This motion vector is supplied to the motion compensation image memory 10. The motion-compensated image memory 10 outputs the motion-compensated reference image to the subtracter 4 by instructing the picture type to the type determination circuit 14 and determining the blocking position of the reference image based on the motion vector. In this case, the subtractor 4 subtracts the image signal from the image rearranging circuit 15 and the motion-compensated reference image from the motion-compensated image memory 10 to output a difference signal (prediction error).

Thereafter, DCT processing and quantization processing are performed on the prediction error. Thereby, the code amount of the B and P pictures is significantly reduced. The output of the inverse DCT transform circuit 8 for the P picture is a prediction error. The adder 9 reads the motion-compensated reference image from the motion-compensated image memory 10 and adds the prediction error to the prediction error.
Restore the original image. The adder 9 gives the restored image for the P picture as a reference image to the motion compensation image memory 10 and stores it.

FIG. 8 is an explanatory diagram showing hierarchical codes of MPEG-Video.

The first layer is a sequence layer, and a sequence header is added to the top of the layer. The sequence header describes a bit rate (bit_rate) at the time of performing the compression encoding, an image size, and the like. Multiple GOs after the sequence header
The P-layer follows, and again the next sequence starts. Therefore, the bit rate and the image size cannot be changed within one sequence.

The GOP layer includes a GOP header and a plurality of picture layers. GOP
Is a unit of random access of a compression-encoded signal, and includes at least one I picture. The GOP header describes a time code, a flag called Closed_GOP indicating whether or not the first image of the GOP has been compressed using information of an image signal included in the previous GOP, and the like.

The picture layer is composed of a picture header and a plurality of slice layers. The picture header is
The coding type (I / P / B) of the picture, the accuracy of the motion vector used in the picture, and the like are described.

The slice layer includes a slice header and a plurality of macroblock layers. The slice header has a unique code called a slice start code, by which the vertical position of the slice (vertical address of the encoded image) can be known. There is also a flag indicating the quantization scale used in this slice.

The quantization scale indicates the quantization characteristics of the quantizer 6 shown in FIG. 7, and must be described at least at the beginning of a slice. Here, as a constraint in forming a slice, overlapping of slices is not allowed, and a slice is usually encoded as a set of macroblocks arranged in a horizontal direction and extends over a plurality of macroblock lines. There is no such thing (it may cross over in the standard).

A macro block is composed of a macro block header and a block layer. The macroblock header describes the coding mode of the macroblock, the horizontal address of the macroblock, and the motion vector of the macroblock.

A macroblock can be skipped (not coded) within a slice, but the first and last macroblocks of a slice cannot be skipped. Therefore,
Thus, the horizontal address of the macroblock at the head of the slice, that is, the horizontal position of the slice can be known.
The block layer is configured by a code obtained by coding DCT coefficients, but a code indicating a difference between DC components of DCT may be added depending on a coding mode.

The grammar generation circuit 11 outputs such an encoded output of the MPEG standard via an output terminal 12,
It is also output to the code amount control circuit 13. The code amount control circuit 13
The generated code amount is monitored, and for example, a feedback signal for smoothing the code amount generated during a certain period is output to the quantizer 6. By determining the quantization coefficient of the quantizer 6 based on the output of the code amount control circuit 13, the code amount is limited to the set code amount.

By the way, in digital broadcasting which has recently started broadcasting, a compression code by MPEG is transmitted, and the number of broadcasting channels is several hundred in some cases. Also, H
The recording capacity of data storage media, such as a DD (hard disk drive) and an optical disk, has increased dramatically. In view of such a situation, digital broadcasting requires a channel for a program selection guide to help a general audience select a program, and a storage medium manages recorded image signals. Therefore, for example, an additional image for displaying a thumbnail is required.

For example, in a program selection guide channel, a screen is divided into a plurality of sub-screens,
There is a demand for a function that can reduce the image of each channel and simultaneously refer to a plurality of broadcast channels. In addition, since the additional image of the compressed image storage medium is similarly converted into a small screen and recorded, a large number of recorded images can be easily grasped by the additional image.

The MPEG system is generally used as an encoding / decoding device for digital broadcasting or a large-capacity compressed image storage medium. In the above-described conventional image compression / encoding device, one type of input image is used. Or it is necessary to encode a plurality of broadcast channel image signals after reducing and synthesizing them in advance.

Therefore, for example, in the case of digital broadcasting,
If the program to be broadcast at the same time is changed by one program, or if the broadcasting of a certain channel is interrupted, it is necessary to perform the reduction synthesis of all the broadcasting channels again and then encode. In addition, in a storage medium, in order to create an additional image from a recorded compressed image, it is necessary to decode all the images once, perform reduction synthesis, and then encode again.

Note that since a GOP always includes one I picture, a frame coded image can be edited in the time domain by using a GOP as a unit. However, as described above, in order to edit a coded image in a two-dimensional area, it is necessary to restore the image once.

[0026]

As described above, in the above-described conventional image compression encoding apparatus, a part or configuration of an encoded image obtained by encoding a reduced composite image based on a plurality of images is changed. In such a case, there is a problem in that the reduced composite image must be encoded after being recreated by the reduced composite processing. Also, in order to create and encode a reduced composite image based on an encoded image that has already been encoded, it is necessary to decode the already encoded image to obtain a reduced composite image, and then perform encoding. Has to be performed.

The present invention has been made in view of such a problem, and has as its object to provide an image compression encoding apparatus capable of editing an encoded image on a stream.

It is another object of the present invention to provide an image compression encoding apparatus capable of creating an encoded image of a reduced composite image based on a plurality of original images from a plurality of encoded images. And

[0029]

An image compression encoding apparatus according to the present invention is provided with an input image corresponding to a partial area of a screen, compression-encodes the input image and outputs an encoded image. Encoding means, encoding control means for controlling the encoding means to complete encoding of the encoded image in a predetermined area unit of a screen, and a base image corresponding to the entire area of the screen and the input image. Synthesizing means for synthesizing an encoded image based on the input image with a base encoded image, which is encoded data of the base image, in units of a predetermined area of the screen to obtain an encoded image of the synthesized image. It is provided.

In the present invention, the input image to be synthesized with the base image is supplied to the encoding means and encoded. The encoding control means completes the encoding of the input image in a predetermined area unit of the screen. The synthesizing unit synthesizes the encoded image of the input image with the encoded base image in a predetermined area unit of the screen.

[0031]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of an image compression encoding apparatus according to the present invention. In FIG. 1, the same components as those in FIG. 7 are denoted by the same reference numerals.

In the present embodiment, editing is performed in units of slices starting at an arbitrary horizontal position in one macroblock line. The encoded data is encoded using images that are temporally and spatially shifted. In order to perform editing on a stream, the stream must be completed in edit units. In the present embodiment, the head macroblock of the slice is configured by an intra macroblock (a macroblock without a reference image) by utilizing the fact that an address is added at the head of the slice and the DC component and the like are also initialized. Thus, encoding is completed in slice units, and editing in slice units is enabled.

In the present embodiment, the present invention is applied to an image editing apparatus, in which an input image is encoded and then edited on a stream. It should be noted that a coded image that has already been coded may be input, and editing work may be performed on this coded image.

An input image is input to the input terminal 1. This input image is supplied to the image processing circuit 21. The image processing circuit 21 performs predetermined processing on the input image signal based on the processing control signal input via the terminal 22 and outputs the processed signal to the scanning line conversion circuit 2. The processing control signal gives the coding type, the position and area ratio of the image to be added on the screen, and the like. For example, when the processing control signal instructs to encode the input image as it is, the image processing circuit 21
Outputs the input image signal to the scanning line conversion circuit 2 as it is.

If the processing control signal indicates that a reduced image based on the input image signal is to be added to a part of the image before editing (hereinafter referred to as a base image), Scanning line conversion circuit 2 after performing reduction processing
Output to Note that, in this case, the image processing circuit 21
The image size is reduced to the image size specified by the processing control signal. The image processing circuit 21 can also output a child image obtained by extracting only a part of the image. In this case, for example,
Pan _ & _ S defined in ISO / IEC13818
Cut out at the position of can.

The scanning line conversion circuit 2 converts the input line-sequential or interlaced image signal into a macroblock composed of 16 horizontal pixels × 16 vertical lines and outputs the macroblock to the image rearranging circuit 15 and the motion detecting circuit 3. . The motion detection circuit 3 detects the motion of the input image and outputs a motion vector to the motion compensation image memory 10 and the grammar generation circuit 23.

The image rearranging circuit 15 includes a type determining circuit 14
, The encoding type is instructed, the images are rearranged and output to the subtractor 4. When generating encoded images of P and B pictures, the subtracter 4 subtracts the image from the image rearranging circuit 15 and the motion-compensated reference image from the motion-compensated image memory 10 to output a prediction error. However, when generating an encoded picture of an I picture, the picture from the picture rearranging circuit 15 is output as it is.

The output of the subtractor 4 is supplied to a DCT conversion circuit 5. The DCT conversion circuit 5 converts the input image signal into D
CT processing is performed and output to the quantizer 6. The quantizer 6 is supplied with a feedback signal for determining a quantization coefficient from the code amount control circuit 13, quantizes the input DCT output, and sends the quantized output to the grammar generation circuit 23 and the inverse quantizer 7. Output.

The inverse quantizer 7 obtains a reference image by:
The quantized output is inversely quantized and output to the inverse DCT transform circuit 8. The inverse DCT transform circuit 8 performs an inverse DCT process on the input inverse quantized output, restores the image before the DCT process, and outputs the restored image to the adder 9.

The adder 9 outputs the restored image based on the I picture to the motion compensated image memory 10 as it is.
For the prediction error based on the picture, the original image is restored by adding to the motion-compensated reference image from the motion-compensated image memory 10 and output to the motion-compensated image memory 10. The motion-compensated image memory 10 holds the input restored image as a reference image, and instructs the coding type from the type determination circuit 14 and subtracts the reference image motion-compensated based on the motion vector into the subtracter 4 and the adder Output.

The type determination circuit 14 converts the coding type corresponding to the base coded image supplied from the base coded image supply circuit 25 into an image rearrangement circuit 15, a grammar generation circuit 11, and a motion compensation image memory. Instruct 10

The grammar generation circuit 23 converts the quantized output from the quantizer 6 into a code array corresponding to the format of the coded data to be output, and outputs the coded data. For example, the grammar generation circuit 23 performs the hierarchical encoding shown in FIG. In the present embodiment, the grammar generating circuit 23 generates only codes below the slice layer in FIG.

The encoded output from the grammar generation circuit 23 is supplied to the code amount control circuit 13 and the grammar synthesis circuit 24. The code amount control circuit 13 outputs a feedback signal for smoothing the generated code amount to the quantizer 6. In the present embodiment, when a reduced image is added to a base image, an output of a base coded image supply circuit 25 described later is given to the code amount control circuit 13 and added to the entire screen of the base image. The area ratio of the reduced image can be grasped. The code amount control circuit 13 outputs a feedback signal for controlling the code amount of the quantized output generated based on the area ratio.

The base coded image supply circuit 25 outputs a coded output of a base image (hereinafter referred to as a base coded image) based on the processing control signal from the terminal 22. The base coded image supply circuit 25 receives the output of the grammar synthesizing circuit 24, for example, and outputs a base coded image. The base coded image from the base coded image supply circuit 25 is supplied to the type determination circuit 14, code amount control circuit 13, grammar generation circuit 23, and grammar synthesis circuit 24.

The base coded image supply circuit 25 outputs a base coded image in which the boundary of each area matches the boundary of the slice layer. Even if the processing control signal indicates a region that does not match the boundary of the slice layer as the region into which the image to be added is fitted, the base coded image supply circuit 25 matches the boundary of the slice layer as the region into which the image to be added is fitted. An area having a boundary is set. Furthermore, if the region into which the image to be added is to be included by the processing control signal includes the region where the image already exists, the base coded image supply circuit 25
An area that is a non-image part including the area specified by the processing control signal is searched, and an area into which an image to be added is set is set based on the search result.

The grammar synthesizing circuit 24 synthesizes the output of the grammar generating circuit 23 and the base coded image from the base coded image supply circuit 25 in a slice layer, thereby forming a reduced image based on the base image and the input image signal. Is synthesized on the stream and output. The output of the grammar synthesizing circuit 24 is output through the output terminal 12 and is also supplied to the base coded image supply circuit 25.

Next, the operation of the embodiment configured as described above will be described with reference to FIGS. 2 and 3. FIG. FIG. 2 shows a display example of a composite image, and FIG. 3 shows the structure of a slice.

It is assumed that the base coded image output from the base coded image supply circuit 25 is obtained by coding the base image including the reduced images A, B, and C in FIG. That is, the base images are (X0, Y0) to (X3,
Y1), a reduced image A is arranged in an area (hereinafter, referred to as an A area), and a reduced image B is arranged in an area (X0, Y1) to (X1, Y2) of the screen (hereinafter, referred to as a B area). The reduced image C is arranged in the region (X2, Y1) to (X4, Y2) (hereinafter, referred to as C region), and the other region is a non-image portion.

Now, in the upper right portion of the base image, on the screen, an area of (X3, Y0) to (X5, Y1) (hereinafter, referred to as "X5, Y1").
It is assumed that an encoded image to which the reduced image D shown in FIG. 2 is added at a position corresponding to the D region). Input terminal 1
Is input with an input image which is a source of the reduced image D. The image processing circuit 21 is supplied with a processing control signal via a terminal 22 to designate a screen size, reduces the input image signal to the designated screen size, and outputs the reduced image signal to the scanning line conversion circuit 2. Note that the size of the reduced image is the macro block (16
It is set to an integral multiple of (line × 16 dots).

When a value which is not an integral multiple of the macroblock is specified as the size of the reduced image to be added by the processing control signal, the image reducing circuit 21 changes this value to an integral multiple of the macroblock. Process.

The scanning line conversion circuit 2 converts the input reduced image signal into a signal in units of macro blocks and outputs the signal to the image rearranging circuit 15. The images are rearranged by the image rearranging circuit 15. In this case, the type decision circuit
14 controls the image rearranging circuit 15 according to the coding type of the base coded image specified by the processing control signal so that the image added by the same coding type as the base coded image is compressed. Rearrange the images.

When generating an I picture, the subtractor 4
Outputs the output of the image rearrangement circuit 15 to the DCT conversion circuit 5 as it is. The DCT conversion circuit 5 performs DCT processing on the input image signal in units of macroblocks and outputs it to the quantizer 6. The DCT transform output is quantized by the quantizer 6 and supplied to the grammar generating circuit 23.

The output of the quantizer 6 is also output to the inverse quantizer 7 to obtain a reference image. The image signal that has been inversely quantized by the inverse quantizer 7 and subjected to inverse DCT processing by the inverse DCT transform circuit 8 is supplied to the motion compensation image memory 10 via the adder 9. On the other hand, the motion detection circuit 3 detects the motion of the input image signal and outputs a motion vector to the motion-compensated image memory 10.
The reference image is motion-compensated based on the motion vector, and the motion-compensated reference image is output to the subtractor 4 and the adder 9.

Thus, the output of the image rearranging circuit 15 and the motion-compensated reference image are subtracted by the subtractor 4 to obtain a prediction error. When generating P and B pictures,
DCT processing and quantization processing are performed on the prediction error from the subtractor 4.

Since the area ratio of the image area added to the entire screen is known, the code amount control circuit 13 controls the code amount according to this ratio. For example, if the bit rate of the base coded image is 4 Mbps and the added image area ratio is 1/4, the code amount control circuit 13
Controls the code amount of the added image to be 1 Mbps.

Next, the relationship between the bit rate of the base code and the code amount of an individually compressed image will be described. In the conventional example, since the code amount is controlled for the entire screen, the code amount matches the bit rate indicated by the sequence layer. By the way, in MPEG, the relationship between the amount of code generated instantaneously and the bit rate varies. That is,
In MPEG, a change in the code amount for each image to be compressed is allowed, and vbv_bu is used to absorb the change.
The ffer model is defined.

That is, in the MPEG, in consideration of the fact that the generated code amount of the encoded data differs depending on the picture and the picture type, the decoder side decodes each picture in real time. Has a vbv buffer for absorbing the
Assuming a v-buffer, the generated code amount is controlled so that the vbv buffer does not underflow or overflow.

When the transmission path is a fixed bit rate (hereinafter, CBR)
(Referred to as “Constant Bit Rate”), the code amount is allowed to vary within a range in which the vbv buffer does not underflow or overflow.

On the other hand, when the transmission rate is a variable bit rate (hereinafter, referred to as VBR (Variable Bit Rate)), the code generation is performed at the bit rate indicated by the sequence layer in FIG. 8 assuming vbv_buffer. Done in In this case, when the generated code amount is small, vbv_buffer is in a state where data is always stored to the full capacity, and the decoding decoder uses vbv_buffer for each picture layer as a decoding unit.
Read the encoded data from the buffer. In VBR encoding, the amount of generated code for each image does not need to be controlled as strictly as CBR encoding, and the amount of code in the case of performing compression encoding of one image only needs to be larger than the vbv_buffer size.

In this embodiment, the code amount is controlled for each image to be synthesized. Specifically, the bit rate of the base coded image is 4 Mbps, and the imposed image area ratio is 1
If the reduced image to be added is encoded at 1 Mbps in the case of / 4, the amount of code is controlled using a model in which the vbv_buffer size is also increased by 倍 according to the ratio.

Accordingly, the code amount of the entire base code is close to the value indicated by the bit rate. However, since the code amount of each reduced image is controlled independently, the CBR
Does not completely match the model of coding. Therefore, in the present embodiment, the bit rate of the VBR code model is described as the bit rate described in the sequence layer.

As described above, in the present embodiment, V
Since the BR encoding model is used, it is not necessary to consider the shortage of the code amount corresponding to the non-image area of the base image.

In this embodiment, the type determining circuit
According to 14, the type is determined so that the leading block of the slice becomes an I picture, and the DC component and the like are initialized in the leading block of the slice. The grammar generation circuit 23 generates, based on the output of the quantizer 6, data below the slice layer in the range of the D region in FIG.

Thus, in the present embodiment, the code is completed in units of slice layers. The slice layer data of the grammar generation circuit 23 is supplied to a grammar synthesis circuit 24.

The grammatical synthesizing circuit 24 is supplied with the base coded image from the base coded image supply circuit 25, and the grammatical synthesizing circuit 24 generates a slice of a slice layer corresponding to the D region in FIG. The data below the slice layer from the grammar generation circuit 23 are arranged as a stream.

FIG. 3 shows the configuration of a slice layer in a base coded image by horizontal and vertical lines that divide a screen. In FIG. 3, the number of slice layers in the vertical direction is different from the actual example. As shown in FIG. 3, in the slice layer of the base image, a boundary is formed at least at the boundary of each region.

The grammar synthesizing circuit 24 arranges the output of the grammar generating circuit 23 as data below the slice layer in the D area in FIG. Since the data of the slice layer and below from the grammar generation circuit 23 is supplied with the address of the reduced image to be synthesized, the grammar synthesis circuit 24 determines the vertical and horizontal addresses of the slice at the position where the reduced image to be synthesized is fitted. (The position of the D area). Also, for layers above the picture layer, there is no need to convert the information such as the coding type so that the base coded image and the compressed code of the reduced image match. In addition, when the display mode or the like changes before and after editing, each value of the corresponding stream is changed.

In FIG. 3, one slice layer is composed of one macro block so that a reduced image can be inserted at an arbitrary position and size in an area corresponding to a non-image part of the base image. The encoding of the base image is performed as described above. As a result, the layout can be freely set. Even when the free area of the base image does not constitute one slice layer with one macroblock, by adding data constituting one slice layer with one macroblock instead of the reduced image to be added. The layout can be set freely.

In this way, the grammar synthesizing circuit 24 synthesizes the base coded image and the coded output of the reduced image D on the stream and outputs them to the output terminal 12. Note that the grammar synthesis circuit 24
Is supplied to the base coded image supply circuit 25, so that the image shown in FIG. 2 can be used as a base image for further editing on a stream. The grammar synthesis circuit 24
Indicates information indicating an area where an image is synthesized and an area where an image is not synthesized, according to ISO / IEC 11172 or 13
818 may be added to the user data (user_data) area of the sequence layer, GOP layer or picture layer.

As described above, in the present embodiment, the encoding process is completed in units of slice layers, and
By combining the encoded output of the created slice layer and the base encoded image in slice layer units,
It allows editing on the stream. Accordingly, when creating a coded image of a composite image using a plurality of reduced images, even when a composite image including a plurality of reduced images is not supplied, or when an image serving as a source of the plurality of reduced images is not supplied at one time However, by combining the encoded images of the reduced images, an encoded image of the combined image can be obtained.

Although the above embodiment has been described assuming that the area D is fixed as the area into which the reduced image is to be inserted, the position on the screen of the image to be added can be moved temporally.

It is also possible to change the image size of the area where the reduced image is to be fitted. In this case, the size can be changed in GOP units shown in FIG. Further, in this case, clos which does not use the information of the previous GOP is used.
Compression encoding is performed with ed_GOP. Accordingly, the motion vector can be always detected between images of the same size in the motion detection circuit 3 that performs bidirectional or forward prediction. The area where no image is fitted is a non-image area, but it is also possible to fit an image signal of a fixed pattern stored in advance in this area.

In the above embodiment, one macroblock is defined as one slice in an area where no image is fitted (non-image part). However, a plurality of macroblocks may be defined as one slice. . Then,
The horizontal size of the image to be synthesized is 1
Limited to integer multiples of a slice, but one macroblock = 1
When a slice is formed, the number of slice headers required for each macroblock is reduced, and the code amount can be reduced.

FIGS. 4 and 5 are explanatory diagrams for explaining another embodiment of the present invention. FIG. 4 shows an example of synthesizing a screen, and FIG. 5 shows an example of the configuration of a slice. The circuit configuration of the present embodiment is the same as FIG.

In the present embodiment, the newly synthesized image D
Shows an example in a case where the screen size of the area where the reduced image D is fitted in the non-image portion of the base image is larger.

In this case, the grammar synthesizing circuit 24 changes the image size of the base coded image in the sequence layer shown in FIG. 8, and then synthesizes an additional screen. I have. Also, at this time, the sequence layer also changes the bit rate by the ratio before and after the change, so that when the image size is greatly changed, the correlation between the bit rate indicated by the sequence layer and the actual code amount is obtained. Will be higher.

FIG. 6 is a block diagram showing another embodiment of the present invention. 6, the same components as those in FIG. 1 are denoted by the same reference numerals, and description thereof will be omitted.

This embodiment is different from the embodiment of FIG. 1 in that a type change circuit 31 is added. Type change circuit 31
Changes the encoding type of the base encoded image supplied from the base encoded image supply circuit 25, and
And a grammar generating circuit 23.

In the embodiment configured as described above, only the method of determining the coding type of the reduced image to be added is different from the embodiment in FIG. According to the MPEG standard, as described above, there are three types of codes: a coding type using bidirectional prediction (B picture), a coding type using forward prediction (P picture), and an intra-frame coding type (I picture). There are different types.

For a B picture, an encoding mode using bidirectional prediction, an encoding mode using forward prediction, and an encoding mode using intra-frame prediction can be appropriately selected for each macroblock depending on the image content. In the case of a P picture, an encoding mode using forward prediction and an encoding mode using intra-frame prediction can be appropriately selected according to image content. In the case of an I picture, all macroblocks are encoded in an encoding mode using intra-frame prediction.

Therefore, for example, a predetermined area in an image composed of B pictures may be coded by a coding mode using forward prediction and a coding mode using intra-frame prediction. In some cases, a predetermined area in an image composed of P pictures is coded only in a coding mode using intra-frame prediction.

In this embodiment, the type change circuit
When the base coded image supplied from the base coded image supply circuit 25 is a B picture, the reference 31 designates a P or I picture as the picture type to the type determination circuit 14 and the grammar generation circuit 23. Also, the type change circuit 31
If the base coded image supplied from the base coded image supply circuit 25 is a P picture, the I picture is instructed to the type determination circuit 14 and the grammar generation circuit 23.

The type determining circuit 14 and the grammar generating circuit 23
Encoding is performed according to the encoding type specified by the type changing circuit 31.

When a change that is difficult to predict and encode occurs in an image such as a scene change, encoding the scene change as an I or P picture may increase the quality of the compressed image. Are known. On the other hand, the coding type of the base coded image is fixed irrespective of the change of the image of the newly added image. Therefore, without changing the coding type of the base coded image as in the present embodiment, the compression of the synthesized screen is performed by coding using forward prediction or coding using intra-frame prediction. Improves image quality.

As described above, in the present embodiment, the same effects as those of the embodiment of FIG. 1 can be obtained, and the quality of the compressed image can be improved.

The present invention is not limited to the above embodiments, and various modifications are possible. For example,
It is conceivable that a region where an image has already been synthesized is designated as a region to which a new image is to be added by the processing control signal. In this case, for example, for a region where an image has already been combined, the screen position of the combined image is changed by changing the horizontal and vertical addresses in the screen, and then input to the region specified by the processing control signal. Encoding and synthesizing processing may be performed so as to add the reduced image.

[0087]

As described above, according to the present invention, an encoded image can be edited on a stream, and a reduced composite image based on a plurality of original images from a plurality of encoded images. Has the effect of being able to create an encoded image of.

[Brief description of the drawings]

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

FIG. 2 is an explanatory diagram for explaining the embodiment in FIG. 1;

FIG. 3 is an explanatory diagram for explaining the embodiment in FIG. 1;

FIG. 4 is an explanatory diagram for explaining another embodiment of the present invention.

FIG. 5 is an explanatory diagram for explaining another embodiment of the present invention.

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

FIG. 7 is a block diagram showing a conventional image compression encoding apparatus.

FIG. 8 is an explanatory diagram showing a stream configuration of the MPEG standard.

[Description of Signs] 14 ... Type determination circuit, 21 ... Image reduction circuit, 23 ... Grammar generation circuit, 24 ... Grammar synthesis circuit, 25 ... Base coded image supply circuit

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Minoru Okita, Innovator Toshiba Multimedia Technology Research Institute, 8-8 Shinsugita-cho, Isogo-ku, Yokohama-shi, Kanagawa Prefecture (72) Inventor Kazuhiko Yamauchi 3-3-1 Shimbashi, Minato-ku, Tokyo No. 9 Toshiba Abu E Co., Ltd. F term (reference) 5C059 KK37 KK39 LA06 LA09 MA00 MA04 MA05 MA14 MA23 MA31 MA32 MB21 MC11 NN01 PP05 PP06 PP07 RB09 RB13 TA60 TB07 TB11 TB17 TC01 TC25 TC27 TC37 TD06 UA02 UA32

Claims (7)

[Claims] An input unit that receives an input image corresponding to a partial area of a screen, compresses and codes the input image, and outputs a coded image; Encoding control means for completing encoding of the encoded image in units of a predetermined area, and an encoded image of a synthesized image of the input image and a base image corresponding to the entire area of the screen, An image compression encoding apparatus comprising: a synthesizing unit that synthesizes an encoded image based on the input image with a base encoded image, which is encoded data, in units of a predetermined area of the screen and outputs the combined image. 2. A coding type determining means for determining a coding type based on a coding type of a base coded image which is coded data of a base image corresponding to an entire area of a screen, and a predetermined area in the base image. Is specified as a region of an additional image in units of macroblocks, and an input image of an image size corresponding to the region of the additional image is compression-encoded by the encoding type determined by the encoding type determining means and slice layers and below. Encoding means for outputting an encoded image of the above; combining the encoded image from the encoding means with the base encoded image in units of the slice layer, and encoding a composite image of the base image and the input image And a synthesizing means for obtaining a coded image. 3. The image compression coding method according to claim 2, wherein said combining means adds data indicating that said base coded image is variable bit rate coded to said slice layer. apparatus. 4. The encoding device according to claim 2, wherein the encoding unit controls a code amount of the encoded image based on the input image according to a ratio of an image size of the base image to the input image. 3. The image compression encoding apparatus according to claim 1. 5. When the area specified as the area of the additional image includes an area outside the area of the base image, the base of the base image is reset to include the area outside the area. 3. The image compression encoding apparatus according to claim 2, wherein an encoded image is supplied to said combining means. 6. When the encoding type of the base encoded image is an encoding type based on bidirectional predictive encoding, the encoding type determining means includes not only an encoding type based on bidirectional predictive encoding but also an encoding type based on bidirectional predictive encoding. The encoding type according to the one-way predictive encoding and the encoding type according to the intra-frame encoding are also determined. If the encoding type of the base encoded image is the encoding type according to the unidirectional predictive encoding, The image compression encoding apparatus according to claim 2, wherein not only the encoding type by predictive encoding but also the encoding type by intra-frame encoding is determined. 7. The base coded image corresponding to a non-image area of the base image completes coding by one macroblock, and the synthesizing unit includes a base macro image of the area where the input image is synthesized. 3. The apparatus according to claim 2, wherein a slice header is added to a next macroblock.