JP2000134622A - Data processor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable parallel processing independent for each small image in data processing based on area division by dividing source image information into plural blocks and compression-encoding the respective blocks at a compression processing part at any arbitrary bit rate. SOLUTION: A data processor 101 inputs image data, divides them into small blocks, independently performs compression processing to the respective blocks and outputs the compressed data of different compressibility (bit rates) in a region of interest(ROI) and a non-interested region in the image. A source image is divided into plural small blocks by an arithmetic unit 102, and the respective blocks are respectively independently processed. Based on data indicating the ROI (ROI information), the arithmetic unit 103 judges whether each block includes the ROI or not and classifies the blocks into the group of blocks including the ROI (interested block group) and the group of blocks including no ROI (non-interested block group) and concerning the interested block group, reversible compression is performed but concerning the non-interested block group, reversible or irreversible compression is performed.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data processing apparatus for compressing image data.

[0002]

2. Description of the Related Art With the development of networks and the development of multimedia applications, the need to handle data having a large amount of information, such as image data, has been increasing, and image data is effective because it has a large amount of information itself. Compression is needed.

[0003] Conventional image information compression techniques include DC
Examples include JPEG and MPEG compressed using T. DCT is technically saturated, and in recent years, instead of DCT, compression by region division has been attracting attention as a method capable of achieving a high compression ratio.

[0004]

However, according to the conventional compression method, for example, reference (1) (“Medical image
Ge compression with lossl
ess region of interest ", E
LSEVIER, Signal Processin
g 59, 1997, p. 155-171), a region of interest (ROI: Region Of) in the image.
Interest) and the non-interest area are not sufficiently considered (in the above-mentioned document, the ROI is defined only as a set of rectangles and circles).
It has been practically difficult to separately encode independently of OI and ROI.

[0005] Further, for example, in Reference (2) (Japanese Patent Application No.
Compression and decompression by area division such as 210264 ″) have poor seamlessness in the spatial direction, and the ROI and non-RO
Since the resolution (or bit rate) with I was set to a different resolution and bits could not be preferentially allocated to the ROI, the compression of the image including the ROI was insufficient in terms of the compression ratio.

[0006]

SUMMARY OF THE INVENTION The present invention provides an ROI in an image.
In order to easily describe a region other than the ROI and the ROI, the image is divided into some small images (blocks) and classified into those belonging to the ROI and those not belonging to the ROI. In addition, in the compression encoding by region division, processing is performed independently for each block, thereby making it possible to have seamlessness in the spatial direction.

Further, the compression ratio is improved by utilizing the correlation between the blocks, predicting the luminance value information of the segment obtained by region division, or using a special segment designation symbol in segment coding at the time of region division. It is intended.

More specifically, the data processing apparatus according to claim 1 comprises a plurality of pixels, each of which is a natural number k.
What is claimed is: 1. A data processing apparatus for compressing original image information associated with (k ≧ 2) -bit luminance values to generate compressed encoded data, comprising: an input unit, a data processing unit, and an output unit; The unit receives the original image information, the data processing unit divides the original image information into a plurality of blocks, and compresses and encodes each of the blocks by a compression processing unit at an arbitrary bit rate. The unit outputs the compression-encoded data that has been compression-encoded by the data processing unit.

According to a second aspect of the present invention, in the data processing apparatus,
Original image information and attention area information indicating an attention area of the original image information are input, the data processing unit divides the original image information into a plurality of blocks, and divides each block into a block group including the attention area information. (To be referred to as a block of interest) and a block that does not include the region of interest information (to be referred to as a block of non-target), and each of the blocks is sent to the compression processing unit at an arbitrary bit rate. Compression encoding.

According to a third aspect of the present invention, in the data processing unit, the data processing unit losslessly compresses and encodes the target block group by a compression processing unit and losslessly or irreversibly compresses the non-target block group by a compression processing unit. The output unit outputs the compressed coded data of the block group of interest before the compressed coded data of the block group of non-target block.

According to a fourth aspect of the present invention, the data processing apparatus has a first compression processing section and a second compression processing section as compression processing sections, and the first compression processing section is one or more of the original image information. A compression process is performed on a first plane group including a plurality of bit planes by a first region division method, the first region division information indicating a result of the region division, and a segment corresponding to each segment obtained as a result of the region division. First compression information including first luminance information indicating a luminance value is created, and the second compression processing unit belongs to the highest bit plane lower than the (n-1) th plane group in the original image information. The n-th plane group consisting of one or a plurality of continuous bit planes is targeted, and the second area division method is used based on the area division result up to the (n-1) -th plane group in the original image information. Compression processing In the process of creating n-th compression information including n-th region division information indicating a result of region division and n-th luminance information indicating a luminance value corresponding to each segment obtained as a result of the region division, n is defined as n = 2 to an arbitrary natural number j not exceeding k.

According to a fifth aspect of the present invention, the data processing apparatus is characterized in that the first area division method and the second area division method are quad-tree divisions.

According to a sixth aspect of the present invention, in the data processing apparatus, the data processing unit has a context determining unit, and the context determining unit is configured to encode one or more blocks around the divided block when encoding each block. The state of the block is classified based on the distribution of luminance values inside the block, and the data processing unit performs the compression encoding for each state classified by the context determination unit.

According to a seventh aspect of the present invention, in the data processing device, the data processing unit has a context determining unit, and the context determining unit is configured to encode one or more blocks around the divided block when encoding each block. The state of the block is classified based on the distribution of luminance values inside the block, and the data processing unit performs the compression encoding for each state classified by the context determination unit.

In the data processing device according to the present invention, the data processing unit has a context determining unit, and the context determining unit is configured to encode one or a plurality of blocks around the divided block when encoding each of the divided blocks. The state of the block is classified based on the distribution of luminance values inside the block, and the data processing unit performs the compression encoding for each state classified by the context determination unit.

According to a ninth aspect of the present invention, in the data processing device, the data processing unit has a context determining unit, and the context determining unit is configured to encode one or more of the divided blocks when encoding each of the divided blocks. The state of the block is classified based on the distribution of luminance values inside the block, and the data processing unit performs the compression encoding for each state classified by the context determination unit.

According to a tenth aspect of the present invention, in the data processing device, the data processing unit has a context determination unit, and the context determination unit is configured to encode one or more blocks around the divided block when encoding each block. The state of the block is classified based on the distribution of luminance values inside the block, and the data processing unit performs the compression encoding for each state classified by the context determination unit.

In the data processing apparatus according to the present invention, when the context determination unit encodes each divided block, one or more blocks around the block on the same bit plane as the block are encoded. And classifying the state of the block on the basis of the distribution of luminance values inside a block (referred to as an upper block) at the same position as the block on the bit plane higher than the block, and performing the data processing. The unit performs compression encoding for each state classified by the context determination unit.

In a twelfth aspect of the present invention, in the data processing apparatus,
A compression processing unit having a first segment luminance value prediction unit, the second compression processing unit having a second segment luminance value prediction unit,
The first segment luminance value prediction unit is configured to calculate one or a plurality of segments adjacent to each segment (referred to as an encoded segment) obtained as a result of the region division.
A luminance prediction value is generated by a predetermined method based on the size and the luminance value, and a prediction error between the luminance value of the coded segment and the coded luminance prediction value is coded by a predetermined method.
As the luminance information, the second segment luminance value prediction unit generates a luminance prediction value by a predetermined method based on the size and the luminance value of one or a plurality of segments adjacent to the encoded segment, and And a prediction error between the luminance value and the encoded luminance prediction value is encoded by a predetermined method, and is used as the n-th luminance information.

According to a thirteenth aspect of the present invention, in the data processing device, the first
A compression processing unit having a first segment luminance value prediction unit, the second compression processing unit having a second segment luminance value prediction unit,
The first segment luminance value prediction unit is configured to calculate one or a plurality of segments adjacent to each segment (referred to as an encoded segment) obtained as a result of the region division.
A luminance prediction value is generated by a predetermined method based on the size and the luminance value, and a prediction error between the luminance value of the coded segment and the coded luminance prediction value is coded by a predetermined method.
As the luminance information, the second segment luminance value prediction unit generates a luminance prediction value by a predetermined method based on the size and the luminance value of one or a plurality of segments adjacent to the encoded segment, and And a prediction error between the luminance value and the encoded luminance prediction value is encoded by a predetermined method, and is used as the n-th luminance information.

According to a fourteenth aspect of the present invention, in the data processing apparatus, the first
A compression processing unit having a first segment luminance value prediction unit, the second compression processing unit having a second segment luminance value prediction unit,
The first segment luminance value prediction unit is configured to calculate one or a plurality of segments adjacent to each segment (referred to as an encoded segment) obtained as a result of the region division.
A luminance prediction value is generated by a predetermined method based on the size and the luminance value, and a prediction error between the luminance value of the coded segment and the coded luminance prediction value is coded by a predetermined method.
As the luminance information, the second segment luminance value prediction unit generates a luminance prediction value by a predetermined method based on the size and the luminance value of one or a plurality of segments adjacent to the encoded segment, and And a prediction error between the luminance value and the encoded luminance prediction value is encoded by a predetermined method, and is used as the n-th luminance information.

According to a fifteenth aspect of the present invention, there is provided the data processing apparatus according to the first aspect.
4, the first segment luminance value prediction unit and the second segment luminance value prediction unit determine the state of the block classified by the context determination unit when generating the luminance prediction value. , One or a plurality of segments adjacent to each segment (referred to as an encoded segment) obtained as a result of the region division,
The method is characterized in that a luminance prediction value is generated by a predetermined method based on the size and the luminance value.

According to a sixteenth aspect of the present invention, there is provided a data processing apparatus for compressing original image information comprising a plurality of pixels and each of the pixels being associated with a luminance value of a natural number k (k ≧ 2) bits to generate compressed encoded data. A data processing device, comprising: an input unit, a data processing unit, and an output unit, wherein the input unit receives the original image information, and the data processing unit includes a first compression processing unit and a second compression processing unit as a compression processing unit. A second compression processing unit, wherein the first compression processing unit performs compression processing on a first plane group including one or a plurality of continuous bit planes in the original image information by a first area division method. To generate first compression information including first region division information indicating a result of the region division and first luminance information indicating a luminance value corresponding to each segment obtained as a result of the region division. The part is the original image information The n-th plane consisting of one or a plurality of continuous bit planes belonging to the highest order among the bit planes lower than the (n-1) -th plane group in the target, and the (n-1) th plane in the original image information is targeted. A) performing compression processing by the second area division method based on the area division result up to the plane group, n-th area division information indicating the result of area division, and a luminance value corresponding to each segment obtained as a result of the area division Is recursively repeated from n = 2 to an arbitrary natural number j that does not exceed k, and the second compression processing unit When performing the compression process by the area division method, if the shape of the segment generated in the (n-1) th plane group in the original image information is the same as the shape of the segment generated in the nth plane group, A symbol indicating that the segments are the same is issued as the area division information, and when all the segments generated in the n-th plane group are the segments of the minimum unit, it indicates that all the segments are the segments of the minimum unit. It is characterized by issuing a symbol.

In a preferred embodiment, the data processing section divides the original image information into a plurality of blocks, and compresses and codes each of the blocks by the compression processing section. .

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS In the present embodiment, lossless compression means a compression method in which data of an original image (area corresponding to each block) is completely stored, and irreversible compression means an original image. (Area corresponding to each block) is a compression method in which data is not always completely stored.

(First Embodiment) FIG. 1 is a schematic configuration diagram of a data processing apparatus according to a first embodiment. The data processing apparatus 101 receives image data, divides the image data into small blocks, and performs compression processing on each of the blocks independently, so that the compression ratio (bit rate) between the region of interest (ROI) and the non-region of interest of the image is reduced. An input device 102 for outputting different compressed data and inputting data indicating an original image and a region of interest, a computing device 103 for performing compression processing such as blocking and region division, and converting the original image data and region divided data It comprises a storage device 104 for storing and an output device 105 for outputting compressed data.

The original image is divided into a plurality of small blocks by the arithmetic unit 103, and each block is processed independently. The arithmetic unit 103 determines whether each block includes the ROI based on the data indicating the region of interest (region of interest information), and determines a group of blocks including the ROI (a group of blocks of interest) and a group of blocks not including the ROI. The group is classified into groups (non-attention block group), and lossless compression is performed for the attention block group, and lossless or irreversible compression is performed for the non-attention block group.

As a method of reversible or irreversible compression performed by the arithmetic unit 103, for example, a hierarchical area division (HS: Hierarchical Segme) described in reference (2).
ntation) coding can be used. The HS is configured as shown in the flowchart of FIG. First, a bit plane of at least one layer is cut out in a predetermined range (block) where processing in the input image is required (this is called an upper plane group) (S201, 20).
2).

Then, a predetermined area division is performed on the upper plane group to generate a segment, and area information including area division information and luminance value information is obtained (S203). next,
In the same image block, at least one layer of bit planes is cut out from the bit planes belonging to the lower order of the upper plane group in the bit order with the MSB as the highest order (this is called a lower plane group) (S204).

Further, the lower plane group is divided into regions based on the segments of the upper plane group. Specifically, it is determined whether or not a segment that is further subdivided than the segment obtained in the upper plane group is generated (S205). In the lower plane group, when the division further proceeds from each segment of the upper plane group, area information including area division information and luminance value information is generated (S207). On the other hand, if the division does not proceed from each segment of the upper plane group up to that point, area information consisting only of luminance value information is generated, but in some cases, area division information indicating that division is not necessary is also required as area information. (S206). Hereinafter, the processed upper plane group and lower plane group are combined into a new upper plane group, and the processing is performed recursively until the required bit plane is encoded (S208).

As a result, lossless compression and lossy compression can be performed. That is, the most significant bit plane (MSB)
Plane) to the least significant bit plane (LSB plane), lossless compression is performed when all bit planes are processed. In other cases (such as when some lower bit planes are not coded), lossless compression is performed. Become.

As described above, in this specification, selecting only a bit plane component of a certain layer of the original image data to be processed is referred to as “cutting out a bit plane”.

As a method of dividing the area in the arithmetic unit 103, for example, quadtree division can be used. 4
Split tree division is a method of dividing an area into four child areas as shown in FIG. In this case, the state of each segment can be represented by encoding "1" when dividing, and "0" when not dividing, and an arbitrary region shape can be encoded. This corresponds to the above-described area division information. The above-mentioned luminance value information is a luminance value assigned to each segment. For lossless compression, it is sufficient that each segment has the same luminance value as the original image. That is, the division may be performed recursively until each segment is configured with the same luminance value.

Conventionally, the shape of an ROI has been described as a set of polygons and circles. For example, in the case of document (1), ROI
Is represented as a set of rectangles and circles. This example is shown in FIG. FIG. 7 shows an example of specifying the shape of the ROI according to the first embodiment. In the present embodiment, the ROI shape can be approximated only by determining whether the ROI is included for each small block, whereas in the conventional example, parameters (vertex, center, radius, etc.) of a polygon and a circle that match the ROI area
I have to ask. Also, in the conventional ROI area designation method, since the shape of the ROI is not known in advance, the ROI
Although parallel processing of I and non-ROI is difficult, in this example, the block group of interest and the block group of non-interest including the ROI can be defined as a set of preset small blocks. In addition, parallel processing can be speeded up by individual processing.

(Embodiment 2) Since the overall configuration of a data processing apparatus according to the present embodiment is the same as that of Embodiment 1,
The numbers of each part shown in FIG. 1 are used. Data processing device 1
01 inputs image data, divides the image data into small blocks, and performs compression processing on each block independently.
An input device 102 for inputting data indicating an original image and a region of interest, which outputs compressed data having different compression ratios (bit rates) between a region of interest (ROI) and a non-region of interest of an image; An arithmetic unit 103 performs compression processing such as area division, a storage device 104 that stores original image data and area division data, and an output device 105 that outputs compressed data.

FIG. 4 is a schematic diagram of the processing in the arithmetic unit 103. Image data input device 40 from input device 102
The image data input from 1 is divided by the blocking unit 402 into
It is divided into a plurality of small blocks as in the first embodiment.
Each small block is classified by the context determining unit 403 based on the division shape of the peripheral block by the region divider 405. For example, the HS described in the first embodiment can be used as a method of region division by the region divider 405. The result of the area division by the area divider 405 is input to the context determining unit 403 through the delay unit 404.

FIG. 8 shows an example of referring to a peripheral block. In this example, the determination of the context of the block C currently being processed is the same as that of the block C adjacent to the top, the block B adjacent to the left, and the upper bit plane (group) input from the delay unit 404. This is performed based on the information of the block P located in the area.

The context is, for example, $ PLAIN, C
It is classified into ALM, EDGE, NOISY, etc. The number of categories to be classified and the classification method are arbitrary. In this case, when it is determined that the division of the area in the peripheral block is considerably advanced, the context of this block is determined to be NOISY, and If an edge exists in the block, it is determined to be EDGE. The CALM is applied when the division of the peripheral block has not progressed much, and the PLAIN is applied when the upper plane or the peripheral block does not exist, and when the division is not performed at all in the peripheral block.

As a specific example, using the above-described quad-tree division of FIG. 3 as the region division method, {PLAIN, EDGE, OT
FIG. 9 shows an example of classification into three contexts of HERSＥＲ. The classification rule here is PLAIN when no peripheral block is divided at all, EDGE when an edge exists in the peripheral block, and O in other cases.
It is assumed to be classified as THERS.

In FIG. 9A, since all of the peripheral blocks (blocks A, B, and P) are not divided at all, block C is classified as PLAIN. In FIG. 9B,
Since an edge is recognized in block A, it is classified as EDGE.

The blocks C thus classified are divided into regions by the region divider 405, and the encoder 406 encodes the region division information based on the respective contexts. The encoder 406 encodes both the area division information and the luminance value information, but the encoding performed based on the context is only the area division information. As a result, in the related art, simply {split / unsplit} is reduced to {0
The region division information recursively represented using the symbol of 1｝ is coded more efficiently.

This will be described with reference to an example of the encoding of the block C in FIG. In the block C of FIG. 9B, since a vertical edge is confirmed in the block A, the context is determined to be EDGE. Generally, it is known that image information has a strong correlation with neighboring pixels, and the periphery of a flat portion (a portion with few divisions) is also less divided, and the edge has a strong tendency to continue near a pixel where an edge exists. ing.

Here, FIG. 10 shows a typical example when a block corresponding to the block C in FIG. 9B is divided into regions.
The edge of the block A also occurs continuously in the block C of FIG. If the area division information of the block C is encoded by a conventional method, a total of 33 bits are required as shown in FIG. On the other hand, in the method according to the second embodiment,
It can be encoded by the bits required for the “Edge” symbol + 16 bits. The “Edge” symbol requires information such as the direction (horizontal / vertical) and position of the edge. Even when these are considered, in this case, 4 bits at most are sufficient for the “Edge” symbol. (1 bit to indicate direction, 2 bits to indicate position, 1 bit to indicate presence of edge). That is,
What conventionally required 33 bits can be represented by 20 bits. This is because the context is determined to be an edge by classifying it into a context, and the fact that an edge actually exists is used.

The area division information encoded by the encoder 406 is input to the storage device 104 or the compressed data output unit 407 to the output device 105 together with the luminance value information encoded in the same manner. .

(Embodiment 3) Since the overall configuration of the data processing apparatus of the present embodiment is the same as that of Embodiment 2, the numbers of the respective parts shown in FIG. 1 are used. The data processing apparatus 101 receives image data, divides the image data into small blocks, and performs compression processing on each of the blocks independently, so that the compression ratio (bit rate) between the region of interest (ROI) and the non-region of interest of the image is reduced. An input device 102 for outputting different compressed data and inputting data indicating an original image and a region of interest, a computing device 103 for performing compression processing such as blocking and region division, and converting the original image data and region divided data It comprises a storage device 104 for storing and an output device 105 for outputting compressed data.

FIG. 5 is a schematic diagram of the processing in the arithmetic unit 103. Image data input device 50 from input device 102
The image data input from 1 is divided by the blocking unit 502 into
It is divided into a plurality of small blocks as in the second embodiment.
Each small block is classified by the context determining unit 503 based on the division shape of the peripheral block by the region divider 505. For example, the HS described in the first embodiment can be used as a method of region division by the region divider 505. The result of the region division by the region divider 505 is input to the context determining unit 503 through the delay unit 504.

The operation of the context judging unit 503 is the same as the operation of the context judging unit 403 described in the second embodiment, except that the context information may also be input to the predictor 506. The case where context information is input to the predictor 506 will be described later.

The predictor 506 predicts the luminance value of each segment from the luminance values of the surrounding segments with respect to the luminance value information to be given to the segment generated in each block. FIG. 11 shows an example of prediction by the predictor 506. In the prediction for the current segment, first, the processing which has been completed among the segments A to F adjacent to the periphery and which has the largest size is selected. For example, when processing is performed by raster scanning, segments D, E, and F are unprocessed, and the largest one is selected from segments A, B, and C. Here, it is assumed that the segment A is selected.

Next, the luminance value of the segment A is used as a predicted value for the luminance value of the current segment. When processing is performed on one bit plane, the luminance value (predicted value) of each segment is 0 or 1. Finally,
Outputs the error between the predicted value and the true value.

When there are a plurality of segments having the largest size, the luminance value is determined by a predetermined method determined in advance. For example, a method of giving the highest priority to the segment at the left portion, obtaining the average luminance value of the largest segment, and rounding down the decimal part is used.

Here, the predictor 506 can use the context information obtained by the context judging unit 503. The predictor 506 that has obtained the context information uses the luminance value of the segment corresponding to the context information as the prediction value, instead of simply referring to the segment whose processing is completed and having the largest size at the time of prediction. FIG. 12 shows an example of the prediction in this case.

In FIG. 12, a certain segment in the block C determined to be an edge in FIG. 9B is a segment to be processed at present. The largest segment A to D adjacent to the current segment is segment B
However, since it is known from the context information that there is an edge in the vertical direction, the luminance value of the segment A is used as a predicted value.

Generally, in a region where an edge exists in image information, it is known that the brightness value correlation between pixels appears strongly in the direction along the edge and weakens in the direction across the edge. Is considered, the prediction accuracy is improved. Examples are shown from FIG. 13, FIG. 14, and FIG. FIG. 1 shows an example of plane group data composed of a single-layer bit plane.
3 is shown. Since it is a single-layer bit plane, the luminance value is 0.
Alternatively, it can be represented by a binary value of 1. FIG. 14 shows an example in which each block in FIG. 13 is divided into regions by quadtree division. As shown in FIG. 14, it is assumed that a vertical edge is confirmed in the surrounding block A in the context of the block C currently being processed, and that the block B is determined to be a “vertical edge” because the block B is not divided.

When encoding the current segment shown in FIG. 14 and trying to predict from the surrounding segments,
The segments A to D shown in FIG. When raster scanning is used, raster scanning is performed sequentially from left to right, one row at a time from the top of the image, so that only segments A and B have already been processed and the luminance value has been determined. Becomes Segments A and B
Since both have the same area, the above-mentioned case is the case where there are a plurality of segments having the largest size. Here, when the prediction segment is determined using the context information,
Since the current context is a "vertical edge",
Accordingly, segments adjacent in the vertical direction are prioritized over segments adjacent in the horizontal direction and become prediction segments.
That is, the predicted luminance value becomes the luminance value “0” of the segment A, and coincides with the luminance value “0” of the current segment.

Of course, this is only an example, but as described above, image information generally has a strong correlation with neighboring pixels.
In the vicinity of the pixel where the edge exists, the edge tends to continue, and the correlation of the pixel value appears strongly in the direction parallel to the edge. It is obvious that the prediction accuracy of the luminance value is improved as compared with the case where "when there are a plurality of segments, the segment at the left is given the highest priority".

(Embodiment 4) The overall configuration of the data processing apparatus of the present embodiment is the same as that of Embodiment 1,
The numbers of each part shown in FIG. 1 are used. Data processing device 1
Reference numeral 01 denotes an input device for inputting image data and performing compression encoding by region division, and an input device 102 for inputting data indicating an original image and a region of interest, and an arithmetic device 103 for performing compression processing such as blocking and region division. , A storage device 104 for storing original image data and area division data, and an output device 105 for outputting compressed data.

The area dividing method performed by the arithmetic unit 103 is as follows.
For example, the HS described in Embodiment 1 can be used. The arithmetic unit 103 encodes whether each segment generated by the HS is divided by the lower plane group, and when one or a plurality of segments are divided into the same shape even in the lower plane group, , A symbol indicating that the symbols are "divided into exactly the same shape" is separately prepared, and this is issued as region division information.

Similarly, when the arithmetic unit 103 encodes whether each segment generated by the HS is divided into lower plane groups, one or a plurality of segments are lower plane groups, In this case, a symbol indicating that “all segments are divided into the smallest unit” is separately prepared, and this is issued as area division information.

As a result, the generated area division information can be reduced compared to the conventional area division information issued for each segment. The effect will be described with reference to an example. In the following, it is assumed that the arithmetic unit 103 uses HS by quadtree division.

In a block composed of 8 pixels × 8 pixels of a certain plane group P1, a segment as shown in FIG.
Suppose that it was obtained by S. One circle represents one pixel,
Assume that white / black corresponds to pixel value 0/1. A plane group that is one level lower than this plane group is defined as P2. P
At 2, it is determined whether this segment is further divided.

Here, assuming that the data is as shown in FIG. 17 in the plane group P2, the shape of the segment in the plane group P2 is exactly the same as that of the plane group P1. At this time, the arithmetic unit 103 displays the symbol SS (Same Segmen) indicating that the symbol is “divided into exactly the same shape”.
Tation). The bit length of the SS symbol varies depending on the design method of the entire codeword.
~ 7 bits. On the other hand, if “no division” is set to 0, the conventional coding method requires 19 bits, the same as the number of segments, as shown in the lower part of FIG. Similarly,
Assuming that the data is as shown in FIG. 18 in the plane group P2, the area of each segment in the plane group P2 is 1 pixel × 1 pixel (minimum unit).

At this time, the arithmetic unit 103 outputs the symbol CS (Compl
issue (Every Segmented). The bit length of the CS symbol varies depending on the design method of the entire codeword, but is generally 3 to 7 bits. On the other hand, if "not split" is set to 0 and "divided" is set to 1, the conventional encoding method requires 15 bits of the remaining segment division information required by the minimum unit as shown in the lower part of FIG.

That is, the SS / CS symbol can indicate the division state with less bits, which is very effective.

[0064]

According to the present invention, independent data processing can be performed for each small image at the time of data processing by region division, and the processing time can be reduced. Further, by the context classification and the segment luminance value prediction, encoding can be performed using the correlation remaining between blocks and between segments, and the compression ratio can be improved.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of a data processing device according to a first embodiment;

FIG. 2 is a flowchart of an area dividing process according to the first embodiment.

FIG. 3 is a schematic diagram of quadtree partitioning.

FIG. 4 is a schematic configuration diagram of region division compression encoding in an arithmetic unit 103 including a context determination unit;

FIG. 5 is a schematic configuration diagram of region division compression encoding in an arithmetic unit 103 including a context determination unit and a predictor;

FIG. 6 is a diagram showing an example of a conventional ROI shape description method.

FIG. 7 is a diagram showing an example of a method of describing the shape of an ROI according to the first embodiment.

FIG. 8 is a diagram illustrating a reference example of peripheral blocks of a context determining unit according to the second embodiment;

FIG. 9 is a diagram showing an example of context determination by a context determination unit.

FIG. 10 is a diagram showing an example of encoding region division information based on context according to the second embodiment.

FIG. 11 is a diagram showing a prediction example of a predictor 506 according to the third embodiment.

FIG. 12 is a diagram showing a prediction example of a predictor 506 according to the third embodiment.

FIG. 13 shows an example of image data.

14 is a diagram showing an example of image data and an example of area division with respect to FIG.

FIG. 15 shows an example of image data and a peripheral view of a current segment in FIG.

FIG. 16 is a diagram showing an example of data according to the fourth embodiment.

FIG. 17 illustrates an example of data according to the fourth embodiment.

FIG. 18 is a diagram showing an example of data according to the fourth embodiment.

[Explanation of symbols]

Reference Signs List 101 data processing device 102 input device 103 arithmetic device 104 storage device 105 output device 401 image data input device from input device 102 402 blocking unit 403 context determination unit 404 delay unit 405 region divider 406 encoder 407 storage device 104 or Compressed data output unit 501 to output device 105 Image data input unit from input device 102 502 Blocking unit 503 Context determination unit 504 Delay unit 505 Region divider 506 Predictor 507 Encoder 508 To storage device 104 or output device 105 Compressed data output section

Claims (17)

[Claims] 1. A data processing apparatus for compressing original image information comprising a plurality of pixels, wherein each pixel is associated with a luminance value of a natural number k (k ≧ 2) bits to generate compressed encoded data. Having an input unit, a data processing unit and an output unit,
The input unit receives the original image information, the data processing unit divides the original image information into a plurality of blocks,
A data processing apparatus, wherein each block is compression-encoded by a compression processing unit at an arbitrary bit rate, and the output unit outputs compression-encoded data that has been compression-encoded by the data processing unit. 2. The input unit according to claim 1, wherein the input unit receives original image information and attention area information indicating an attention area of the original image information, and the data processing unit divides the original image information into a plurality of blocks. The blocks are divided, and the blocks are classified into a block group including the region-of-interest information (referred to as a group of target blocks) and a block group not including the region-of-interest information (referred to as a non-target block group), A data processing apparatus, wherein each block group is compression-encoded by a compression processing unit at an arbitrary bit rate. 3. The data processing unit according to claim 2, wherein the block of interest is losslessly compression-coded by a compression processing unit, and the non-target block group is compression-encoded losslessly or irreversibly by a compression processing unit. A data processing apparatus, wherein the output unit outputs the compressed coded data of the block group of interest before the compressed coded data of the block group of non-target block. 4. The method according to claim 1, 2 or 3.
, The data processing unit has a first compression processing unit and a second compression processing unit as the compression processing unit, wherein the first compression processing unit is configured to include one or a plurality of continuous bit planes in the original image information. For the first plane group consisting of
First compression information including first region division information indicating a result of the region division and first luminance information indicating a luminance value corresponding to each segment obtained as a result of the region division. And the second compression processing unit generates
The n-th plane group consisting of one or a plurality of continuous bit planes belonging to the highest bit plane lower than the (n-1) -th plane group in the original image information is targeted, and Based on the area division result up to the (n-1) th plane group, the compression processing is performed by the second area division method, and the n-th area division information indicating the result of the area division and each segment obtained as a result of the area division A process in which n-th compressed information composed of n-th luminance information indicating a luminance value corresponding to n is recursively repeated from n = 2 to any natural number j not exceeding k. Processing equipment. 5. The data processing apparatus according to claim 4, wherein the first and second area division methods are quadtree division. 6. The data processing unit according to claim 1, wherein the data processing unit has a context determining unit, and the context determining unit is configured to encode one or more blocks around the block when encoding each of the divided blocks. Based on the distribution of the internal brightness value, the state of the block is classified,
The data processing unit, wherein the data processing unit performs compression encoding for each state classified by the context determination unit. 7. The data processing unit according to claim 2, wherein the data processing unit has a context determining unit, and the context determining unit is configured to encode one or more blocks around the block when encoding each of the divided blocks. Based on the distribution of the internal brightness value, the state of the block is classified,
The data processing device, wherein the data processing unit performs compression encoding for each state classified by the context determination unit. 8. The data processing unit according to claim 3, wherein the data processing unit has a context determining unit, and the context determining unit is configured to encode one or more blocks around the block when encoding each of the divided blocks. Based on the distribution of the internal brightness value, the state of the block is classified,
The data processing device, wherein the data processing unit performs compression encoding for each state classified by the context determination unit. 9. The data processing unit according to claim 4, wherein the data processing unit has a context determining unit, and the context determining unit is configured to encode one or more blocks around the block when encoding each of the divided blocks. Based on the distribution of the internal brightness value, the state of the block is classified,
The data processing device, wherein the data processing unit performs compression encoding for each state classified by the context determination unit. 10. The data processing unit according to claim 5, wherein the data processing unit has a context determining unit, and the context determining unit is configured to encode one or more blocks around the block when encoding each of the divided blocks. Based on the distribution of the internal brightness value, the state of the block is classified,
The data processing device, wherein the data processing unit performs compression encoding for each state classified by the context determination unit. 11. The method according to claim 9, wherein, when encoding each of the divided blocks, the context determination unit may include one or more of the blocks around the block on the same bit plane as the block. Based on the distribution of luminance values inside a plurality of blocks and a block located at the same position as the block on a bit plane that is higher than the block (referred to as an upper block), the state of the block is classified, The data processing unit, wherein the data processing unit performs compression coding for each state classified by the context determination unit. 12. The method according to claim 4, wherein the first compression processing section has a first segment luminance value prediction section, and the second compression processing section has a second segment luminance value prediction section. , The first segment luminance value prediction unit performs a predetermined segment based on the size and the luminance value of one or a plurality of segments adjacent to each segment (referred to as an encoded segment) obtained as a result of the region division. Generating a luminance prediction value by a method, encoding a prediction error between the luminance value of the coded segment and the coded luminance prediction value by a predetermined method, and using the encoded error as the first luminance information, the second segment luminance value prediction unit Generates a luminance prediction value by a predetermined method based on the size and luminance value of one or a plurality of segments adjacent to the coding segment, and generates the luminance value of the coding segment and the coding luminance prediction value. A data processing apparatus characterized in that a prediction error from a measured value is encoded by a predetermined method and used as the n-th luminance information. 13. The method according to claim 9, wherein the first compression processing section has a first segment luminance value prediction section, and the second compression processing section has a second segment luminance value prediction section. , The first segment luminance value prediction unit performs a predetermined segment based on the size and the luminance value of one or a plurality of segments adjacent to each segment (referred to as an encoded segment) obtained as a result of the region division. Generating a luminance prediction value by a method, encoding a prediction error between the luminance value of the coded segment and the coded luminance prediction value by a predetermined method, and using the encoded error as the first luminance information, the second segment luminance value prediction unit Generates a luminance prediction value by a predetermined method based on the size and the luminance value of one or a plurality of segments adjacent to the encoded segment, and generates the luminance value of the encoded segment and the encoded luminance. A data processing apparatus characterized in that a prediction error from a prediction value is encoded by a predetermined method and used as the n-th luminance information. 14. The data processing device according to claim 11, wherein the first compression processing unit has a first segment luminance value prediction unit, and the second compression processing unit has a second segment luminance value prediction unit, The first segment luminance value prediction unit performs a predetermined method based on a size and a luminance value of one or a plurality of segments adjacent to each segment (referred to as an encoded segment) obtained as a result of the region division. And a prediction error between the luminance value of the coded segment and the coded luminance prediction value is coded by a predetermined method, and is used as the first luminance information. Generating a luminance prediction value by a predetermined method based on the size and the luminance value of one or a plurality of segments adjacent to the encoded segment, and generating the luminance value of the encoded segment and the encoded luminance prediction value. A data processing apparatus characterized in that a prediction error from a measured value is encoded by a predetermined method and used as the n-th luminance information. 15. The data processing device according to claim 13 or 14, wherein the first segment luminance value prediction unit and the second segment luminance value prediction unit generate the luminance prediction value when the first segment luminance value prediction unit generates the luminance prediction value. Based on the state of the block classified by the context determination unit and the size and luminance value of one or more segments adjacent to each segment (referred to as an encoded segment) obtained as a result of the area division. A data processing device for generating a luminance prediction value by a predetermined method. 16. A data processing apparatus for compressing original image information which is composed of a plurality of pixels and in which each pixel is associated with a luminance value of a natural number k (k ≧ 2) bits to generate compression encoded data. , An input unit, a data processing unit, and an output unit, wherein the input unit receives the original image information, and the data processing unit includes a first compression processing unit and a second compression processing unit as a compression processing unit The first compression processing unit performs a compression process using a first area division method on a first plane group including one or a plurality of continuous bit planes in the original image information, and obtains a result of the area division. And first luminance information including first luminance information indicating a luminance value corresponding to each segment obtained as a result of the region division. The second compression processing unit generates the first compression information. (N-1) th plane group inside The n-th plane consisting of one or a plurality of continuous bit planes belonging to the highest order among lower bit planes is targeted, and a region division result up to the (n-1) th plane group in the original image information is used as a reference. The compression process is performed by the second region division method, and the n-th region division information including the n-th region division information indicating the result of the region division and the n-th luminance information indicating the luminance value corresponding to each segment obtained as a result of the region division The process of creating compression information,
n is recursively repeated from n = 2 to an arbitrary natural number j which does not exceed k, and when the second compression processing unit performs compression processing by the second area division method, the second ( n-1) When the shape of the segment generated in the plane group and the shape of the segment generated in the n-th plane group are the same, a symbol indicating that the segments are the same is issued as the n-th region division information; A data processing device, wherein when all the segments generated in the n-th plane group are segments of the minimum unit, a symbol indicating that the segments are all segments of the minimum unit is issued. 17. The data processing apparatus according to claim 16, wherein the data processing unit divides the original image information into a plurality of blocks, and compresses and codes each of the blocks by the compression processing unit. Data processing device.