JPH0819002A - Image processing unit and its method - Google Patents

Abstract

PURPOSE:To store image data effectively with a segment information table of a minimum scale by using a stored segment number for an address number of a table in which code data of a stage and a segment number to be selected next are stored for an address number. CONSTITUTION:Code data of stages (1)-(4) are written respectively 9 in segments S-1 to S-4. When the segment S-1 is filled, an idle segment S-5 is selected for code data in the stage (1) and the data are stored therein. Then a number '5' of a selected segment is written in the address 1 of the segment information table. Similarly, when the segments S-2, S-3 are filled, idle segments S-6, S-7 are respectively selected and the segment numbers '6', '7' are written in the addresses 2, 3. The segment numbers are stored in a table configured by a single memory.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus and method for compressing / decompressing color image data for storage.

[0002]

2. Description of the Related Art Conventionally, DCT transform (discrete cosine transform)
In the color image compression / expansion method based on, the variable length coding method is used as a coding method, and therefore the compression data amount cannot be controlled by a general application method. Therefore,
There is a proposal for a memory system divided into a plurality of segments and a method for distributing the obtained code data in multiple stages.

The code data distributed in these multiple stages is stored by properly selecting a memory segment for each stage. Then, when there are no more segments to select, among the already stored segments, the segment in which the code data of the low priority stage is stored is invalidated, and overwriting is newly performed to realize compressed data control. are doing. Further, since it is necessary to correctly take out the code data in the segment memory at the time of decompression, a proposal has been made for a segment information table in which the segment information is stored.

[0004]

However, the above-mentioned conventional segment information table requires a table in which the segment number selected for each distribution step is stored, and therefore it is necessary to have the same number of tables as the number of distribution steps. There is. Further, in order to improve the control accuracy of the amount of compressed data, it is necessary to increase the number of stages, and in the above-mentioned conventional segment information table, there is a problem that the scale of the segment information table increases as the number of stages increases. is there.

The present invention has been made in view of the above-mentioned problems, and an object of the present invention is to provide a minimum-scale segment information regardless of the number of stages of distributing encoded data to a memory divided into a plurality of segments. An object of the present invention is to provide an image processing device and method for effectively storing image data in a table.

[0006]

In order to achieve the above object, the invention according to claim 1 converts color image data into spatial frequency components, and quantizes the converted data to obtain the color data. In an image processing device for compressing an image, means for distributing the quantized image data in a plurality of predetermined stages, means for encoding each of the image data distributed in the plurality of stages, and the encoding First storage means divided into a plurality of segments for storing the selected image data, and second storage means for storing a predetermined selection sequence of the plurality of segments,
In order to store the code data in each of the plurality of stages according to the selection sequence, means for selecting a segment of the first storage means, and the selection result in the second
Storing means in the storage means.

In the above structure, the segment number table having a large number of entries is stored in a table constituted by a single memory, and the scale of the segment information table is large even if the number of steps for distributing the code data increases. Therefore, compressed image data can be stored at low cost.
Further, the invention according to claim 5 is further provided with a means for decoding the coded data at each of the steps, and a segment selection for selecting a segment of the first storage means according to a step required by the decoding. And means for inversely quantizing the quantized coefficient obtained by the decoding and inversely converting the inversely quantized image data of the spatial frequency component into color image data.

With this configuration, one segment information table can be used for compression, and the accumulated image data can be expanded and decoded.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings. FIG. 1 is a block diagram showing the arrangement of an image processing apparatus according to an embodiment of the present invention. A segment information table described later and its control device are applied to the device shown in FIG.

In FIG. 1, reference numeral 1 is a color conversion unit, 2 is a sub-sampling unit, 3 is a DCT (discrete cosine transform) unit, and 4 are
Is a quantization unit, 5 is a quantization table, 6 to 9 are Huffman coding units, and 10 is a Huffman table. Further, 11 is a segment control unit, 12 is a segment information table, 13
Is a compression memory, and 14 is a CP for controlling the entire apparatus.
U (central control unit), 15 is a ROM for storing data such as an operation program of the CPU 14, and 16 is a RAM used as a work area for various programs.

The color conversion unit 1 shown in FIG. 1 obtains Y, U, and V from RGB image data by 3 × 3 linear matrix conversion represented by the following equation (1). Here, Y indicates a luminance component, and U and V indicate chromaticity components.

Further, the sub-sampling unit 2 utilizes the fact that the sensitivity characteristic of the human eye is more sensitive to the luminance component Y than to the chromaticity components U and V, so that the sub-sampling of Y, U and V is performed. Is performed. In this case, Y: U: V = 4:
Conversion and sampling with a ratio of 1: 1, or Y: U: V = 4: 2: 2 are performed. The DCT unit 3 performs discrete cosine transform on each of 8 × 8 blocks for each of Y, U, and V to transform into spatial frequency components.
Note that here, the coefficients thus transformed are called DCT coefficients.

In the quantizer 4, the above DCT coefficient is 8
Quantization is performed using the quantization table 5 for each × 8 block. The coefficient quantized in this way is called a quantized coefficient. Then, the 8 × 8 two-dimensional quantized coefficients are one-dimensionally rearranged from low frequency components to high frequency components by zigzag scanning, as shown in FIG. In progressive coding, the quantized coefficients arranged in one dimension are divided into several stages (stage (1) to stage (4) in FIG. 3) from low frequency components to high frequency components, as shown in FIG. To do. The stage (1) has a Huffman encoding unit 6
The stage (2) is Huffman-encoded by the Huffman encoder 7, the stage (3) is Huffman-encoder 8 and the stage (4) is Huffman-encoder 9.

However, these Huffman code units 6 to 9
Are all encoded using the same Huffman table 10. Note that, here, the data obtained by Huffman coding is referred to as coded data. The compression memory 13, as shown in FIG.
00 kBytes) is divided into S-1 to SN. Also, the segment selection is selected according to the segment selection transition table shown in FIG. 5 and written in the segment information table shown in FIG. The segment selection transition table shown in FIG.
The information regarding (4) is shown, and the first column of the table shows the selected segment numbers (S-1 to SN) of the compression memory 13 in which the encoded data of the stage (1) is written, It indicates that the image goes from the beginning to the end as it goes from left to right. In addition, "END '" in the table indicates that the encoded data in each stage is completed.

Next, the procedure for writing the encoded data in the compression memory 13 for each stage, that is, the multi-step encoding of the image data and the method for writing the encoded data in the memory divided into a plurality of segments. Will be described. FIG. 7 is a flowchart illustrating a processing procedure for storing code data in the compression memory 13 divided into a plurality of segments according to the present embodiment.

Since the image processing apparatus according to the present embodiment uses the Huffman coding method which is a variable length coding method, the code amount in each stage also differs. The code amount depends on the original image data, but when a natural image is used as the original image, the power is generally concentrated on the low frequency component. Therefore, the stages (1) to
When comparing (4), the amount of codes output from the stage (4) is larger in the stage (1).

As shown in the second column of FIG. 5 (column No. 2), the encoded data output at each stage is segment S-1, S-2, S-3, S-4. Written in. Then, in the stage (1), since the low frequency component data is encoded, the segment S-1 is filled earlier than the other stages, and the writing to the empty segment S-5 starts.

Next, in stages (2) and (3), the writing to the segments S-2 and S-3 is satisfied, and the writing to the segments S-5 and S-6 is started ( Step S31). In the next stage, if stage (1) finishes writing to segment S-5 before stage (4) finishes writing to segment S-4, it leaves empty segment S-8. Is not assigned to stage (4) but to stage (1). This occurs because the code amount generated in the stage (1) is larger than that in the stage (4).

Similarly, in each stage, if the segment to be written is satisfied in each stage, a vacant segment is selected and the encoded data is written therein (step S32, step S33, step S3).
4). In this embodiment, a method of dividing the original image into several stages and encoding is used. Therefore, when the encoded data amount of each stage is smaller than the target compression (code) memory amount, the invalid stage does not appear. However, when the target compressed (coded) data amount is reached during encoding, the designated stage (stage (4) in FIG. 5) is invalidated, and the segment S- 4, the encoded data of the other stages (1) to (3) are written in S-11, S-15, ....

This processing is performed on the 8th to 11th lines (No.
8 to No. 11) and the segment SN is assigned in the stage (2) of the eighth column, all the segments of the compression memory 13 are assigned. Therefore, in order to make up for the shortage of the compression memory 13,
The stage (4) is invalidated (“0”), the segment S-4 used in the stage (4) is used in the stage (1), and the segment S is used as shown in the ninth column. -11 is used for stage (2).

At the stage (2), the code data ends at the segment S-11, and thereafter, the END mark is added as shown in FIG. Also, on stage (1),
As in the tenth column, the segment 15 used in the stage (4) is allocated, and since the code data is completed thereafter, the END mark is added (step S
35).

Next, the segment information table shown in FIG. 6 in which the selected segment number is stored, and FIG.
The control procedure will be described with reference to the flowchart shown in FIG. 6 and the configuration of the segment control unit 11 will be described with reference to FIG. The segment information table shown in FIG. 6 has addresses corresponding to the total number of segments in the compressed memory and
It is a memory configured by the number of bits that can represent the total number of compressed memory segments per address. The contents of the table indicate the selected segment number of the code data at a certain stage, and the code data at the stage also indicate the address number at which the segment number to be selected next is stored.

In this embodiment, since the number of stages of code data is 4, the code data of stages (1) to (4) are written in the segments S-1 to S-4, respectively. The next segment number to select is address 1
Is written to. Similarly, stages (2), (3),
The segment numbers selected next to (4) are written in addresses 2, 3 and 4, respectively.

In the flowchart shown in FIG. 8, in step S11, the code data of stages (1) to (4) are written in the segments S-1 to S-4, respectively. Here, since the code data has not been completed (NO in step S12), the presence or absence of the filled segment is determined in step S13. That is, since the segment S-1 is satisfied (YES in step 13), the empty segment S-5 is selected and stored as the code data of the stage (1). Therefore, in step S14, the selected segment number "5 '" is written in the address 1 of the segment information table (see FIG. 6).

Similarly, when the segments S-2 and 3 are filled, the vacant segments S-6 and 7 are selected, and the segment numbers "6 ',"7'are written to the addresses 2, 3. Be done. When the segment S-5 is filled, the vacant segment S-8 is selected as the code data of the stage (1), and the segment number -8 '
Is written in the address 5.

Similarly, if a segment is filled, a free segment is selected, and the selected segment number is written at the same address as the filled segment number. When the target compression (code) memory amount is reached during encoding, the code data of the designated stage is invalidated, and other valid coded data is written in the segment assigned to that stage.

In the segment information table shown in FIG. 6,
Since the segment number N is written in the address N-3, it means that all the compressed data segments have been selected. Then, as described above, since the segment to which the stage (4) is assigned is invalid, the segments S-4, S-11, S-15, ... Are in a free state, and the segment S-4 is newly added. It is assigned to the encoded data of stage (1). Therefore, the segment number "4 '" is written in the address N-1 of the segment information table.

Further, the segment S-11 is assigned to the encoded data of the stage (2), and the segment number "11 '" is written in the same address N as the segment number N which has been accessed so far. Finally, when the segment S-4 is filled, the vacant segment S-15 is selected and the segment number "1" is assigned to the address 4.
5'is newly written (step S15 to step S18).

Next, the structure of the segment controller 11 will be described in detail. In the segment control unit shown in FIG. 9, the free segment generation unit 20 outputs an unused segment number. During compression, the next segment latch 21 temporarily stores the invalid segment number after all the segments have been selected, and that segment is selected next. When decompressing, the segment number to be selected next is taken out from the segment information table and temporarily stored.

Gates 22 and 27 are gate portions for opening and closing signals, and are stage 1 latch 23, stage 2 latch 24, stage 3 latch 25, stage 4 latch 2
In 6, the segment numbers of the code data of the stages 1, 2, 3, 4 which are currently selected are written. In the segment controller shown in FIG. 9, the stage 1 latch 23, the stage 2 latch 24,
The stage 3 latch 25 and the stage 4 latch 26 respectively have the segment numbers “1 ′” and “2” to be selected first.
',' 3 ', and'4' are written. Then, the compression memory 13 has a stage latch (23 to 26) at that stage.
A segment is selected on the basis of the output of, and the code data is written.

When the segment S-1 is satisfied, the output 1 of the stage 1 latch 23 is used as the address of the segment information table 12, and the empty segment number "5 '" output from the empty segment generator 20 changes the gate 27. It is written in the segment information table 12 via The empty segment number 5'is newly written in the stage 1 latch 23, and the code data of the stage 1 is the newly written segment number 5 '.
'Is stored in the compression memory 13.

Similar processing is performed for the stages 2 and 3, and if the segments S-2 and 3 are respectively filled, the stage 2 latch 24 and the stage 3 latch 25 are provided.
The outputs 2 and 3 of 1 are used as the addresses of the segment information table 12, and the empty segment numbers “6 ′” and “7 ′” output from the empty segment generator 20 are newly written in the segment information table 12 via the gate 27.

Also, these empty segment numbers-6
′ And ‘7’ are newly written in the stage 2 latch 24 and the stage 3 latch 25, respectively.
Thereafter, the code data of is stored in the compression memory 13 based on the newly written segment number. In this way, until all the segments are selected, the segment selection processing is performed in the same manner as above, and if all the segments are selected, the invalid stage code is stored in the next segment latch 21. The first invalid segment number is written. As described above, in this embodiment, since the stage (4) is invalid, the segment number "4 '" is written in the next segment latch 21. Then, if segment S- (N-1) of stage (1) is satisfied, the following three operations are performed. That is, (1) With the output N-1 of the stage 1 latch 23 as the address of the segment information table 12, the output 4 of the next segment latch 21 is newly written.

(2) Output of the next segment latch 21 to the stage 1 latch 23 Selected empty segment number `4
′ Is newly written. (3) The selected empty segment number 4 is set to the gate 22
Through the segment information table 12 as the address of the segment information table 12, the next segment latch 21 is read from the segment information table 12.
To be stored.

Hereinafter, the same processing is repeated until the code data is completed. On the other hand, when decompressing the compressed data stored in the compression memory 13, the image data is decoded using only the valid stages (1) to (3). The Huffman coding units 6 to 9 shown in FIG. 1 are provided with four Huffman decoding units, the Huffman table 10 includes a Huffman decoding table for decoding the coding of the tables, and the quantization unit 4 includes The inverse quantization unit and the quantization table 5 are provided with an inverse quantization table, and the DCT unit 3 is provided with an inverse DCT unit, respectively.

In this decompression processing, the segment control unit 11
Reads the coded data from the compression memory 13 according to the segment information table 12. The processing steps will be described below. In this embodiment, since the code data of the stage (4) is invalid, the initial setting is the stage 1 latch 23, the stage 2 latch 24, and the stage 3 latch 2.
5, the segment numbers "1 '", "2'" and "3 '" to be initially selected are written in 5, respectively. Then, for example, if the code data of the stage (1) is requested according to the stage number of the code data requested by the Huffman decoding unit, the code of the segment number `1 ′ stored in the stage 1 latch 23 The data is transferred from the compression memory 13 to the Huffman coding section.

At that time, the output 1 of the stage 1 latch 23 is used as the address of the segment information table 12, and the segment number 5 to be selected next to the stage (1) is selected.
′ Is written from the segment information table 12 to the next segment latch 21. When all the segments S-1 have been read, the segment number 5'of the next segment latch 21 is written in the stage 1 latch 23, and the code data of the stage (1) is stored in the compression memory 13
Is read from the segment S-5.

Thereafter, similarly, the coded data of each stage is read from the compression memory 13 and transferred to the above four Huffman decoding units. Then, in these four Huffman decoding units, the encoded data of each stage is decoded using the above Huffman decoding table, and transferred to the dequantization unit. In this inverse quantization unit, inverse quantization is performed for each 8 × 8 block from the decoded data of all valid stages.

Further, the inverse DCT unit performs an inverse DCT transform on the data of the inverse quantization unit for each 8 × 8 block. And in the sub-sampling unit,
Enlargement processing corresponding to the sub-sampling unit is performed, and the color conversion unit performs the inverse conversion process of the above equation (1). As described above, according to the present embodiment, in the segment information table, the stored segment number is stored as the segment number selected next by the code data of that stage among the code data distributed in multiple stages. By using the address number of the table to be stored, it is not necessary to prepare segment information tables for the number of stages, and the compressed data amount can be controlled with only one segment information table, and even if the number of stages for distributing the code data increases. The size of the segment information table does not become large, and control of the amount of compressed data can be realized at low cost.

The present invention may be applied to a system composed of a plurality of devices or an apparatus composed of a single device. Further, it goes without saying that the present invention can be applied to the case where it is achieved by supplying a program to a system or an apparatus.

[0041]

As described above, according to the first aspect of the present invention, the segment number having the entries for the total number of segments is stored in the table constituted by the single memory, and the code data is distributed. Even if the number of stages increases, the size of the segment information table does not increase, and compressed image data can be stored at low cost.

According to the invention described in claim 5, 1
It is possible to perform compression by using one segment information table, decompress the accumulated image data, and perform decoding processing.

[Brief description of drawings]

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

FIG. 2 is a diagram showing a zigzag scan.

[Fig. 3] Fig. 3 is a diagram illustrating the number of stages of progressive coding and a quantization coefficient.

FIG. 4 is a diagram showing a structure of a compression memory 13 divided into segments.

FIG. 5 is a diagram showing a segment selection transition table.

FIG. 6 is a diagram showing a segment information table.

FIG. 7 is a flowchart showing a segment selection procedure.

FIG. 8 is a flowchart showing a storage procedure of selected segment numbers.

9 is a block diagram showing a configuration of a segment control unit 11. FIG.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 color conversion part 2 sub-sampling part 3 DCT (discrete cosine transform) part 4 quantization part 5 quantization table 6-9 Huffman coding part 10 Huffman table 11 segment control part 12 segment information table 13 compression memory 14 CPU (central control part) ) 15 ROM 16 RAM

 ─────────────────────────────────────────────────── --Continued Front Page (72) Inventor David Reid Denman Street, Tramla, 2074, New South Wales, Australia 11

Claims (7)

[Claims] 1. An image processing apparatus for converting color image data into a spatial frequency component, quantizing the converted data and compressing the color image, wherein the quantized image data is subjected to a plurality of predetermined steps. A means for allocating each of the image data distributed to the plurality of stages, a first storage means partitioned into a plurality of segments for storing the encoded image data, Second storage means for storing a predetermined selection sequence of the plurality of segments; and the first storage means of the first storage means for storing the code data at each stage of the plurality of stages according to the selection sequence. An image processing apparatus comprising: a unit that selects a segment; and a unit that stores the selection result in the second storage unit. 2. The second storage means has addresses for the total number of segments of the plurality of segments that divide the first storage means, and one unit of the address is a bit size capable of expressing the total number of segments. The image processing apparatus according to claim 1, further comprising: 3. The selection sequence of the segments stored in the second storage means stores the selection sequence of the segment to be selected next by the code data at a predetermined stage of the plurality of stages. The image processing apparatus according to claim 1, wherein the address of the second storage unit is indicated. 4. The means for determining whether or not the data amount of the code data has reached a predetermined data amount during the compression processing, and the data amount of the code data has reached a predetermined data amount. In this case, the predetermined segment includes means for invalidating image data assigned to a predetermined step among the plurality of steps, and segment storage means for storing the number of the segment corresponding to the invalid image data. 2. The image processing apparatus according to claim 1, wherein the image data at the stage where is satisfied is written in a segment having an address corresponding to the segment stored in the segment storage unit as a free segment. 5. A means for decoding the coded data at each of the steps, a segment selecting means for selecting a segment of the first storage means according to a step required by the decoding, and the decoding. 2. The image according to claim 1, further comprising: a unit that inversely quantizes the quantized coefficient obtained by the quantization, and inversely converts the image data of the dequantized spatial frequency component into color image data. Processing equipment. 6. The segment selecting means further holds, for each code data of each of the plurality of steps, a number of a segment currently selected by each of the steps, and the current selection. The image processing apparatus according to claim 5, further comprising means for holding the number of the segment to be selected next, when all the segments are read out. 7. An image processing method for converting color image data into spatial frequency components, quantizing the converted data and compressing the color image, wherein the quantized image data is subjected to a plurality of predetermined steps. And a step of encoding each of the image data distributed in the plurality of stages, and storing the encoded image data in a first memory divided into a plurality of segments for storing the encoded image data. A first storing step; a second storing step of storing a predetermined selection sequence of the plurality of segments in a second memory; and the code data for each of the plurality of stages according to the selection sequence. An image comprising: selecting a segment of the first memory for storage in a second memory; and storing the selection result in the second memory. Processing method.