JPH06319047A - Information processing equipment - Google Patents

Abstract

(57) [Summary] [Purpose] There is a compression method that is more suitable for the data to be compressed.
If an inappropriate compression method is used, the size of the compressed data will be larger than the original data in the worst case. For this reason, it is inefficient to select another compression method if compression fails once. Select the optimal compression method and combine and compress without test compression. [Structure] The image data input to RAM 1 (16) is
The data is transferred to the work RAM 2 (19) by the DMA 17.
During the transfer, the counter provided in the DMA 17
The appearance frequency of the specific pattern is measured, the CPU 11 selects an appropriate compression method according to the value of the counter, compresses the data in the work RAM 2 (19), and stores it in the compressed data storage RAM 3 (18). To do.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information processing device,
In particular, it relates to an information processing device that compresses and outputs information.

[0002]

2. Description of the Related Art Conventionally, as an information processing apparatus of this type, if the original data is all compressed and the size of the compressed data becomes larger than the original data, or if the compressed data memory overflows, the current data Compression method such as copying or switching to another compression method without compression, or compressing a part of the original data and comparing the compressed data with the original data as shown in Japanese Patent Laid-Open No. 62-259183. However, a method of compressing all original data only when the compressed data has a smaller size than the original data has been adopted.

FIG. 4 is a simplified block diagram of a conventional image reading apparatus. The known CPU 41 includes a ROM 43, an image binarization circuit 44, and an image storage R via a bus 42.
AM1 (46), work RAM2 (49), RAM1
DMA for data transfer from (46) to RAM2 (49)
47, connected to the RAM 3 (48) for storing compressed data. The CPU 41 repeats the image reading and compression work by the program stored in the ROM 43 in advance. When the binarization circuit 44 receives a read command from the CPU, it converts the image information received from the image sensor 45 into two.
It is digitized and stored in RAM1 (46). When the predetermined number of data is written in the RAM1 (46), the binarization is stopped and the binarization end flag is set to notify the CPU 41 that the binarization is completed.
When the CPU 41 confirms the binarization end flag set by the binarization circuit 44 by polling or the interrupt flag, it transfers the image of the RAM 1 (46) to the work RAM (49) by the known DMA 47. The CPU 41 compresses the original data in the working RAM (49) to the RAM 3 (48).
To store.

FIG. 5 shows an R of an image reading apparatus as a conventional example.
It is a figure which shows the flow of the image reading compression program memorize | stored in OM43. When the reading is started in S5.1, a reading command is issued to the binarizing circuit 44 in S5.2, and the binarization completion flag is set and the data is waited for in S5.3. Once the data is prepared C
The PU 41 gives an instruction to the DMA 47 in S5.4 to transfer the data in the RAM 1 (46) to the work RA.
Transfer to M2 (49). When transfer is completed, work R
RAM by compressing the data of AM2 (49) by the CPU 41
3 (48).

Next, the sizes of the original data stored in the RAM1 (46) and the compressed data stored in the RAM3 (48) are compared in S5.6. If the original data is larger than the compressed data, the next processing is performed. Proceed to. If the size of the current data is smaller than that of the compressed data, the compressed data in the RAM3 (48) is discarded and the original data in the RAM2 (49) is copied as it is in S5.7. Next decided Li
The processing of S5.2 to S5.8 is repeated until the reading is completed in S5.8 by reading the ne number. The above is a brief description of the conventional example.

[0006]

However, in the conventional example, in the worst case, the transfer must be performed twice in S5.5 and S5.7, which is inefficient, and the reading time for one line is determined. If so, the read time may not be reached. Even with the method disclosed in Japanese Patent Application Laid-Open No. 62-259183, there is no guarantee that a part of the compression rates used in the test will match the overall compression rate. If the compression ratios are significantly different,
It also had the drawback of being counterproductive.

[0007]

The information processing apparatus of the present invention is, in an information processing apparatus having at least one type of compression method, means for measuring the appearance frequency of a specific data pattern in original data, and the data It is characterized in that the compression method is changed in accordance with the number of measured patterns to perform compression.

The means for measuring the appearance frequency includes means for transferring data from area 1 to area 2 of the storage device,
And a means for measuring the appearance frequency of the specific pattern provided in the transfer means.

[0009]

In contrast to the configuration of the conventional example, a new means is provided in the DMA to measure the frequency of the specific data pattern appearing in the original data being transferred, and the compression method is selected according to the number of measurements. The data in the storage area is compressed and stored in the RAM 3 (48).

[0010]

An embodiment of the present invention will be described in detail below with reference to the drawings. FIG. 1 is a schematic configuration diagram of an image reading apparatus according to an embodiment of the present invention. 4, which is a configuration diagram of a conventional example, includes a CPU 11, a CPU 41, and a ROM 13.
And ROM 43, binarization circuit 14 and binarization circuit 44, RA
M1 (14) and RAM1 (44), RAM2 (19) and RAM2 (49), DMA17 and DMA47, RAM3
(18) corresponds to RAM3 (48). The DMA 17 transfers the image data of the RAM 1 to the RAM 2 according to the instruction from the CPU 41. At the same time, it also has a function to measure how many "00" were in the transferred data. C
After the transfer is completed, the PU 41 can read the value obtained by measuring the appearance frequency by the DMA and can also clear the value.

In the conventional example, the CPU 41 uses the RAM 1 (46)
The image data of RAM1 (46) was compressed and stored in RAM3 (48).
After transferring to RAM2 (19) by A17, RAM2
The data of (19) is compressed and stored in the RAM 3 (18).

FIG. 2 shows an image reading apparatus R according to an embodiment.
It is a figure which shows the flow of the image reading compression program memorize | stored in OM13. FIG. 5 is a flowchart of a conventional example.
And S2.1 and S5.1, S2.2 and S5.2, S
2.3 and S5.3, S5.8 and S2.8, S5.9 and S
2.9 is supported.

From S2.1 to S2.3, the conventional S
Same as 5.1 to S5.3, but when the data is ready in S5.3, the CPU 11 sets D in S2.4.
RAM1 (16) to RAM2 (1
9) is instructed to transfer the data, and the DMA 17 performs the transfer. At the same time, the number of "00" data in the transfer data is measured. If the number is less than half of the transferred data in S2.5, the CPU 11 determines the RAM 2
The data is copied from (19) to the RAM 3 (18) without compression. If it is half or more, the CPU 11 determines that the RAM 2 (1
The data of 9) is compressed and stored in the RAM 3 (18).
The processing from S2.2 to S2.8 is repeated until the reading is completed in S2.8.

In this embodiment, RAM2 (19) and RAM3
Although (18) is divided, it is also possible to use one RAM and allocate different areas to RAM2 (19) and RAM3 (18).

Next, the compression method used in this embodiment will be described with reference to FIG. 31 is an example of the current data, and 3
2 is the compressed data. The first byte of the compressed data is a compression flag that determines whether the data is compressed data or uncompressed data, and if 0, the subsequent data is uncompressed data.
If 1, it is compressed data. The second byte is a byte indicating the length of compressed data. In this embodiment, 1 byte is allocated, but 2 bytes to several bytes can be allocated. Next, regarding compressed data, this is a method of combining data other than "00" with a pointer and data. If there is data "n2" in the n1st byte, "n"
Expressed as a combination of 1 "and" n2 ". For example, if the first byte of the original data contains" 11 ", it becomes" 01 "and" 11 ". Is 200%, and if "00" in the original data is less than half, the compressed data will be larger than the size of the original data, so "00" in S2.5.
If is less than half of the transferred data, it is compressed, and if it is less than half, it is uncompressed.

In this embodiment, the next line is not read until the compression of one line of data is completed. However, when the DMA transfer is completed in S2.4, the read command is issued to read and compress the image. By simultaneously performing the above, higher speed processing is possible. If the binarization circuit 14 and the image sensor 15 have sufficiently high functions, R
DM directly from the binarization circuit 14 except AM1 (16)
A configuration in which data is transferred to the RAM 2 (19) via A17 is also possible.

In this embodiment, the binarized image data is the target of the original data, but in a storage device such as a hard disk, the appearance frequency of a specific pattern in the data transferred from the host side by SCSI or the like is measured. However, it is effective for applications such as the controller compressing and storing in a drive, and other various data to be compressed.

In this embodiment, the specific pattern is "00".
The means for measuring the
Can also measure while transferring data from RAM1 (16) to RAM2 (19).
It is also possible to prepare a counter inside.

In this embodiment, the specific pattern is one pattern having a length of 1 byte "00", but "FF", "A5".
Or "0000" / "FFFF" / "1234567
8 ”multi-byte pattern check method, multiple patterns prepared, total appearance frequency or each appearance frequency check method, and pattern for measuring appearance frequency.
By providing a means for arbitrarily selecting or changing the length or the like by setting the length or the like in a register or the like, it is possible to change the pattern to be measured according to the situation.

In this embodiment, the number of branch points of compressed or uncompressed is half, but a method of changing the branch points according to the division method such as 1/3, 2/3 or 20 bytes and the circumstances before and after is also possible. Is.

In the present embodiment, only two types, compressed and uncompressed, were selected, but it is also possible to prepare a compression system 1, a compression system 2 and a compression system 3 and select which system according to the number of specific patterns. It is possible.

In this embodiment, the compression result is stored in the RAM 3
Although stored in (18), it is also possible to transmit the compressed data to another information processing device or the like by providing a communication means.

[0023]

As described above, according to the present invention, it is possible to judge the effect of compression at an early stage, minimize unnecessary time, and enable flexible compression.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an image processing apparatus of this embodiment.

FIG. 2 is a flowchart of a compression algorithm of the image processing apparatus of this embodiment.

FIG. 3 is an explanatory diagram of a compression method used in this embodiment.

FIG. 4 is a configuration diagram of a conventional image processing apparatus.

FIG. 5 is a flowchart of a compression algorithm of a conventional image processing apparatus.

[Explanation of symbols]

 11: CPU 12: Bus 13: ROM 14: Binarization circuit 15: Image sensor 16: Image storage RAM1 17: DMA with data pattern appearance frequency measurement function 18: Compressed data storage RAM3 19: Work RAM2 31: Original data 32: Compressed data sample diagram 41: CPU 42: Bus 43: ROM 44: Binarization circuit 45: Image sensor 46: Image storage RAM1 47: DMA 48: Compressed data storage RAM3 49: Working RAM2

Claims (2)

[Claims] 1. An information processing apparatus having at least one type of compression method, wherein a compression method is changed in accordance with a means for measuring the frequency of appearance of a specific data pattern in original data and the number of measurements of said data pattern. An information processing device characterized by compression. 2. The information processing apparatus according to claim 1, wherein the means for measuring the appearance frequency is means for transferring data from area 1 to area 2 of the storage device, and a specification provided in the transfer means. An information processing apparatus comprising: a unit that measures the frequency of appearance of a pattern.