JPH05260315A - Data recovery device - Google Patents

Abstract

PURPOSE:To reduce the memory capacity for a data storing device which expands and restores the compressed data on the color static images. CONSTITUTION:A decoding circuit 11 outputs and stores the data obtained by adding a single bit serving also as an underflow/overflow flag to the decoding data having the prescribed number of bits into a memory 12. Therefore the bit with of the memory 12 can be reduced by one bit compared with a conventional case where a single bit is required exclusively for each of underflow and overflow flags.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a data restoration device,
In particular, it relates to a data decompression device for decompressing and decompressing compressed data of a color still image.

The band compression encoding of a color still image is being internationally standardized so that it can be applied to a wide range of applications such as a print image, a captured image of an electronic still camera, and a color fax, and a discrete cosine transform (DCT).
Adaptive Discrete Cosine Tr (ADCT) that combines Transfrom) and Huffman coding
ansform) coding scheme has been proposed.

In this ADCT coding method, image data is two-dimensionally DCTed, linearly quantized, and then Huffman coded to obtain compressed data. Huffman decoding is performed in order to perform quantization during the compression process. Even if decoding is performed in the order of inverse quantization and inverse DCT, the original image cannot be completely restored, but the difference between the original image and the restored image cannot be discriminated by human eyes by setting the conditions. You can

A compressed image data coded by this ADCT coding system may be further provided with a hierarchical coding function called progressive coding.
This progressive coding is a method that first codes a rough overall image and then encodes the image so that the resolution gradually becomes finer. The resolution increases with the passage of. Thereby, the outline of the image can be grasped in a short time, and the efficiency of searching the image database can be greatly improved.

In such an operation for decompression / decompression of progressively encoded compressed image data, an overflow or underflow may occur in which the grayscale value exceeds the upper limit value or the lower limit value. Flow without increasing memory capacity
Need to be identified.

[0006]

2. Description of the Related Art FIG. 3 is a block diagram showing an example of a conventional data restoration device. In the figure, the progressively-encoded compressed image data is the decompression large-scale integrated circuit (LS).
I) is input to 1 and is decompressed (decoded) here. That is, the decompression LSI 1 decodes the input compressed image data by referring to the Huffman coding table, performs dequantization by referring to the quantization table, and further performs inverse DCT operation to obtain ADCT decoded data.

The ADCT decoded data consists of three types of 10-bit data including 8-bit image data, 1-bit underflow bit, and 1-bit overflow bit. That is, there are three types of 8-bit image data, for example, for luminance signals and two types of color signals, so that the bit width of one pixel is 30 bits. The underflow bit is "1" when an underflow has occurred, "0" when it has not occurred, and the overflow bit is "1" when an overflow has occurred and "0" when it has not occurred.

The 10-bit decoded data relating to the luminance signal is input to the image memory 2 _Y, and the 10-bit decoded data relating to the two types of color signals is input to and stored in the image memories 2 _R and 2 _B , respectively. It The decoded data read out from the image memories 2 _Y , 2 _R and 2 _B , respectively, are supplied to the overflow / underflow processing circuits 3 _Y , 3 _R and 3 _B , where based on the values of the overflow bit and the underflow bit. When there is an overflow, the image data is changed to the upper limit value of the gradation value, and when there is an underflow, it is changed to the lower limit value of the gradation value.

The 8-bit image data respectively taken out from the overflow / underflow processing circuits 3 _Y , 3 _R and 3 _B are supplied to a matrix circuit 4 where red (R) and green are calculated by a known matrix operation. After being converted into three types of primary color data of (G) and blue (B), it is output to a monitor device (not shown).

[0010]

However, in the above-described conventional apparatus, the decoded data supplied to the image memories 2 _Y , 2 _R and 2 _B are 10 bits each, whereas the commercially available image memory is used. Since the bit width is usually only 8 bits or 9 bits, two 8-bit width memories must be used as each image memory in order to store the above 10-bit decoded data. It is uneconomical.

The present invention has been made in view of the above points,
It is an object of the present invention to provide a data restoration device that solves the above problems by representing the underflow bit and the overflow bit by one bit.

[0012]

FIG. 1 shows a block diagram of the principle of the present invention. As shown in the figure, the present invention is a device for restoring data that may cause an overflow exceeding an upper limit value and an underflow exceeding a lower limit value at the time of data decoding, when either an underflow or an overflow occurs, Decoding circuit 11 that adds a 1-bit flag having a predetermined logical value to the decoded data and outputs the decoded data.
And a memory 12 for storing the output of the decoding circuit 11, and a judging means 13 for judging whether an underflow or an overflow has occurred in the decoded data based on the output of the memory 12.

[0013]

It is a well known fact that changes in brightness and color of natural images are generally smooth. This means that if the above data is natural image data, it will not overflow immediately after the decoded data underflows,
It also shows that underflow does not occur immediately after overflow.

Further, by comparing the values of the decoded data between the adjacent values before and after, the direction of change of the value of the decoded data can be known. Therefore, in the present invention, attention is paid to the above points, and one bit also serves as a flag indicating whether overflow or underflow occurs in decoded data.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 2 shows a block diagram of an embodiment of the present invention. In the figure, the decompression LSI 21 corresponds to the decoding circuit 11,
Similar to the decompression LSI 1 shown in FIG. 3, the progressive-coded compressed image data is decoded, and the luminance signal Y, the first color signal C _R , and the second color signal C _{B are obtained for each image.}
The decoded data composed of the image data and the flag is output in this order. However, unlike the decompression LSI 1, the decompression LSI 21 has a 1-bit flag for identifying underflow / overflow added to each 8-bit image data.

Therefore, 8-bit image data and 9-bit decoded data consisting of a 1-bit underflow / overflow flag are extracted from the decompression LSI 21 and supplied to the image memory 22 for storage. Here, as described above, the decoded data from the decompression LSI 21 is a total of three types for the luminance signal and two types of color signals, and 8-bit image data relating to the luminance signal and 1-bit underflow / overflow for each pixel. 9-bit data consisting of a dual-purpose flag, 8-bit image data relating to the first color signal C _{R and} 9-bit data comprising a 1-bit underflow / overflow dual-purpose flag, and 8-bit image relating to the second color signal C _B 9-bit data consisting of data and a 1-bit underflow / overflow combined flag is taken out, and these three types of data are individually restored by a dedicated circuit section. In the embodiment of FIG. For simplicity of illustration, only one system of the circuit section is shown.

The decoded data written in the image memory 22 includes 8 bits of image data and underflow /
Since the overflow / combined flag is 1 bit, which is a total of 9 bits, the image memory 22 can be composed of one commercially available memory having a bit width of 9 bits.

The image data and the underflow / overflow dual-use flag read from the image memory 22 for displaying the image data are supplied to the image data / flag bit division circuit 23 to be divided into the image data and the flag. The data is stored in the previous value data storage register 24 and delayed by one pixel data transfer period.

Image data / flag bit division circuit 23,
The underflow / overflow determination circuit 25, which constitutes the determination means 13 together with the previous value data storage register 24, receives the image data extracted from the image data / flag bit division circuit 23 and the underflow / overflow dual-use flag. On the other hand, the image data extracted from the previous value data storage register 24 is input.

The underflow / overflow dual-use flag is set to "1" in both cases of occurrence of underflow and overflow of image data, and "0" otherwise. It Therefore, when the value of the underflow / overflow dual-use flag is "1", it is necessary to determine which of underflow and overflow has occurred.

Therefore, the underflow / overflow determination circuit 25 compares the value of the image data from the dividing circuit 23 with the value (previous value) of the image data from the register 24 one pixel before. When the gradation value has changed to a value close to the maximum value or more, and underflow /
When the value of the overflow combined use flag is "1", it is determined that an overflow has occurred, and when the gradation value has changed to a value close to the minimum value or more by a predetermined value and the value of the underflow combined use / overflow combined use flag is "1". When, it is determined that an underflow has occurred.

The switch circuit 26 receives the image data from the dividing circuit 23 at a terminal a, and the underflow data from an underflow data generator 27 at a terminal b (here, as an example, the minimum value "00" in hexadecimal). ”), The overflow data from the overflow generator 28 (here, the maximum value of hexadecimal value“ FF ”) is input to the terminal c, and according to the determination output from the underflow / overflow determination circuit 25. , An input signal of one of these terminals a to c is selected and terminal d is selected.
Output as an image output to.

That is, the switch circuit 26 selectively outputs the image data from the division circuit 23 when the determination result of the underflow / overflow determination circuit 25 indicates that neither underflow nor overflow has occurred. On the other hand, the switch circuit 26 generates the underflow data from the underflow data generator 27 when the determination result of the underflow / overflow determination circuit 25 indicates underflow, and generates the overflow data when it indicates overflow. The overflow data from the container 28 is selected and output.

As described above, according to this embodiment, the image memory, which has conventionally required two memories having a bit width of 8 bits, is
It can be configured with one memory having a bit width of 9 bits. This means that when the data length of each of Y, C _R , and C _B is 8 bits and the image size is 512 pixels in the horizontal direction and 400 pixels in the vertical direction, the cell size of the conventional image memory is 512 × 25.
9 in total memory 8-bit width _{6 (Y, C R, C} B
However, according to the present embodiment, a total of 6 memories each having a cell size of 512 × 256 bits and a 9-bit width (2 for each of Y, C _R , and C _B ) are required. It is only necessary to prepare the image memory, and the image memory can be configured with a memory that is ⅔ times as many as the conventional memory.

[0025]

As described above, according to the present invention, since the flag indicating whether overflow or underflow has occurred in the decoded data is also used by one bit, the number of bits of image data can be reduced. The memory capacity can be significantly reduced without reduction, and the device can be configured at low cost without degrading image quality.

[Brief description of drawings]

FIG. 1 is a principle block diagram of the present invention.

FIG. 2 is a block diagram of an embodiment of the device of the present invention.

FIG. 3 is a configuration diagram of an example of a conventional device.

[Explanation of symbols]

 11 Decoding Circuit 12 Memory 13 Determining Means 21 Decompression LSI 22 Image Memory 23 Image Data / Flag Bit Division Circuit 24 Previous Value Data Storage Register 25 Underflow / Overflow Determination Circuit 26 Switch Circuit

Claims (3)

[Claims] 1. A device for restoring data which may cause an overflow exceeding an upper limit value and an underflow exceeding a lower limit value at the time of decoding data, and when the underflow or the overflow occurs, a predetermined logical value is set. A decoding circuit (11) for adding a 1-bit flag to the decoded data and outputting the decoded data, a memory (12) for storing the decoded data and the flag output from the decoding circuit, and an output decoded data and flag for the memory (12) Determination means (13) for determining whether underflow or overflow has occurred in the decoded data based on
A data recovery device comprising: 2. The judging means (13) comprises a dividing circuit (23) for respectively dividing the decoded data and the flag from the memory (12), and a register (24) for storing the decoded data from the memory (12). ) With the decoded data from the division circuit (23) as the previous value with the decoded data from the register (24), and the comparison result and the value of the flag from the division circuit (23) are compared to the predetermined logic. 2. The data restoration device according to claim 1, further comprising a determination circuit (25) for determining whether underflow or overflow has occurred depending on whether the value is a value or not. 3. The data restoration device according to claim 1, wherein the decoded data is progressively displayed image data that has been ADCT decoded.