JP2002159008A - Method and device for restoring image - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suppress the quality deterioration of a restored image that occurs when the total number of DC component data and AC component data in a block is abnormal. SOLUTION: An image restoring device is provided with a count data control circuit 50 which successively counts AC component data in order from the low frequency component side to the high frequency component side starting from DC component data at every block before increase DCT transformation is started and, in addition, while individual DC component data and AD component data constituting the blocks are restored and, at the same time, ignores the AC component data of a block when the counted number of the data to be contained in the block is a prescribed number or larger.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention divides a two-dimensional matrix of pixels constituting an original image into a plurality of rectangular blocks, and performs DCT (Discrete Cosine) for each block.
The present invention relates to an image restoring method used for restoring an image file whose data amount has been compressed after compression and then decompressing and inverse DCT transforming the image file into an image composed of a two-dimensional matrix of pixels. (Direct Cu
rrent) component data and AC (Alternating Current)
The present invention relates to an image restoration method and an image restoration apparatus capable of suppressing a decrease in the quality of a restored image which occurs when the total number of component data is abnormal.

[0002]

2. Description of the Related Art For transmission using communication means such as the Internet and storage using a storage medium such as a DVD (Digital Video Disc), JPEG (Joint Photo
An image file compressed in accordance with regulations such as the Tographic Experts Group) is used. By compressing image files to be transmitted or stored, transmission time can be shortened, the storage capacity of storage media can be maximized, and costs can be reduced. is there.

In JPEG and others, 2
The pixel group in a dimensional matrix is divided into a plurality of rectangular blocks, and the amount of data is compressed after DCT conversion for each block. For example, the data amount is reduced by Huffman coding, quantization processing, entropy coding, or the like. Specifically, in JPEG, Huffman coding is performed after DCT conversion to compress the data amount.

[0004] The image file compressed in this way is again restored to an image using a two-dimensional matrix of pixel groups by an image restoration method or an image restoration apparatus as the object of the present invention. Specifically, decompression processing corresponding to the compression processing is performed. After that, the image is expanded again and subjected to inverse DCT transform to restore the image to a two-dimensional matrix of pixels.

[0005]

However, the image file is transmitted after the data amount is compressed as described above,
When data is stored, an error may occur in the data for some reason. Also, due to the compression, if such an error occurs, not only data in the image file may be rewritten, but also an abnormality may occur in the number of data. That is, erroneous data may be generated or data may be lost, and the number of data may increase or decrease.

In JPEG or the like, a two-dimensional matrix of pixels forming an original image is divided into a plurality of rectangular blocks each having (8 × 8 = 64) pixels which are usually adjacent to each other. Thereafter, the processing is performed in block units.
That is, for each block, the data amount is compressed after DCT conversion to obtain an image file for transmission or storage. In addition, data obtained from the image file is expanded and inversely DCT-transformed for each block to restore the image to a two-dimensional matrix of pixels.

A block of (8 × 8 = 64) pixels is DCT
When converted, one DC component data and 63 AC
It becomes a block composed of a total of 64 data by component data. The block is again restored to an image of a block of (8 × 8 = 64) pixels by performing an inverse DCT transform.

In an image file compressed according to JPEG or the like, AC component data is run-length encoded. That is, data indicating the number of data whose consecutive values are "0" (the length of the data is continuous and called the run length), and data indicating the value of data other than "0" following the data indicating the value. It is encoded so as to be expressed by. Therefore,
If an error occurs in the data indicating the run length due to an error during transmission or storage as described above, the number of AC component data becomes 63 or more or decreases.

When the number of data increases, excess AC component data exceeding the prescribed number of data to be included in one block is erroneously replaced by DC component data or AC component data of the next block. It will be. Then, in the prior art, the decoded data is assumed to be naturally 64 data, and even if an error occurs, it is treated as it is. In such a case, an obstacle such as color shift occurs.

In the prior art, such obstacles could not be removed or reduced. Therefore, conventionally, when a failure such as an image shift or a color shift occurs, the image file is transmitted again or stored and read again. Alternatively, conventionally, processing or the like has been performed as it is.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned conventional problems.
It is an object of the present invention to provide an image restoring method and apparatus capable of suppressing a decrease in the quality of a restored image, which is caused when there is an abnormality in the total number of C component data.

[0012]

First, an image restoration method according to the first invention of the present application divides a two-dimensional matrix of pixels constituting an original image into a plurality of rectangular blocks, and performs DCT conversion for each block. In an image restoration method used to decompress an image file whose data amount has been compressed and then perform an inverse DCT transform to restore an image composed of a two-dimensional matrix of pixels, before performing the inverse DCT transform, At the stage where the individual DC component data and AC component data constituting the data are restored, the number of these data is counted for each block, and if there is an excess of data exceeding the specified number of data to be included in the block, , By ignoring the AC component data by the excess,
This has solved the above-mentioned problem.

Next, the image restoration apparatus of the second invention of the present application
An image file conforming to the JPEG specification, in which a two-dimensional matrix of pixels constituting an original image is divided into a plurality of rectangular blocks, DCT-converted for each block, and the amount of data is compressed, is decompressed again and inversed. In an image restoration apparatus used for restoring a reproduced image by a two-dimensional matrix pixel group by DCT transformation, individual DC component data and AC component data constituting a block are restored before inverse DCT transformation. At each stage, the AC component data is sequentially counted from the low-frequency component side to the high-frequency component side in order from the DC component data for each block. If data is counted, a count data control circuit that ignores the AC component data of the corresponding block is provided. By the, it is obtained by solving the above problems.

Hereinafter, the operation of the present invention will be briefly described.

According to the present invention, an image file in which a two-dimensional matrix of pixels constituting an original image is divided into a plurality of rectangular blocks, DCT-converted for each block, and the data amount is compressed is decompressed again. The present invention is directed to an image restoration method and apparatus used for performing an inverse DCT transform to restore an image by a two-dimensional matrix pixel group.

In the present invention, before performing the inverse DCT transform,
And the individual DC component data and A
When the C component data is restored, the number of these data is counted for each block. In this way, if there is an excess of data exceeding the prescribed number of data to be included in the block, the AC component data is ignored by the excess.

Therefore, according to the present invention, even if an error occurs during transmission or storage and the number of DC component data or AC component data in a block increases or decreases, the influence is less likely to be exerted on other blocks. Therefore, it is possible to suppress adverse effects when an error occurs.

As described above, according to the present invention, it is possible to suppress a decrease in the quality of a restored image which occurs when the total number of DC component data and AC component data in a block is abnormal.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a block diagram showing the overall configuration of an image restoration apparatus to which the present invention is applied and an image compression / encoding apparatus used together therewith.

In this figure, first, in order to transmit and save an image file, an image compression encoding apparatus that compresses and encodes an input image (original image) in accordance with the JPEG regulations is a blocking circuit. 12 and an encoding circuit 14
It is composed of The image file generated by the image compression encoding circuit is a JPEG image, and can be used for transmitting and storing the image.

First, the blocking circuit 12 divides a two-dimensional matrix pixel group constituting the original image into a plurality of rectangular blocks each having usually (8 × 8 = 64) pixels adjacent to each other. Output a block image.

Next, the encoding circuit 14 receives the block image, performs DCT conversion for each block, and generates an image file in which the data amount is compressed. The encoding circuit 14
Is composed of an F-DCT circuit 30, a quantization circuit (Quantizer) 32, a Q table 34, an entropy encoding circuit 36, and an E table 38, as shown in FIG.

The F-DCT circuit 30 has (8 × 8 = 64)
Blocks of pixel data into one DC component data,
Is converted into a block composed of a total of 64 data based on and 63 AC component data. The quantization circuit 32 performs a quantization process using the data of the Q table 34. In the entropy encoding circuit 36, an E table 38
Is used to compress the amount of data.

Next, referring to FIG. 1, an image restoration apparatus used for restoring an image file generated by the image compression / encoding circuit and restoring the image file into a two-dimensional matrix of pixels by performing inverse DCT transform is as follows. It comprises a decoding circuit 16 and an image construction circuit 18.

First, the decoding circuit 16 decompresses the image file generated by the image compression / encoding device again and performs inverse DCT.
Convert and output a blocked image.

Next, the image construction circuit 18 restores the block image to an image composed of a two-dimensional matrix of pixels and outputs an output image (restored image).

Here, the decoding circuit 16 comprises an entropy decoding circuit 40 and an E table 42 as shown in FIG.
, A count data control circuit 50, a dequantizer circuit (Dequantizer) 44, a Q table 46, an I-DC
And a T circuit 48.

The entropy decoding circuit 40 uses the E-table 42 to compress and expand the data amount. The E-table 42 stores the same data as the E-table 38 described above.

Conventionally, after the processing by the entropy decoding circuit 40, the processing by the inverse quantization circuit 44 is performed. However, in the present embodiment, the present invention is applied, and the entropy decoding circuit 40 and the inverse quantization Between the circuits 44, a count data control circuit 50 having an internal circuit as shown in FIG. 4 is provided.

The inverse quantization circuit 44 performs an inverse quantization process using the data of the Q table 46. The Q table 46 stores the same data as the Q table 34.

The I-DCT circuit 48 divides a block composed of a total of 64 data of one DC component data and 63 AC component data into (8 × 8 = 64) pixel data. Convert to a block.

FIG. 4 is a block diagram showing a configuration of the count data control circuit 50 described above.

As shown, the count / data control circuit 50 includes a switch 52, a counter circuit 54, and a reset determination circuit 56.

First, the switch 52 is normally turned on, and the data output from the entropy decoding circuit 40 is input to the inverse quantization circuit 44 as it is. When the switch 52 is turned off by a signal from the counter circuit 54, the entropy decoding circuit 40
The data output from is blocked by the inverse quantization circuit 4
4 is no longer input. That is, the data is ignored in the inverse quantization circuit 44.

The counter circuit 54 has a reset input R
Is reset to "0".
After the reset, the entropy decoding circuit 40 outputs the DC of each block which the entropy decoding circuit 40 sequentially outputs.
Each time the component data and the AC component data are input, the count value is incremented by "1". Then, when the count value becomes "64", the counter circuit 54 outputs a signal for turning off the switch 52 until the count value is reset to "0" by the input of the reset input R. .

When the start of inputting the DC component data is normally determined in each block, the reset determination circuit 56 outputs a signal to the reset input R of the counter circuit 54, resets the counter circuit 54, and resets the DC component. The switch 52 is turned on before data reaches the switch 52. The above-mentioned normal input is determined based on end data of the Huffman coded signal (end of the previous block), a marker inserted as needed in a JPEG image file, and the like.

In the present embodiment, before the inverse DCT transform, that is, before the I-DCT circuit 48 and at the stage where the individual DC component data and AC component data constituting the block are restored, At a stage subsequent to the entropy decoding circuit 40, the number of these data is counted for each block by the counter circuit 54 in the count / data control circuit 50. If there is an excessive amount of data exceeding the prescribed number of data to be included in the block, the switch 5 in the count data control circuit 50
By turning 2 off, the AC component data can be ignored by the excess.

The installation location of the count data control circuit 50 may be any stage before the inverse DCT transformation and at the stage where the individual DC component data and AC component data constituting the block are restored. Therefore, the present invention is not limited to the arrangement between the entropy decoding circuit 40 and the inverse quantization circuit 44 as shown in FIG. For example, as shown in FIG. 5, it may be provided between the inverse quantization circuit 44 and the I-DCT circuit 48.

As described above, according to the present embodiment, the present invention can be effectively applied. Therefore, it is possible to suppress the deterioration of the quality of the restored image, which is caused when the total number of DC component data and AC component data in the block is abnormal.

[0041]

According to the present invention, it is possible to suppress a decrease in the quality of a restored image that occurs when the total number of DC component data and AC component data in a block is abnormal.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an overall configuration of an image restoration apparatus to which the present invention is applied and an image compression / encoding apparatus used therewith.

FIG. 2 is a block diagram showing a configuration of an encoding circuit used in the embodiment.

FIG. 3 is a block diagram showing a configuration of a decoding circuit used in the embodiment.

FIG. 4 is a block diagram showing a configuration of a count data control circuit used in the embodiment.

FIG. 5 is a block diagram showing a configuration of a decoding circuit used in a modification of the embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 12 ... Blocking circuit 14 ... Encoding circuit 16 ... Decoding circuit 18 ... Image construction circuit 30 ... F-DCT circuit 32 ... Quantization circuit 34 ... Q table 36 ... Entropy encoding circuit 38 ... E table 40 ... Entropy decoding Circuit 42 E table 44 Inverse quantization circuit 46 Q table 48 I-DCT circuit 50 Count data control circuit 52 Switch 54 Counter circuit 56 Reset determination circuit

Claims (2)

[Claims] An image file in which a pixel group of a two-dimensional matrix constituting an original image is divided into a plurality of rectangular blocks, DCT-converted for each block, and an image file in which a data amount is compressed is decompressed again and inverse DCT-converted. In an image restoration method used to convert the image into a two-dimensional matrix pixel group, individual DC component data and AC component data constituting a block are restored before the inverse DCT transform. At the stage, the number of these data is counted for each block, and if there is an excess of data exceeding a prescribed number of data to be included in the block, the AC component data is ignored by the excess. Image restoration method. 2. An image file conforming to the JPEG specification, in which a two-dimensional matrix pixel group constituting an original image is divided into a plurality of rectangular blocks, DCT-converted for each block, and the data amount is compressed. In an image restoration apparatus used for decompressing and performing inverse DCT transform to restore a reproduced image by a two-dimensional matrix pixel group, before the inverse DCT transform, and individual DC component data and AC At the stage where the component data is being restored, the DC component data is sequentially counted from the low frequency component side to the high frequency component side for each block, and the AC component data is sequentially counted, and the rule to be included in the block is specified. Count data control circuit that ignores the AC component data of the corresponding block when the data of more than An image restoration apparatus characterized by comprising: