JP2000134459A - Image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To convert a compressed image by rotating the image in a frequency region so as to reduce a calculation quantity and to convert to high speed mode when a direction of the compressed image is changed through rotation without converting the image compressed through frequency conversion into an image in a real space. SOLUTION: External compressed image data are received (S1). Quantized frequency image data are extracted by conducting inverse entropy coding (S2). Rotation processing of the quantized frequency image data is executed (S3). First the image data are moved in a rotating direction and rearranged for each block of components of the quantized frequency image data (S3(1)). Processing of top bottom inversion is conducted as required by moving the image data in the rotating direction and rearranging the data as to internal data of each block configuring the quantized frequency image data (S3(2)). After the rotation processing is conducted in a state of the quantized frequency image data, entropy coding is conducted to generate compressed image data (S4) and the compressed image data in the rotated state are outputted (S5).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image processing method, and more particularly to an image processing method capable of performing a suitable rotation process after compression.

[0002]

2. Description of the Related Art JPEG (Joint Photographic Experts Group) is a technique for compressing digital image data.
Etc. are known.

[0003] To perform such compression processing of digital image data, the digital image data is divided into a plurality of blocks, and the blocks are subjected to DCT (Descrete Cosine Transform).
m: Discrete cosine transform) to convert the frequency image data into frequency domain image data in the frequency domain, etc., ・ Convert the frequency image data to quantized frequency image data by referring to the quantization table, ・ Encode the quantized frequency image data Thus, compressed image data is generated.

In order to restore the compressed image data, the original image data can be taken out by going through the reverse procedure. The image formats using the JPEG compression described above include JPEG (JFIF), Exif,
CIFF, FlashPix, and the like are widely used as compression methods for natural images.

[0005]

In an image format created by a digital camera such as Exif or many scanners, image data having a fixed aspect ratio is set regardless of the vertical and horizontal directions set by a user. It is recorded as.

For this reason, image data or the like photographed by the user in the vertical position is not always displayed in the intended direction when the user opens the image file using software compatible with JPEG.

For this reason, when the image file is opened, the user needs to rotate the image data by 90 degrees. However, since browsing software (hereinafter, referred to as a browser) used on the Internet or the like cannot rotate an image, an image that should be in a vertical position must be observed in a horizontal position, which is very inconvenient.

[0008] In the above-mentioned Exif, tag information for designating the direction of an image exists, but there are few software programs that change the direction of an image using this tag information.

Further, in many cases, actual image data is compressed. To recompress and save such an image after reorienting the image in a normal direction, it is necessary to convert the compressed image into a real image by decompressing the real image. A series of processes such as rotation of the image and compression of the rotated actual image are required.

[0010] As for the rotation of an image, a method for rotating an actual image is described as described in US Patent No. 5,111,192. For example, FIG. 11 shows how JPEG-format compressed image data is rotated. Here, the compressed image data is converted into quantized frequency image data by the inverse entropy encoding unit 10, converted into frequency image data by the inverse quantization unit 20, and further converted into image data by the inverse frequency conversion unit 30. The processing so far is the same as the normal decompression of the compressed image data.

Then, in this state, the rotation processing section 40 rotates the image data under the instruction of the control section 1. The rotated image data is converted to frequency image data by the frequency conversion unit 50, converted to quantized frequency image data by the quantization unit 60, and further entropy-encoded by the entropy encoding unit 70 to obtain compressed image data. The processing of the frequency conversion unit 50 to the entropy coding unit 70 is the same as the generation of normal compressed data.

Due to the processing described above, there is a problem that it takes a long time to decompress and compress. Further, since requantization is performed by the quantization unit 60 at the time of recompression, there is a problem that the image is deteriorated when the quantization coefficient is different from that when the compressed image data is first created.

Accordingly, an object of the present invention is to realize an image processing method capable of reducing the amount of calculation and performing high-speed conversion when the direction of a compressed image is changed by rotation, without causing image quality deterioration. It is.

[0014]

That is, the present invention for solving the above problems is as described below. (1) The invention according to claim 1 is a method for processing an image compressed using frequency conversion, wherein the image compressed using frequency conversion is not converted into an image in a real space. An image processing method characterized by rotating an image.

According to the present invention, since the image compressed by using the frequency conversion is rotated in the frequency domain without being converted into the image in the real space, the direction of the compressed image is changed by the rotation. In addition, it is possible to reduce the amount of calculation and perform high-speed conversion.

(2) According to a second aspect of the present invention, in a state where an image compressed using frequency conversion is quantized,
The image processing method according to claim 1, wherein the image is rotated.

According to the present invention, the image compressed by using the frequency conversion is rotated in the frequency domain in a state where the image is quantized. The amount can be reduced and conversion can be performed at high speed, and the image quality does not deteriorate.

(3) The invention according to claim 3 is the image processing method according to claim 1, wherein the image compressed using the frequency conversion uses a discrete cosine transform.

According to the present invention, since the image compressed using the frequency conversion by the discrete cosine transform is rotated in the frequency domain without being converted into the image in the real space, the direction of the compressed image is rotated. Therefore, when changing, it is possible to reduce the amount of calculation and perform high-speed conversion. That is, since the left and right portions of the actual image can be inverted by inverting the sign of the AC portion of the block component, the amount of calculation is reduced and the speed can be increased.

(4) The invention according to claim 4, wherein the image compressed by using the frequency conversion is rotated by changing the zigzag table in the frequency domain. This is an image processing method.

According to the present invention, the image is rotated in the frequency domain by changing the zigzag table in the frequency domain without converting the image compressed using the frequency conversion by the discrete cosine transform into an image in the real space. Therefore, when the direction of the compressed image is changed by rotation, the amount of calculation can be reduced and high-speed conversion can be performed. That is, since only a single zigzag table exists for an image, the speed can be increased.

(5) The invention according to claim 5 is characterized in that, for an image compressed using frequency conversion, the image is rotated by inverting the sign of the AC component of the data. Image processing method.

According to the present invention, the image is rotated in the frequency domain by inverting the sign of the AC component of the data without converting the image compressed using the frequency conversion by the discrete cosine transform into the image in the real space. Thus, when the direction of the compressed image is changed by rotation, the amount of calculation can be reduced and high-speed conversion can be performed.
That is, since the left and right portions of the actual image can be inverted by inverting the sign of the AC portion of the block component, the amount of calculation is reduced and the speed can be increased.

(6) The invention according to claim 6 is the image processing method according to claim 1, wherein the image compressed by using the frequency conversion uses a wavelet transform.

According to the present invention, the image compressed using the frequency transform by the wavelet transform is rotated in the frequency domain without being converted into the image in the real space. When changing, it is possible to reduce the amount of calculation and perform high-speed conversion. That is, in the wavelet transform, since there is no complex number component, the rotation can be easily performed.

(7) The image processing method according to the first aspect, wherein an image photographed by a digital camera is rotated. In the present invention,
Since the image is rotated in the frequency domain without converting the image compressed using the frequency conversion into an image in the real space, a desired image can be displayed at high speed in accordance with the shooting condition in the digital camera. Will be able to do it.

(8) The invention according to claim 8 is a step of displaying a plurality of compressed images in a list, a step of selecting an image to be rotated from the list, and a step of rotating the selected image by rotating the original image. 2. The image processing method according to claim 1, further comprising a step of executing a process.

According to the present invention, an image to be rotated is selected from the list of compressed images, so that the processing can be efficiently performed even when a large number of images exist.

(9) The image processing method according to the ninth aspect, wherein the images to be displayed in a list are reduced images attached to a compressed image. According to the present invention, since the image to be rotated is selected from the list of reduced images attached to the compressed image, the processing can be efficiently performed even when a large number of images exist. .

(10) The image processing method according to claim 8, wherein the images to be displayed in a list are reduced images created by low frequency components in a frequency domain. is there.

According to the present invention, the image to be rotated is selected from the list of reduced images created from the low-frequency components of the compressed image. Therefore, even if there are many images, the processing can be efficiently performed. You will be able to proceed.

(11) According to the eleventh aspect of the present invention, an image is formed based on image direction information recorded in an image file or image direction information recorded on a recording medium on which an image file is recorded. 2. The image processing method according to claim 1, wherein the image is rotated.

In the present invention, the image is rotated based on the image direction information recorded in the image file or the image direction information recorded on the recording medium on which the image file is recorded. Thus, a large amount of image files can be automatically processed.

(12) According to the twelfth aspect of the present invention, when the image is rotated, the information of the image direction recorded in the image file or the information of the image direction recorded on the recording medium on which the image file is recorded is recorded. The image processing method according to claim 11, wherein information is corrected.

In the present invention, since the information on the image direction recorded in the image file or the information on the image direction recorded on the recording medium on which the image file is recorded is corrected, the direction of the image is corrected. It is possible to avoid mistakes when rotating.

(13) According to a thirteenth aspect, when transferring image data from an image storage medium having a small storage capacity to an image storage medium having a larger capacity than the image storage medium, the image is rotated. 2. The image processing method according to claim 1, wherein:

According to the present invention, since the image is rotated when the image data is transferred to the image storage medium having a large storage capacity, a database of the image data in a normal direction can be created.

(14) The image processing method according to claim 1, wherein after the rotation of the image, the image is recompressed without changing a parameter relating to image quality.

According to the present invention, after the image is rotated, the image is recompressed without changing the parameter relating to the image quality. Therefore, the image quality of the recompressed image is maintained, and the image is correctly oriented in the same image quality as the original image. Can be obtained at high speed.

[0040]

Embodiments of the present invention will be described below in detail. First Embodiment First, referring to FIG. 1, an image processing apparatus 1 used in an embodiment of an image processing method of the present invention will be described.
00 will be described. FIG. 1 is a functional block diagram showing an overall electrical schematic configuration of an image processing apparatus 100 according to an embodiment of the present invention.

In FIG. 1, reference numeral 101 denotes a control unit for controlling the entire apparatus, and reference numeral 102 denotes an inverse entropy encoding unit for performing inverse entropy encoding to generate quantized frequency image data from compressed image data (JPEG image data). ,
Reference numeral 103 denotes a rotation processing unit which performs rotation processing on the quantized frequency image data in the frequency domain, and 104 performs entropy coding processing to generate compressed image data (JPEG image data) from the quantized frequency image data. It is an entropy coding unit.

Here, the operation of the embodiment of the present invention will be described with reference to the block diagram of FIG. 1, the flowchart of FIG. 2, and the explanatory diagrams of FIG. 3 and FIG. First, compressed image data is externally supplied to the image processing apparatus 100 in FIG. 1 (S1 in FIG. 2).

In this embodiment, it is assumed that each block is compressed image data (such as JPEG compressed image data) that is compressed by using DCT transform. Since the compressed image data is entropy-coded (Huffman-coded), the compressed image data is
Performs inverse entropy encoding. As a result, image data in the frequency domain quantized by the predetermined quantization table (this is called quantized frequency image data) is extracted (S2 in FIG. 2).

Next, rotation processing of the quantized frequency image data is executed (S3 in FIG. 2). Here, attention is paid to the fact that the compressed image data is divided into blocks when the compressed image data is created from the original image data.

Here, it is assumed that the block of the quantized frequency image data before rotation is composed of six blocks as shown in FIG. Each block is composed of, for example, 8 × 8 pixel data.

First, for each block constituting the quantized frequency image data, the image is moved in the rotating direction and rearranged (S3 in FIG. 2). In this case, the blocks in the state as shown in FIG. 3A are rearranged as shown in FIG.
If the rotation is degrees, the images are rearranged as shown in FIG. 3C. If the rotation is 270 degrees clockwise, the images are rearranged as shown in FIG. 3D.

If the blocks in the state as shown in FIG. 4A are to be flipped horizontally (to create a left and right mirror image), they are rearranged as shown in FIG. ), Rearrange as shown in FIG. 4 (c).

This operation is unnecessary when dealing with compressed data that is not divided into blocks such as JPEG. Next, if necessary, the data inside each block constituting the quantized frequency image data is subjected to a relocation inversion process such that the image is moved in the direction of rotation and rearranged (S3 in FIG. 2).

Here, as a processing method in the block, a process of transposing the image data in the block (here, inversion of the image corresponding to the transposition matrix (inversion centering on the upper left and lower right axes)) is performed. ), And a process of exchanging codes.

In the process of making the reversal, it is possible to actually move the parameters. However, in the case of JPEG compression, pay attention to the fact that the data in each block of the image data is arranged in accordance with the zigzag table. The calculation process can be simplified by changing the zigzag table. It should be noted that, for a JPEG image, the zigzag table is 1 for each image, so that this processing can be executed at high speed.

By changing the zigzag table, the scanning direction of the block data is changed as shown in FIG. 5 or FIG. As a process of making a mirror image, in the case of DCT, a left-right or up-down mirror image conversion can be performed by changing the sign of the AC frequency component while leaving the DC frequency component as it is. Such processing can be performed at an extremely high speed as compared with the processing of replacing real data in the real space. It should be noted that a specific example of the transfer reversal process will be described in detail separately.

As described above, after the rotation processing is performed in the state of the quantized frequency image data, entropy coding is performed again to generate compressed image data (S3 in FIG. 2), and the rotated compressed image data is output (S3 in FIG. 2). (FIG. 2S5).

As described above, in the present invention, the image compressed using the frequency conversion is not converted into the image in the real space, but the image is rotated in the frequency domain. Therefore, when changing, it is possible to reduce the amount of calculation and perform high-speed conversion.

Further, according to the present invention, the image compressed by using the frequency conversion is rotated in the frequency domain in a quantized state. Therefore, when the direction of the compressed image is changed by rotation, , The amount of calculation can be reduced and conversion can be performed at high speed, and image quality does not deteriorate.

Further, in the present invention, since the image compressed using the frequency conversion by the discrete cosine transform is rotated in the frequency domain without being converted into the image in the real space, the direction of the compressed image is changed. When the rotation is changed by rotation, the amount of calculation can be reduced and the conversion can be performed at high speed. That is, since the left and right portions of the actual image can be inverted by inverting the sign of the AC portion of the block component, the amount of calculation is reduced and the speed can be increased.

Further, in the present invention, the image compressed in the frequency domain by changing the zigzag table in the frequency domain without converting the image compressed using the frequency conversion by the discrete cosine transform into the image in the real space is rotated. Therefore, when the direction of the compressed image is changed by rotation, the amount of calculation can be reduced and high-speed conversion can be performed. That is, since only a single zigzag table exists for an image, the speed can be increased.

Further, according to the present invention, the AC
Since the left and right portions of the real image can be inverted by inverting the sign of the portion, the amount of calculation is reduced and the speed can be increased. here,
The rotation process will be described using a specific example.

(1) Left and right mirror images: In the case of left and right mirror images, as shown in FIG. 4B, blocks are rearranged in the quantized frequency image data. Then, a process of inverting the sign of the AC component is performed for each block of the quantized frequency image data. Assuming that the original state of the block is as shown in FIG. 7, by inverting the sign of the AC component as shown in FIG. 8, left and right inversion within the block is realized.

(2) Rotation of 90 degrees clockwise: 90 clockwise
In the case of the degree rotation, as shown in FIG.
Block rearrangement is performed on the quantized frequency image data in the degree direction. Then, the zigzag table is changed for each block of the quantized frequency image data as shown in FIG. 5 or FIG. Then, the zigzag table is changed as shown in FIG.

(3) Rotation 180 degrees clockwise: 1 clockwise
In the case of the 80-degree rotation, as shown in FIG. 3C, the blocks are rearranged in the quantized frequency image data in the right 180-degree direction. Then, for each block of the quantized frequency image data, a change in the horizontal inversion of the zigzag table as shown in FIG. 8 and a change in the vertical inversion of the zigzag table as shown in FIG. 9 are performed. As a result, FIG. The sign of the AC component for exchanging the upper, lower, left and right as shown is inverted. That is, as shown in FIG.
By inverting the sign of the component, upside down, left and right in the block is realized.

(4) Rotation 270 degrees clockwise: clockwise 2
In the case of 70-degree rotation, as shown in FIG. 3D, block rearrangement is performed on the quantized frequency image data in the right 270-degree direction. Then, for each block of the quantized frequency image data, the zigzag table is inverted upside down as shown in FIG. 9 to invert the sign of the AC component.

By performing the above (1) to (3),
When the direction of the compressed image is changed by rotation, the amount of calculation can be reduced and high-speed conversion can be performed. <Second Embodiment> For example, in Exif version 2.0 which is an extended format of JPEG,
Information on the orientation of the image is recorded. Using this, if the orientation of the image is different from normal, the above-described rotation of the image is performed and the information of the image orientation is also corrected. Such an operation searches all the image data in a certain disk volume and automatically performs this processing. Note that efficient processing can be performed by automatically performing this processing in the background. By performing such processing, an image in a normal direction can be obtained by software that simply reads only JPEG.

<Third Embodiment> When there is no image orientation information, a list of images is displayed for each folder, and the operator is allowed to specify a film frame to be subjected to image rotation processing. . The image at this time uses data of a reduced image called a thumbnail attached to the original image, so that the operator can make a judgment. For an image for which rotation processing has been designated, rotation processing is performed on the original image data. By doing so, the display operation is quick and efficient.

<Fourth Embodiment> Also, an image compressed using frequency conversion by wavelet transform is rotated in the frequency domain without being converted into a real space image. When the wavelet transform is used, the display operation can be sped up by using a reduced image created from the low frequency components of the compressed image.

<Fifth Embodiment> When image data captured by a digital camera or the like and captured in a memory card is stored, and the stored image data is copied or moved to a large-capacity recording medium, By performing the above-described rotation processing, it becomes possible to create an image database using image data in a normal direction.

<Sixth Embodiment> When the original image is decompressed and displayed by changing its direction, the above processing can be applied. That is, it is faster to invert the sign of the AC component, and invert the sign of the AC component by inverting the zigzag table. Also, this operation may be partially performed and combined with the processing on the actual image.

<Other Embodiments> In each of the above embodiments, the AC component has been described. The DC component may be inversely encoded, and the DC component may be moved in blocks as shown in FIGS.

[0068]

As described above, the present invention has the following effects. (1) According to the first aspect of the present invention, since an image compressed using frequency conversion is not converted into an image in a real space, the image is rotated in a frequency domain.
When the direction of the compressed image is changed by rotation, the amount of calculation can be reduced and high-speed conversion can be performed.

(2) According to the second aspect of the present invention, the image is rotated in the frequency domain in a state where the image compressed using the frequency conversion is quantized. , The amount of calculation can be reduced and conversion can be performed at high speed, and image quality does not deteriorate.

(3) According to the third aspect of the present invention, the image is rotated in the frequency domain without converting the image compressed using the frequency conversion by the discrete cosine transform into the image in the real space. When the direction of the compressed image is changed by rotation, the amount of calculation can be reduced and high-speed conversion can be performed. That is, since the left and right portions of the actual image can be inverted by inverting the sign of the AC portion of the block component, the amount of calculation is reduced and the speed can be increased.

(4) According to the fourth aspect of the present invention, the zigzag table in the frequency domain is changed without converting the image compressed using the frequency conversion based on the discrete cosine transform into the image in the real space. Since the image is rotated in the region, when the direction of the compressed image is changed by rotation, the amount of calculation can be reduced and high-speed conversion can be performed. That is, since only a single zigzag table exists for an image, the speed can be increased.

(5) According to the fifth aspect of the present invention, the sign of the AC component of the data is inverted without converting the image compressed using the frequency conversion by the discrete cosine transform into the image in the real space. Since the image is rotated in the frequency domain, when the direction of the compressed image is changed by rotation, the amount of calculation can be reduced and high-speed conversion can be performed. That is, since the left and right portions of the actual image can be inverted by inverting the sign of the AC portion of the block component, the amount of calculation is reduced and the speed can be increased.

(6) In the invention described in claim 6, since the image compressed using the frequency transform by the wavelet transform is not converted into the real space image, the image is rotated in the frequency domain. When the direction of the compressed image is changed by rotation, the amount of calculation can be reduced and high-speed conversion can be performed. That is, in the wavelet transform, since there is no complex number component, the rotation can be easily performed.

(7) According to the seventh aspect of the present invention, since the image compressed in the frequency domain is rotated without converting the image compressed using the frequency conversion into the image in the real space, the digital camera can A desired image can be displayed at high speed in accordance with the state of shooting.

(8) In the invention described in claim 8, the image to be rotated is selected from the list of compressed images, so that the process can be efficiently performed even when a large number of images exist. Will be able to

(9) According to the ninth aspect of the present invention, the image to be rotated is selected from a list of reduced images attached to the compressed image, so that even when a large number of images are present, it is efficient. The process can proceed to

(10) According to the tenth aspect, the image to be rotated is selected from a list of reduced images created from the low-frequency components of the compressed image. Thus, the processing can be efficiently performed.

(11) According to the eleventh aspect of the present invention, an image is formed based on image direction information recorded in an image file or image direction information recorded on a recording medium on which an image file is recorded. Since the rotation is performed, a large amount of image files can be automatically processed.

(12) According to the twelfth aspect, the information on the image direction recorded in the image file or the information on the image direction recorded on the recording medium on which the image file is recorded is corrected. This makes it possible to avoid mistakes when rotating the direction of the image.

(13) In the invention according to claim 13, since the image is rotated when the image data is transferred to the image storage medium having a large storage capacity, a database of image data in a normal direction is created. It becomes possible to do.

(14) In the invention according to claim 14, since the image is rotated and recompressed without changing the parameters related to the image quality, the image quality of the recompressed image is maintained and the original image is maintained. It is possible to obtain an image of the same image quality and the correct direction at high speed.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an image processing apparatus used in an embodiment of an image processing method according to the present invention.

FIG. 2 is a flowchart illustrating a processing procedure according to the embodiment of the present invention.

FIG. 3 is an explanatory diagram showing a processing state of block rearrangement in the embodiment of the present invention.

FIG. 4 is an explanatory diagram showing a processing state of block rearrangement in the embodiment of the present invention.

FIG. 5 is an explanatory diagram showing a state of a zigzag table according to the embodiment of the present invention.

FIG. 6 is an explanatory diagram showing a state of a zigzag table according to the embodiment of the present invention.

FIG. 7 is an explanatory diagram showing a state of processing in a block according to the embodiment of the present invention.

FIG. 8 is an explanatory diagram showing a state of processing in a block according to the embodiment of the present invention.

FIG. 9 is an explanatory diagram showing a state of processing in a block according to the embodiment of the present invention.

FIG. 10 is an explanatory diagram showing a state of processing in a block according to the embodiment of the present invention.

FIG. 11 is a block diagram illustrating a configuration of a conventional image processing apparatus that performs image rotation.

[Explanation of symbols]

 Reference Signs List 101 control unit 102 inverse entropy encoding unit 103 rotation processing unit 104 entropy encoding unit

Continued on the front page F-term (reference)

Claims (14)

[Claims] 1. A method for processing an image compressed using frequency conversion, wherein the image is rotated in a frequency domain without converting the image compressed using frequency conversion into a real space image. An image processing method characterized by the following. 2. The image processing method according to claim 1, wherein the image is rotated while the image compressed using the frequency conversion is quantized. 3. The image processing method according to claim 1, wherein the image compressed using the frequency conversion uses a discrete cosine transform. 4. The image processing method according to claim 3, wherein the image compressed by using the frequency conversion is rotated by changing a zigzag table in a frequency domain. 5. The image processing method according to claim 3, wherein the image is rotated by inverting the sign of the AC component of the data with respect to the image compressed using the frequency conversion. 6. The image processing method according to claim 1, wherein the image compressed by using the frequency conversion uses a wavelet transform. 7. The image processing method according to claim 1, wherein an image taken by a digital camera is rotated. 8. A method comprising: displaying a plurality of compressed images in a list; selecting a rotation target image from the list; and executing a rotation process on an original image of the selected image. 2. The image processing method according to claim 1, wherein: 9. The image processing method according to claim 8, wherein the images to be displayed in a list are reduced images attached to a compressed image. 10. The image processing method according to claim 8, wherein the images to be displayed in a list are reduced images created by low-frequency components in a frequency domain. 11. An image rotating device based on image direction information recorded in an image file or image direction information recorded on a recording medium on which an image file is recorded. Item 10. The image processing method according to Item 1. 12. The method according to claim 1, wherein when rotating the image, the information on the image direction recorded in the image file or the information on the image direction recorded on the recording medium on which the image file is recorded is corrected. The image processing method according to claim 11, wherein 13. The method according to claim 1, wherein the image is rotated when transferring image data from an image storage medium having a small storage capacity to an image storage medium having a larger capacity than the image storage medium. The image processing method described in the above. 14. The image processing method according to claim 1, wherein after the rotation of the image, the image is recompressed without changing a parameter relating to image quality.