JP2003274154A - Image processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing method which minimizes the burden on interpolation processing with no sense of visible degradation without contradicting in image contents. <P>SOLUTION: In the image processing method for extracting data including frequency (AC) components of a plurality of partial pixel areas being a unit of compression when image data compressed with a plurality of pixels as a unit together with frequency (AC) components by using orthogonal conversion to a spatial frequency area is restored from the compressed data to the non- compressed data, feature quantities of extracted frequency (AC) components of a plurality of partial pixel areas are compared with a preliminarily set spatial frequency characteristic threshold. High frequency data blocks and low frequency data blocks are discriminated as frequency characteristics, the result is associated with position information of a plurality of partial pixel areas, at least two kinds of filters for varying magnifications of image data, which have different interpolation precisions, are prepared, and the filters are switched and used with a plurality of partial pixel areas as a unit on the basis of the obtained result when magnifications of image data are varied (image data is enlarged or reduced). <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image interpolation method, and more particularly, to a compressed image data format Jpegfile.
The present invention relates to an optimal method for enlarging / reducing the content of an image.

[0002]

[Prior Art] Jpeg file taken with a digital camera
An image may be printed by a PC printer, a direct printer, or the like, or may be printed by a DPE.
At this time, it is general that the number of pixels of the Jpeg file image to be printed does not match the number of image pixels required to print to the desired print size or print resolution.
Original Jpeg by print driver or application
Output is optimized by performing enlargement / reduction, such as interpolation, with a predetermined filter for all pixels of the file image. Also, as an interpolation method for creating printer data from a compressed image file using the DCT of digital data,
"JP-A-5-261982" is known.

This concrete method is described in Jpeg file, DC, etc.
With respect to a compressed image using T (discrete cosine transform), a desired resolution data is calculated by using an inverse DCT transform matching a designated print size.

As another example, in the case of "Japanese Unexamined Patent Publication No. 9-54825", the frequency data is changed by converting the frequency data in the frequency domain DCT-converted by using a constant prepared in advance. By performing inverse DCT conversion of frequency data in the frequency domain, a sharply corrected image is obtained.

[Problems to be Solved by the Invention] When printing a Jpeg file image taken with a digital camera or the like, the Jpeg image is printed.
Enlargement interpolation processing is performed to convert from File size to print size data, but this interpolation processing is uniformly high-order interpolation (bicubic Etc.) It is normal to execute processing.

Further, as a printing form, not only printing from a PC having a high processing capability, but also a printing environment using a device having a low data processing capability such as direct printing for directly printing from a digital camera to a printer has been formed. . This built-in printer has the problem that the high-order interpolation processing is heavy as a load in addition to the normal print processing, and the time from the print command to the start of print execution is long, resulting in poor usability. did. An interpolation method that does not impair the image quality of a natural image and that is light in overall processing is required. Among the above-mentioned conventional examples, "JP-A-5-261982" has a problem that the inverse DCT processing for performing resolution conversion is heavy, and in another conventional example, "JP-A-9-54825", there is a problem of image correction. Therefore, certain changes are made to the frequency component data, but in this case, the consistency of the joints in each block becomes a problem, and this problem also refers to the use of frequency components. However, it does not have functions such as pixel interpolation for scaling.

The present invention has been made in view of the above problems, and an object of the present invention is to minimize the burden of interpolation processing while preventing visual deterioration from being felt against the contents of an image. An object is to provide an image processing method capable of suppressing.

[0007]

In order to achieve the above object, the present invention proposes to convert compressed image data containing a frequency (AC) component in units of a plurality of pixels into a spatial frequency domain by using orthogonal transformation. From uncompressed data to
In an image processing method characterized by extracting data including frequency (AC) components for a plurality of partial pixel regions that are compression units, a feature amount of frequency (AC) components of the extracted plurality of partial pixel regions and , The preset spatial frequency characteristic threshold value is compared, and the high frequency data block and the low frequency data block are determined as the frequency characteristic, and the result is associated with the position information of a plurality of partial pixel areas to change the image data. At least two types of filters with different interpolation precision for doubling are provided, and when scaling (enlarging or reducing) image data, the scaling filter is switched and used in multiple partial pixel area units based on the previously obtained result. It is characterized by doing.

[0008]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the accompanying drawings.

When image data of Jpeg File format which is a general digital camera data compression recording format is printed, luminance and chromaticity bases are set for each block (8 * 8 pixels) before being restored to uncompressed data. By also obtaining the DCT (discrete cosine transform) data of and determining the content of the spatial frequency component for each block, an interpolation filter suitable for each data of the block can be set. By doing this, the interpolation processing for printing can be used without degrading the image quality when performing the enlargement processing for printing in one image file, and without performing the interpolation processing that requires a uniform and high processing time. Become.

Information omission and encoding / decoding of "Jpeg file" of the most common image compression file at present will be described.

Regarding encoding, generally, in a digital camera, a digital video or the like, it is general to save a still image as a Jpeg file. In this case, the signal from the CCD, which is the light receiving element of the input device, is A / D converted and then taken into the frame memory, RGB or CM
The Y filter information is converted into luminance and chromaticity information. After that, it is divided into 8 * 8 (64) square pixel blocks.

In FIG. 3, the bitmap of the luminance data is 8
* This is an example of data of one block of 8 blocks. In the above, the pixel values of 0 to 255 are level-shifted and converted into signals of -128 to 127. In, the DCT coefficient is obtained by DCT (discrete cosine transform). This is a quantization table in which omission of high-frequency components is increased in consideration of visual characteristics, and using this table, the DCT coefficient that is the result on the left is quantized.
Is the result of quantization. This value is entropy-coded and represented by a Huffman code to generate compressed data which is a coded signal.

Next, in decoding, the reverse process of encoding is performed.

That is, the encoded signal is decoded and quantized D
Decode the value of the CT coefficient. Next, a DCT coefficient is obtained by multiplying the quantization table to perform inverse quantization. After that, the level-shifted image is restored by performing the inverse DCT, and the image of one block is decoded by further adding the value 128 of the inverse level shift.

In the description, the combination of the data divided into the luminance information and the chromaticity information and the conversion into the RGB image is omitted, but the flow in the encoding is as shown in FIG. ) And two chromaticity components (Cb, C
r), and encoding and synthesizing each of them,
Generates a compressed image.

As a method of printing a Jpeg image which is a compressed image data file as described above, the image data from the input device is taken into a PC by a USB or a storage medium, the image is expanded, and the image is corrected when necessary. When sending data to the printer after adding the image, inputting image data from the input device directly to the printer, decompressing the image in the printer, performing image correction if necessary, and printing. There are different kinds of options.

In any case, in order to print the input image, it is necessary to convert the number of pixels of the input image data into the number of pixels which matches the print mode and the print range. Figure 1 is Jpeg
It is a block diagram showing the process of decompressing a file and the information acquired at that time. In the process of converting a Jpeg file into RGB bitmap data, first, entropy decoding is performed using a code table, and the quantization table used for inverse quantization is stored as data in addition to performing inverse quantization. The dequantized data is frequency-converted as data in block units, and this data is acquired together with position information as data for obtaining the image frequency characteristic. After that, inverse DCT processing and inverse level shifting are performed, and Ycc-RGB conversion is performed, so that normal RGB bitmap data is developed.

In this compressed image decompression, dequantized data in block units are acquired, the frequency characteristic is judged, and the interpolation filter is selected. This series of flow is shown in the flowchart of FIG. Show.

At the same time that the DCT data in block units of 8 * 8 pixels is acquired, the image file is expanded into RGB bitmap data.

The acquired frequency component value (AC value) in block units of 8 * 8 pixels is composed of 63 from low frequency to high frequency.

Further, the frequency data used in the present embodiment uses the luminance component, but it can be considered to use the chromaticity component.

In the quantization of the Jpeg file, the high frequency component has a considerably rounded value, and the actual image often has a value of "0" in the high frequency part. Since this quantization weighting exists, in the present embodiment, a value representing the frequency characteristic in the block of 8 * 8 pixels is obtained by adding 63 pieces of AC component data.

However, the value representing the frequency characteristic can be determined by another method using the AC component. For example, addition of values in a low frequency region, summation after multiplying each item of AC components 0 to 63 by a weighting factor, and the like are conceivable. Next, the total number of blocks of 8 * 8 pixels in the original image is obtained and stored as "n".

The detection determination is performed in a block unit of 8 * 8 pixels from the upper left of the image, and is performed until the last lower right block. Select the block that you are paying attention to as "B
C ″, and the determination is started from the upper left block where BC = 1.

First, it is judged whether or not the block to be judged has completed the final block, and if it has not ended, the judgment work is started. AC of the block specified by the value of BC
The added value of the component is compared with the judgment table (T), and if it is larger than the judgment table value, the frequency characteristic of the corresponding block is set to a high frequency, and if it is smaller than the judgment table value, the frequency of the corresponding block is set. The characteristic is set to low frequency. This determination is repeated until the value of BC exceeds "n".

In the present embodiment, "T = 200"
Is set, and when the value of the total sum of AC components in one block exceeds 200, the high frequency is set.

After the completion of the determination work, in the enlarged interpolation process for printing, a high-quality high-order interpolation process (for example, cubic) is performed on the block determined to have the high-frequency according to the determination, and the block is determined to have the low-frequency. Interpolation processing (for example, box filter) that is light in processing is performed on the selected block.

In this way, if necessary, high-order interpolation processing is performed on the entire image in a block having high spatial frequency data, and in other blocks, the processing is simplified to improve the image quality. The amount of expansion interpolation processing is reduced without dropping.

Further, in order to perform smoother interpolation processing at the boundary portion where the interpolation processing is switched, an interpolation processing filter used for a block having a frequency characteristic in the middle of the high band and the low band which is considered to be the boundary part. There is a way to deal with it by adding the one that has the performance between wide area and low area. The flow chart when doing this is shown in FIG.
Shown in. At the same time as obtaining the DCT data in block units of 8 * 8 pixels, the image file is expanded into RGB bitmap data. A value representative of the frequency characteristic of a block unit of 8 * 8 pixels is obtained by adding 63 pieces of AC component data.

Next, the total number of blocks of 8 * 8 pixels in the original image is calculated and stored as "n".

Up to this point, if the value of BC is n or less, the value of the total AC component of the block specified by the value of BC and the table "T1" are used in the same process as described with reference to FIG.
"Compare values of". T1 "is a table value for determining whether the spatial frequency characteristic of the image of the 8 * 8 block shows a large value, and in the present embodiment," 200 ".
When the sum in the block is larger than 200, the block designated by BC is set as a high frequency block.

If "T1" or less, the value of "T2" is compared next. “T2” is a threshold table for distinguishing the low frequency characteristic and the intermediate frequency characteristic from the spatial frequency characteristic of the image of 8 * 8 block. In this embodiment, "80" is used.

Here, if the value of the block designated by BC is larger than 80, the block designated by BC is set as the middle frequency block, and if it is smaller than 80, the block designated by BC is Set as a low frequency block.

The above determination is performed until BC for which all the images have been determined exceeds "n", and after the determination work is completed, in the enlarged interpolation processing for printing, in the block determined to have a high frequency according to the determination, High-quality high-order interpolation processing (for example, cubic) is performed, and for blocks that have been determined to have a mid-range frequency, moderate interpolation processing (for example, "Triangle") is performed on blocks that have low performance and processing. In (1), interpolation processing (for example, a box filter) that is light in processing is performed.

FIG. 6 shows an image of a person, the size of which is UXGA (1600 * 1200). The background is Umebayashi, and it can be said that the spatial frequency of this part is high.

FIG. 7 shows an image of a landscape, and the size is UXGA (1600 * 1200). It can be said that the image has a low spatial frequency and a high image compression efficiency.

FIGS. 10 and 11 show the results of the determination of the switching of the enlarged interpolation of the images of FIGS. 6 and 7 according to the flow chart shown in FIG.

The judgment result is that the judgment of each block is replaced with a pixel of 1 dot, and further, when the color information of the dot is brown, it indicates that the judgment detects a low frequency characteristic, and when it is red, The judgment indicates that the mid-range frequency characteristic is detected, and the case of yellow indicates that the judgment detects the high-range frequency characteristic.

In FIG. 10, the skin portion of the person photographed up close is determined to be the low frequency range. A part of clothes such as a trainer is determined to be in the middle frequency range, and a part of the background such as Bairin and the cut portion of the object are determined to be in the high frequency range.

In FIG. 11, it is determined that the high frequency band is the ground part which is the right end of the screen, and the block of the high frequency band and the block of the middle frequency band where the boundary between the blue sky and the cloud is slightly. Mixed. Many other blocks are in the low frequency range, and the image of such composition can maximize the effect of improving the processing efficiency of the present embodiment.

In the multi-interpolation in which the three kinds of interpolation filters in the flowchart of FIG. 5 are switched, smooth image enlargement interpolation can be achieved, but in the case of printing to a large size in which the enlargement ratio of printing is large, Strict observation makes it possible to find image changes due to slight switching at the boundary part. Therefore, in order to deal with this, after determining the high range and the middle range, the range is adjacent to a block (or adjacent to two blocks). A method in which the image quality performance is improved by disclosing it in each block will be disclosed. FIG. 8 is a flow chart for carrying out the widening of the higher frequency to the adjacent block. The data reading and the comparison and judgment of the "T1" and "T2" tables, which are the flowcharts in the left column, are the same as those shown in FIG. 5, and the processing after the block count of BC exceeds "n" Addition occurs. The block count of BC is set to "1" again, the top left corner of the image is returned to, and the data block determined to be a high frequency band exceeding "T1" in the previous determination is detected. After the detection, a reservation for converting to the high frequency band is made regardless of the setting of the block adjacent to the block. This operation is executed until the BC counts up, and after the end, the setting of the block reserved as the high frequency band is changed to the high frequency block.

Next, the block count of BC is set again to "1" and returned to the top left corner of the image, and a data block judged to be a high frequency band exceeding "T2" is detected in the previous judgment. To do.

After the detection, a reservation for converting to the mid-high range is set only when the setting of the block adjacent to the block is set to the low-frequency block. This work is BC
Is executed until the counter counts up, and after that, the block reserved for the middle frequency band is changed to the middle frequency block and the process is completed.

After the completion, in the enlargement interpolation process for printing, high-quality high-order interpolation process (for example, cubic) is performed on the block determined to have the high frequency according to the determination, and the block is determined to have the intermediate frequency. In the block, moderate interpolation processing (eg Triangle) is performed. Interpolation processing (for example, a box filter) that is light in processing is performed on the block determined to be the low frequency.

6 according to the flow chart shown in FIG.
And FIG. 15 and FIG. 16 show the results of the determination of the switching of the enlarged interpolation of the image of FIG. Color coding of the results for the determination is the same as in the case described above. Comparing FIG. 10 and FIG. 11, which are the results performed in FIG. 5, it can be confirmed that the blocks in the high frequency band and the blocks in the middle frequency band both expand the range. When this result is viewed as the execution result of the enlarged interpolation processing by switching, the variation amount at the boundary becomes less noticeable.

Based on this result, the case where the adjacent blocks are the adjacent blocks for two blocks is shown in the flowchart of FIG. 9, and the determination results are shown in FIGS. 17 and 18.

From this result, it can be confirmed that both the blocks in the high frequency band and the blocks in the middle frequency band have expanded the range as in the above case. However, in FIG. 17, it can be seen that blocks in the high frequency band occupy a considerable ratio in the total number of blocks determined.

In the switching enlargement interpolation processing, if the high frequency range occupies the majority of the whole image, the load of judgment and processing for switching is added.
It can be seen that the processing efficiency, which is the original purpose, cannot be expected for a single processing by the high-order interpolation filter for the high frequency band.

As a measure against this, the flow chart of FIG. 13 shows a determination method when a certain ratio of the blocks other than the low frequency band exceeds a certain ratio to the entire image.

In the original image, the total number "n" of blocks of 8 * 8 pixels is obtained, and the block of interest for determination is set to "BC", which is set in the upper left block where BC = 1. Further, the value of the counter "Count1" for counting the total number of blocks in the high frequency band and the value of the counter "Count2" for counting the total number of blocks in the middle and high frequency bands are set to "0". The detection is performed in a block unit of 8 * 8 pixels from the upper left of the image, and is performed up to the last lower right block. When BC exceeds “n”, “Cou
The count of “nt1” and the count of “Count2” are compared with the determination criterion. In the present embodiment, the value of “Count1” exceeds 70% of the total number of blocks, or “Count”.
The sum of "1" and "Count2" is 80% of the total number of blocks.
It has a mechanism to detect the case of exceeding. When the set value is exceeded, the switching enlargement interpolation process is prohibited, and the enlargement interpolation filter in the high frequency range is set to be executed for all images.

However, "Count1" and "Count
As for the setting method and value of 2 ", other values and methods can be executed, but if switching is prohibited in relation to the entire image, the same applies to other methods.

With respect to the contents described so far, if setting is made in any case, all the judgments are made by one operation regardless of the input / output status. Next, the setting of print quality, input size, output size A method of obtaining the enlargement interpolation processing magnification and the print size of the 8 * 8 block in the input image data from the print mode and selecting the enlargement interpolation processing switching table suitable for this will be described.

First, regarding the setting selection method according to the print quality, the table of the set table values is shown in FIG.

The vertical axis represents "high band" for distinguishing between the high band frequency band and the middle band frequency band for executing the three-stage interpolation filter switching shown in FIG. 5, and the middle band frequency band and the low band frequency band. There is a "middle range" column that distinguishes the frequency range.

The horizontal axis is the print quality setting class.
Although there are an unlimited number of expression methods, in the present embodiment, an example of three stages of “fast”, “normal”, and “clean” is provided for easy understanding. In "fast", printing speed is prioritized over image quality, and in "clean", image quality is prioritized. The "normal" setting is somewhere between "fast" and "clean".

From the values in the table, it can be seen that the "fast" setting reduces the ratio of high-frequency and mid-frequency filters used. In this table, "clean"
Although the filter is switched also in the setting of 1, the interpolation may be executed only by the highest level filter that performs the interpolation in the high frequency band.

This table is judged. The flowchart is shown in FIG. The print quality mode set in the printer is acquired, and it is determined whether the mode is "fast".
In the case of "fast", the middle range setting is set to "80" and the high range setting is set to "200", and the process is completed. When the setting is not "fast", it is determined whether it is "normal". In the case of "normal",
Complete with the mid range setting at "60" and the high range setting at "150". If the setting is not "normal", the setting is "clean" by the exclusion method, so the middle range setting is set to "40" and the high range setting is set to "100" to finish the process and wait for a print command.

Next, the enlargement interpolation processing magnification and the print size of 8 * 8 blocks in the input image data are obtained from the print quality setting, input size, output size, and print mode, and the enlargement interpolation processing switching table suitable for this is obtained. The selection method will be described.

FIG. 19 shows a table of set table values.

The horizontal axis represents the enlargement interpolation magnification required for the print size and print mode (resolution) of the input image data. There are three levels of "less than 3 times", "3 or more and less than 5 times" and 5 times or more of the medium.

In the present embodiment, the case where the input image size is UXGA (1600 * 1200) and the print size setting is 203 mm * 152 mm will be described. In the print mode (with resolution setting), 600
It is possible to set 360 (dpi) as the one for dpi (Dot Par Inch). Of these, 48 when printing at 600 dpi
00 * 3600 pixels are required, the interpolation magnification becomes 3 times, and when printing at 360 dpi corresponding to the column of "to * 5", 2877 * 2157 pixels are required and the interpolation magnification is 1.8. It is doubled and corresponds to the column of "to * 3".

The vertical axis is classified according to the size of the length of one side of the 8 * 8 block of the input image data to be printed. In the classification of the present embodiment, 203 mm is 200 (1600 dot divided by 8), which is obtained by the size of the input image and the print size.
It is classified by whether the value divided by is less than 2 or is 2 or more.

In this classification, the “high band” for distinguishing the high band frequency band and the middle band frequency band for executing the three-stage interpolation filter switching shown in FIG. 5 and the middle band frequency There is a "middle range" column that distinguishes the low and high frequencies.

Looking at the values in the table, the larger the print size of the length of one side of the 8 * 8 block of the input image data, the larger the interpolation ratio is set so that the use ratio of the high quality filter is high. Regarding the interpolation magnification, the larger the interpolation magnification is, the higher the use ratio of the high quality filter is set in the enlarged interpolation processing.

This table is judged. A flow chart is shown in FIG. Get the input image size. Then, the print size and mode (including resolution) are acquired. From the acquired information, the enlargement interpolation magnification "HB medium and print length of one side of 8 * 8 block of input image data" PS "is calculated.

It is determined whether the value of "PS" is less than 2 or more than 2, and when it is less than 2, "Cubic" is used for the high-frequency interpolation filter and "Tri" is used for the middle-frequency interpolation filter.
Set "angle". If it is 2 or more, set "Lanczos" to the high-frequency interpolation filter and "Gaussian" to the middle-frequency interpolation filter. This is a high-quality filter that can be used.

When the setting of the filter is completed, it is judged whether or not the enlarged interpolation magnification "HB medium is less than 3. If it is less than 3, the table value is set. If it is 3 or more,
It is determined whether "3 or more and less than 5" or "5 or more", and the table corresponding to the result is set.

The table to be set has the contents of the table of FIG. 19, but is as follows. "PS <2""Less than 3" Mid range setting = 80, High range setting = 200 "3 or more and less than 5" Mid range setting = 60, High range setting = 150 "5 or more" Mid range setting = 40, High Range setting = 100 "PS = 2 or more""Less than 3" Middle range setting = 60, High range setting = 150 "3 or more but less than 5" Middle range setting = 40, High range setting = 100 "5 or more" Middle range Setting = 30, high frequency setting = 80 In the present embodiment, the filter switching value is determined by switching using a table, but a method of calculating a linear switching value using a function or the like is also conceivable.

In addition, table setting depending on the type of printing paper and the type of ink loaded in the printer can be executed with the same idea.

Although the added value of the AC components (63) of each block was used for selecting the determination table of the original image 8 * 8 block shown in FIG. 5, another method is used as a method for determining the characteristic of the block. Indicates.

In FIG. 22, the AC component is reduced to 1 from the low frequency component.
Items grouped in units of 0 (the 7th group is 61-
63), and the group value is the addition of each item.

For determination, AC components of "1-10", "1"
1 to 20 ", ..." 61 to 63 "are compared with the setting values for each item, and if all match, the result is" Yes ".
The filter in that block can be set as in the flowchart of FIG.

In the present embodiment, a flow chart is shown in which a block of the same image is searched for a plurality of times for switching judgment, but a series of processing is performed once. It goes without saying that it can be achieved by

Further, the present embodiment discloses a method using "Jpeg file" as an image compression file.
It goes without saying that the detection of the image of interest can be realized by a simple process with the same idea for other files using conversion into frequency components such as eg2000 file “.

[0075]

As is apparent from the above description, according to the present invention, in the case of a built-in device having a lower processing capacity than a PC, such as when printing directly from a digital camera, the print size of the printer The process of performing higher-order interpolation of the image data so as to match with the above has a problem that it takes a considerable load and takes time from the print command issuance to the print start. It is possible to obtain the effect that the burden of the interpolation processing can be minimized while preventing the visual deterioration from being felt without going against the contents.

[Brief description of drawings]

FIG. 1 is a conceptual diagram showing a flow of acquiring data necessary for decompressing a Jpeg image according to the present invention.

FIG. 2 is a conceptual diagram showing a flow of a process of converting image data of the present invention into a Jpeg format.

FIG. 3 is 8 * 8 Bloc, which is a Jpeg image compression unit of the present invention.
It is a figure which shows the process of converting into Jpeg format which made k into the example.

FIG. 4 is a flowchart from Jpeg image decompression to block determination according to the present invention.

FIG. 5: Jpeg corresponding to three kinds of interpolation filters of the present invention
It is a flowchart from image decompression to block determination.

FIG. 6 is an original image (Atyp for examining the judgment of the present invention.
It is a figure which shows e).

FIG. 7: Original image (Btype) for examining the judgment of the present invention
FIG.

FIG. 8 is a flowchart for grouping adjacent blocks of a high frequency data block according to the present invention.

FIG. 9 is a high frequency data block 2 according to the present invention.
It is a flowchart about grouping of individual adjacent blocks.

FIG. 10 is a diagram showing an A type image determination result of performing frequency determination based on Base Type of the present invention.

FIG. 11 is a diagram showing a B type image determination result of performing frequency determination based on Base Type of the present invention.

FIG. 12 shows the AC frequency data block class of the present invention as n.
It is a flowchart of the frequency determination which has (two or more) steps.

FIG. 13 is a flowchart when the total number of high frequency data blocks of the present invention exceeds the set ratio in the original image.

FIG. 14 is a switching table of the interpolation determination table according to the print quality setting of the present invention.

FIG. 15: Aty for which frequency judgment is performed based on adjacent inclusion type
It is a figure which shows a pe image determination result.

FIG. 16: Bty for which frequency judgment is performed based on adjacent inclusion type
It is a figure which shows a pe image determination result.

FIG. 17 is a diagram showing Atype image determination results obtained by performing frequency determination based on 2Block adjacent inclusion type.

[Fig. 18] Fig. 18 is a diagram illustrating a B type image determination result in which a frequency determination is performed based on a 2Block adjacency inclusion type.

FIG. 19 is a switching table of the interpolation determination table according to the input image size, print size, and print mode of the present invention.

FIG. 20 is a flow chart for dividing the used interpolation filter determination table according to the print quality of the present invention.

FIG. 21 is a flowchart for dividing the used interpolation filter determination table according to the printer settings (print size, print mode) and the size of the original image in the frequency determination of the present invention.

FIG. 22 is a setting UI screen for an AC frequency data block class of the present invention.

Continued front page    F term (reference) 5B057 AA11 BA02 CD06 CE05 CE06                       CG05 CH07 CH18 DC01 DC16                 5C076 AA21 AA22 AA32 BA05 BA06                       BA07 BB25 BB40                 5C078 AA04 BA57 CA22 CA31 CA34                       DA02

Claims (8)

[Claims] 1. A plurality of compression units are used when decompressing image data compressed from compressed data into uncompressed data by compressing image data containing a frequency (AC) component in a unit of a plurality of pixels using orthogonal transformation into a spatial frequency domain. In the image processing method, which is characterized by extracting the data including the frequency (AC) component of the partial pixel region of, the feature amount of the frequency (AC) component of the plurality of extracted partial pixel regions is set in advance. The spatial frequency characteristic threshold value is compared and the high frequency data block and the low frequency data block are determined as the frequency characteristic, and the result is associated with the position information of a plurality of partial pixel areas, and the interpolation accuracy for image data scaling is calculated. At least two types of different filters are provided, and when scaling (enlarging or reducing) image data, a scaling filter is performed in units of a plurality of partial pixel areas based on the previously obtained result. An image processing method characterized by using by switching the filter. 2. A high-precision and high-order filter is used in a partial pixel area determined to be a high-frequency data block, and a low-precision filter is used in a partial pixel area determined to be a low-frequency data block. The image processing method according to claim 1, which is used. 3. A high frequency band including a partial pixel region which is determined to be a high frequency data block and which is adjacent to the partial pixel region or adjacent to n (n is a natural number) partial pixel regions. 3. The image processing method according to claim 1, wherein the image processing method is divided into a data block group and another low frequency data block group. 4. A plurality of compression units are used when decompressing image data compressed from compressed data to uncompressed data by compressing image data containing a frequency (AC) component in units of a plurality of pixels using orthogonal transformation to the spatial frequency domain. Data including a frequency (AC) component for the partial pixel region of is extracted, and the feature amount of the frequency (AC) component of the plurality of extracted partial pixel regions is compared with a preset spatial frequency characteristic threshold value, As frequency characteristics, n (n is 3 or more) frequency data block classes are determined and divided, the results are associated with position information of a plurality of partial pixel regions, and at least n filters with different interpolation precision for image data scaling are used. When the image data is scaled (enlarged or reduced) with different types, the scaling filter should be switched in units of multiple partial pixel areas based on the result obtained earlier. Image processing method according to claim. 5. When the total number of detected high frequency data blocks exceeds a set ratio or value occupying the entire image, interpolation is set uniformly for high frequency data block interpolation for all image data. The image processing method according to claim 1, wherein a filter is used. 6. A frequency (alternating current) of a plurality of partial pixel regions in an original image, comprising means for acquiring printer settings, and based on at least the setting contents and original image size information.
The image processing according to any one of claims 1 to 5, wherein a plurality of preset spatial frequency characteristic threshold values for classifying components into frequency data block classes are provided, and the threshold values are switched according to the setting and executed. Method. 7. The image processing method according to claim 1, wherein the compressed image data is Jpeg File data. 8. A print size corresponding to one pixel of the input image is calculated based on the number of pixels of the input image and the output print size, and a filter for the interpolation process is calculated from the calculation result. The image processing method according to claim 1, which is used by switching.