JP4415650B2 - Image processing method and image processing apparatus - Google Patents

Description

  The present invention relates to an image processing method and an image processing apparatus, and in particular, an image processing method and an image processing for performing correction processing on an image file including original image data and shooting information generated at the time of shooting the original image data. Relates to the device.

In recent years, digital still cameras (hereinafter referred to as electronic cameras) have become widespread. An electronic camera uses a semiconductor element that reacts to light, such as a charge-coupled device (hereinafter referred to as CCD), and converts the captured image into an electrical signal, which is converted into digital data (hereinafter referred to as electronic data) to a magnetic storage medium or semiconductor memory Remember.
Since an electronic camera is usually equipped with a liquid crystal display, a user can check a captured image on the spot or delete electronic data that is not suitable for the purpose. In addition, the original image data captured by an electronic camera can be output after being subjected to various correction processes using a general-purpose personal computer monitor or an image output device such as a printer. It's getting on.

  Along with such expansion of the range of use, there is a growing demand for higher image quality for electronic cameras. As the image quality is improved, it is essential to improve the number of pixels and the sensitivity in order to maintain the image quality satisfied by the user. This will be described with reference to Table 1 and FIG.

Table 1 shows the sensitivity, the number of pixels, and the user satisfaction rate. The ISO sensitivity is arranged in the column, the number of pixels is arranged in the row, the right side in the column shows the percentage of people who are satisfied with the image quality, and the left side shows the percentage of people who are not satisfied.
Looking up at the top of the table, it can be seen that as the number of pixels increases, the proportion of users satisfied with the image increases. That is, it simply indicates that the number of pixels needs to be improved in order to achieve high image quality to satisfy the user's request.
However, a lot of noise recorded in the original image data is included.

FIG. 15 is a diagram illustrating a correlation between sensitivity and perceived noise amount.
When the number of pixels is increased, more noise is recorded in the original image data, and this figure shows this tendency. In this figure, the vertical axis represents the amount of noise, the horizontal axis represents the signal value of the original image data, and the line type represents the sensitivity.
From this figure, it is clear that the amount of perceived noise gradually increases by increasing the sensitivity (see Non-Patent Document 1).

As a conventional technique for solving the above problem, for example, there is an invention related to an imaging apparatus that reduces noise in an imaging signal while reducing a release time lag (see Patent Document 1).
In addition, there is an invention related to a digital camera that can remove fixed pattern noise in a shorter time than conventional (see Patent Document 2).
Each technique is a concept that eliminates fixed pattern noise that occurs due to variations in CCD output characteristics and uneven dark current. Specifically, it detects fixed pattern noise by driving the image sensor without exposure. The data is stored in the memory, and the fixed pattern noise is removed by subtracting the data recorded in the memory from the output signal of the image sensor obtained by actually photographing the subject.
JP 2003-1116064 A JP 2003-51991 A The Number of Effective Pixels for Digital Still Cameras and its Picture Quality., Miyasaka et.al., pp.147-150, IS &T's 2003 PICS Conference Yuichi Ota, Structuring Color Image Information by Region Division, Information Processing, Information Processing, Vol.19, No.12, pp.1130-1136 Picture Segmentation Using A Recursive Segmentation Splitting Method, Comp.Graph. And Image Proc., RON Ohlander et.al., pp.313-333 Principals of COLOR TECHNOLOGY, ROY S BERNS, pp.191-197 (2001)

However, when the above-mentioned conventional technique is applied to image data that has a so-called rough feeling in which noise is conspicuous in a high-saturation gradation part, in order to perform noise removal on the entire area of the image, sharpness and gradation are improved. A poor image may be obtained. For example, it is assumed that noise is perceived in an empty area (A) in FIG.
In the above technique, noise is removed from the area (A) and the area (B) in FIG. 2, so that the noise in the area (A) is removed, but the sharpness and gradation of the area (B) are reduced. It had a problem of being lost. In particular, the loss of gradation and sharpness in a color range in which a color range deeply related to human preference in original image data is referred to as a memory color is markedly a reduction in image quality.

Therefore, it has been desired to reduce noise while maintaining gradation and sharpness in the memory color range. For this purpose, it is possible to reduce the noise only in the region (A) in FIG. 2B, but this solution is difficult with the conventional noise removal processing.
In general, for original image data that includes memory colors, as image processing that emphasizes memory colors, correction that is adjusted to a much brighter and more vivid hue than the electronic data of the original captured original image data Since processing is performed, noise in the memory color area is further perceived.
Therefore, a technique for maintaining gradation and sharpness in the memory color range while reducing the amount of noise has been desired.

(Background to Invention)
Here, the smoothing process according to the photographing information of the original image data will be described.
FIG. 16 is a flowchart showing a main part of the electronic camera.
The electronic camera converts the captured optical signal output from the CCD into a digital signal by an A / D converter (step S1). The imaging signal converted into the digital signal is separated into R, G, and B color component data by the color separation unit, interpolated by the signal interpolation unit, and then gain adjustment unit for each of the R, G, and B color components. (Step S2).
After gain adjustment by the gain adjustment unit (step S3), the image data is subjected to smoothing processing (step S4). Here, the smoothed image signal is subjected to predetermined conversion processing (for example, YCbCr conversion) and then stored or transmitted to each processing unit (step S5). That is, the image data is stored in a storage unit for recording image data on a recording medium such as a flash memory, and transmitted to a display unit for displaying an image on a liquid crystal display device or the like.

Here, the mechanism of the gain adjustment process and the smoothing process and the background of the present invention will be described.
The gain adjustment unit includes an R gain adjustment unit, a G gain adjustment unit, and a B gain adjustment unit, which are amplification processes for digital signals, and can perform different amplification processes on each wavelength. is there. For example, as shown in FIG. 17, when the three gain adjustment units are arranged with a blue filter, a green filter, and a red filter, each from a short wavelength, these three filters may perform different amplification processes. Is possible.
FIG. 17 is a characteristic diagram showing the relative sensitivity with respect to the three primary colors of the image signal.
For example, the Blue filter is amplified as indicated by the arrow in the figure, the other Green and Red filters are fixed, and original image data in which only the B component of the original image data is sensitized is generated. .
If this is corresponded to the case where the scene shown in FIG. 2 is imaged, it corresponds to sensitizing the Red filter in order to match the color balance of the entire image. Therefore, it is considered that the R component includes a large number of noises.

On the other hand, when shooting an image, light source information is set according to the shooting situation and recorded in the image file as shooting information. For example, when a bluish image is taken, a C light source (6,500 [K]) corresponding to the image is recorded. Therefore, since the sensitized signal component can be inferred by referring to the light source information, strong smoothing processing is applied to the sensitized image data, and to the unsensitized image data. Ordinary smoothing can be performed.
Thus, while the inventor has intensively studied, it has been devised that smoothing processing is effective in accordance with imaging information such as light source information and sensitivity information, and the present invention has been achieved.

  In view of the above circumstances, the present invention has been made in view of the future improvement in image quality of electronic cameras, and an image file including original image data and shooting information generated at the time of shooting the original image data. It is an object of the present invention to provide an image processing method and an image processing apparatus that perform correction processing and output.

In order to solve the above-mentioned problem, the invention according to claim 1 is an image processing method for generating output image data based on an image file composed of original image data and photographing information of the original image data. An original image data dividing step for dividing the original image data, and a statistic calculation for calculating at least one of color information (tristimulus value) statistics and physical quantities of the original image data for each of the divided areas. An image feature amount calculation step for calculating an image feature amount of the original image data, and a threshold value of the noise removal filter based on the sensitivity information included in the shooting information and the calculated image feature amount a threshold calculating step of calculating a included in the a priority determination step of determining the priority of smoothing based on statistics or the physical quantity of the original image data, the shooting information for each region obtained by the division An identifier assigning step for assigning an identifier from the light source information and the color information relating to the memory color stored in advance, and an important region estimation step for estimating a divided region to be emphasized based on the results of the priority determination step and the identifier assigning step. wherein characterized in that it has a smoothing step of performing smoothing on the basis of the calculated threshold.

According to a second aspect of the invention, in the image processing method according to claim 1, wherein the identifier assigning step converts the color information on the memory color, the color information of the light source information of the photographing information (tristimulus values) Color An information conversion step, a relative distance calculation step for calculating a relative distance of color information in each region divided by the memory color and the division step, and an identifier for which the relative distance is minimum And an identifier determining step for determining the identifier .

  According to the present invention, an image generation step for generating output image data from an image file composed of original image data and shooting information, and based on the light source information and sensitivity information of the shooting information in the image generation step. By having the smoothing step for performing the smoothing process, it is possible to generate a highly sensitive image file while reducing noise.

The network configuration of the image processing system is such that an electronic camera and an information processing device such as a personal computer are connected by data communication means (for example, a USB interface) capable of mutual data communication. A configuration in which a printer (inkjet system or electrophotographic system) is connected by data communication means (for example, a printer dedicated serial line, a LAN line, etc.) capable of mutual data communication is common. This is because a personal computer can perform advanced image processing due to its processing performance, and image processing software has been introduced to realize this.
The image processing method according to the present invention is most preferably applied to the network configuration as described above. However, recently, an electronic camera and a color printer can be connected to each other by a wireless LAN, infrared communication, or a dedicated connection line. It is also possible to print an image.
Hereinafter, various embodiments of an image processing method according to the present invention will be described with reference to the accompanying drawings.

(Summary of Example 1)
Embodiment 1 describes an embodiment in which an original image data captured by an electronic camera is subjected to image processing according to the present invention by a personal computer and output by a printer in the image processing system having the apparatus configuration of FIG. It is.
About this embodiment, “1-1. Image processing system” and “1-2. Image processing method” are outlined, and a threshold value of a smoothing filter used for smoothing processing of original image data is calculated. -3. Threshold calculation method "and" 1-4. Smoothing processing method "according to the calculated threshold amount will be described in this order.
In this embodiment, the original image data shown in FIG. 2A is subjected to smoothing processing according to the threshold value calculated from the photographing information. The resulting smoothed image is shown in FIG.

(1-1. Configuration of Image Processing System)
(1-1-A. Configuration of Image Processing System)
FIG. 1 is a diagram showing a schematic configuration of an image processing system according to the present invention.
The image processing system includes an electronic camera 11 as an image file generation device, a personal computer 12 as an image processing device, and a color printer 13 as an image output device.
The electronic camera 11 can set various shooting information. Shooting information means acquisition conditions of original image data such as shutter speed, exposure, and aperture. The electronic camera 11 performs shooting (original image data generation) under shooting conditions set by the photographer, and generates an image file that integrally includes image data and shooting information. The generated image file is stored in a recording medium (for example, a memory card).
The personal computer 12 obtains an image file from the electronic camera 11 via, for example, a USB interface, and then performs image region division, identifier assignment, and smoothing processing.
The color printer 13 receives an image file from a personal computer via a data communication means such as a dedicated printer cable or a LAN (Local Area Network), and outputs the image file as an image. A known output format can be applied to the output operation of the color printer at this time.

(1-1-B. Generation of image file)
Here, the generation process of the image file 100 in the electronic camera 11 will be described.
First, the photographer sets a photographing mode prior to photographing. The shooting mode setting affects the original image data to be shot, such as exposure, white balance, shutter speed, aperture value, etc., suitable for each shooting scene prepared in advance, such as “Portrait”, “Landscape”, and “Sports”. The parameter that gives is automatically set. Regardless of the shooting mode, a “full auto mode” automatically set by the electronic camera 11 at the time of shooting and a “manual mode” set by the photographer according to their preference are also prepared. The electronic camera 11 applies this.
An image file is generated using a parameter value that is automatically set from a shooting condition (shooting mode) in response to a shooting request, for example, a depression of a shutter button.
That is, the image file 100 including the original image data and the shooting information is generated with reference to each parameter according to the set shooting mode. The image file generated here is an image file stored in a predetermined format including the image data and shooting information such as shooting conditions. This file structure will be described below.

(1-1-C. Configuration of image file)
In this embodiment, the image file 100 is described as an Exif format file, but the file format of the image file is not limited to this, and the image file 100 and the image processing control data are integrally provided. What is necessary is just to take a structure.

FIG. 4 is a diagram conceptually illustrating an example of the configuration of an image file in the present invention.
The image file 100 has a file structure according to the image file format standard (Exif) for electronic cameras. The specifications of the Exif file are determined by the Japan Electronics and Information Technology Industries Association (JEITA).
The image file 100 includes an original image data storage area 101 that stores original image data, and a shooting information storage area 102 that stores various types of additional information related to the stored original image data. In the original image data storage area 101, original image data is stored in JPEG format. Shooting information is stored in the TIFF format in the shooting information storage area 102. As is well known to those skilled in the art, the Exif format file uses various tags to identify each data, and tags stored in the original image data storage area 101 are tags. Various areas are assigned as names.

FIG. 5 is a diagram showing an example of shooting information data in the present invention.
In the shooting information storage area 102 of the image file 100, the tag name illustrated in this figure and its set value (not shown) are stored. Each tag name is followed by a “0x00” termination code indicating a character termination sequence. The storage area 102 includes an identifier indicating that a parameter is stored, a parameter designation number indicating the number of designated parameters, and a value for designating (identifying) a parameter number allocated in advance for each parameter. The stored parameter number, parameter setting value information storing the parameter setting value of the designated parameter number, and the like are stored. The parameter number is, for example, information stored in a 2-byte area, and the parameter setting value is information stored in a 4-byte area. In the image output apparatus, each parameter value of the shooting information can be acquired using this tag as an index.
For example, for “shooting mode”, when the tag name is “landscape mode”, “+0” is stored as the setting parameter. In addition, as a numerical value corresponding to “ISO sensitivity”, when a preset reference value such as “1” representing “800 equivalent” is prepared, it represents a correction amount from the reference value. Stores numeric values.

With the above file structure, the image file 100 stores the image file 101 including the shooting information set by the photographer. A predetermined image processing apparatus can acquire the original image data and the photographing information by acquiring the image file and analyzing the file with respect to the image file.
The image processing method will be described below.

(1-2. Image processing method)
The image processing method can take the form of a program installed in an information processing apparatus such as a personal computer, for example. The personal computer 12 acquires the image file generated in “1-C. Configuration of Image File” and performs image processing on the image file. The processed image file is transmitted to the color printer 13 which is an image output device. The color printer 13 is a printer that outputs a color image. For example, cyan (C), light cyan (light cyan, LC), magenta (M), light magenta (light magenta, LM), yellow (Y), dark yellow. This is an ink jet printer that forms an image by ejecting seven color inks of (DY) and black (K) onto a printing medium to form a dot pattern. In addition, various printers such as an electrophotographic printer that forms an image by transferring color toner onto a printing medium and heat-fixing the so-called color laser printer can be applied.
The outline of the operation of the image processing method applied in the image processing apparatus will be described below based on the flowchart of FIG.

FIG. 6 is a flowchart showing an image processing method according to the present invention.
The CPU of the personal computer 12 as the image processing apparatus extracts original image data from the acquired image file 100 (step S100). The electronic camera 11 stores image data as a JPEG file, and the JPEG file stores image data using a YCbCr color space in order to increase the compression rate.
The CPU executes a first matrix operation for converting image data based on the YCbCr color space into image data based on the sRGB color space (step S110). For the image data based on the sRGB color space obtained in this way, the CPU calculates a threshold value used for the smoothing filter based on the RGB value of the original image data and the shooting information (step S120). Smoothing processing is executed using the threshold value (step S130).

Then, it is converted into a CMYK color space corresponding to the RGB color space stored in the ROM (step S140). This conversion look-up table (LUT) is a process for changing the color space of image data from the RGB color space to the CMYK color space. , The color used in the color printer 13, for example, data of gradation values of seven colors of C, M, Y, K, LC, LM, and DY. Finally, the CPU forms the obtained image output data on a recording medium and outputs it to obtain an output image.
Hereinafter, details of each process in steps S120 and S130 will be described.

(1-3. Threshold calculation method)
A characteristic value τ indicating the color bias of the image data output from step S110 “first matrix calculation” in FIG. 6 is calculated according to the following equation (1).
τrg = ΣR / ΣG,
τbg = ΣB / ΣG,
τbr = ΣB / ΣR,
ΣR = Σr _IJ / nr,
ΣG = Σg _IJ / ng,
ΣB = Σb _IJ / nb (1)
However,
Σr _IJ , Σg _IJ , Σb _IJ
Are the values of R, G, B at coordinates (I, J), respectively.
nr, ng, nb
Are the numbers of R, G, and B pixels, respectively.

The characteristic value obtained by the equation (1), for example, τrg represents the ratio of the red component to the green component. When the characteristic value τrg is greater than 1, the entire image is interpreted as reddish. The
Hereinafter, this embodiment will be described with reference to FIG.

7, Ru conceptual diagram illustrating a method for calculating a characteristic value of the xy chromaticity diagram according to the present invention.
First, the characteristic value representing the color deviation is plotted as a point D on the xy chromaticity coordinates (point * in the figure). At this time, chromaticity points of RGB primary color light can be determined by various standards, but in this embodiment, NTSC chromaticity coordinates are adopted (points R, G, B ).
From the R, G, and B points of each primary color light, the points internally divided at τrg: 1, τbg: 1, and τbr: 1 are calculated for the line segment RG, line segment BG, and line segment BR, respectively ( An intersection of three straight lines drawn from the vertices of points P1, P2, P3) and ΔRGB to the internal dividing point (P1, P2, P3) is defined as a D point (point * in the figure). In this embodiment, τrg = 3.91, τbg = 0.90, P1 = (0.160, 0,260), and P2 = (0.428, 0.530), so that D = (0.260, 0.360).

Further, the light source information of the photographing light source is acquired from the photographing information of the image file, and this is plotted on the xy chromaticity coordinates. The light source plotted at this time is set as point C. The vector connecting the points C and D is decomposed into a vector parallel to the line connecting the coordinates of each light source color and the internal dividing point. That is, the vector CD is decomposed into a vector Cp3 and a vector Cp1. A difference vector of the vector components obtained by the decomposition is calculated. In the present embodiment, from p1 = (0.314, 0.2781) and p3 = (0.2556, 0.3088), Cp3 = 0.0549 and Cp1 = 0.0382, and this difference vector is | Cp3-Cp1 | = | 0.0549-0.0382 |
= 0.0167
It is.

  Subsequently, the threshold value of the filter to be smoothed is calculated from the difference vector obtained in the previous step and the sensitivity information of the photographing information. In the present embodiment, it can be seen from the above-described difference vector that the image contains a lot of blue components as the color of the entire image. Therefore, it is assumed that a red component is added to the entire image in order to obtain a color balance according to the amount of the blue component. Therefore, the smoothing process is performed on the red component as the color component to be smoothed. Sensitivity information is acquired from the shooting information of the image file, and the ISO sensitivity is 800 in this embodiment. Next, a threshold value THR applied to the smoothing filter is calculated.

For example, an expression expressed by the following expression (2) is assumed.
THR
= Const./(|Cp3-Cp1|×I.Sensitivity) (2)
However,
Const .: Constant
I. Sensitivity: ISO sensitivity.
That is, when the difference vector and the ISO sensitivity are high, the threshold value THR becomes small, and a steeper smoothing process is performed. In this embodiment, the threshold value THR is
THR = 50 / (0.0167 × 800)
= 3.74
Met. Further, the constant in the present embodiment is in the range of about several tens.

(1-4. Smoothing processing method)
FIG. 8 is a diagram showing the concept of the smoothing filter processing in the present invention.
Smoothing processing is performed using the threshold value calculated in the previous step. An example of a filter that performs this process is shown in FIG. In this figure, a is the pixel value of the target pixel. Further, α, α ′, β, β ′, γ, γ ′, δ, and δ ′ are pixel values of pixels around the pixel a. Here, a point-symmetric set is defined as follows.
(Α, α '), (β, β'),
(Γ, γ '), (δ, δ')
Difference information is calculated by the following equation (3) for the point-symmetric group defined as described above.
Δα = abs (α−a)
+ Abs (α′−a) (3)
However,
abs (xy): Absolute value of xy.
Further, Δβ, Δγ, and Δδ are similarly calculated.

Of Δα, Δβ, Δγ, and Δδ calculated by Equation (3), the value that takes the minimum value is Δmin.
Here, if Δmin <THR for a certain threshold value THR, a is replaced with a three-point weighted average value for Δmin (setting of the threshold value THR will be described later). Specifically, when Δα is Δmin, the weighted average is expressed by the following equation (4a). This is defined as new pixel data a. Alternatively, when noise removal is gently performed on the original image, it can be obtained by the following equation (4b).
Weighted average = (α + a + α ′) / 3 (4a),
Weighted average = (α + 2a + α ′) / 4 (4b)

  If comprised in this way, there exists an effect which reduces strong image noise, leaving the edge of image information. Further, according to this embodiment, since the information difference larger than the threshold value is not smoothed, the edge information is likely to remain. In the above description, the case where the smoothing filter processing is applied to the eight points around the pixel of interest has been described. However, the present invention is not limited to this, and the pixels around a are skipped and the next pixel is skipped. Filter processing may be performed.

(Summary of Example 2)
Hereinafter, embodiments of the present invention will be described in the following order based on examples. Note that the generation of the image file and the configuration of the image file are the same as the contents shown in the first embodiment, and thus description thereof is omitted.
In the second embodiment, the original image data captured by the electronic camera 11 is subjected to the image processing according to the present invention by the personal computer 12 and output by the color printer 13 in the image processing system configuration having the apparatus configuration of FIG. It is description about the form to do.

Regarding Example 2, “2-1. Image processing method” is outlined, and “2-2. Original image data area dividing method” and “2 -3. Identifier assignment method ”and [2-4. The smoothing processing method] will be described in this order.
Further, in the second embodiment, the original image data shown in FIG. 2A is compared with the area (A) that is the empty area having the highest image occupying ratio among the areas that are the memory color areas. It is an Example which performs the smoothing process according to the threshold value computed from imaging | photography information. The obtained smoothed image is shown in FIG.

(2-1. Outline of image processing method)
The image processing method can take a form in which image processing is performed using a program installed in the personal computer 12, for example. The personal computer 12 acquires the image file generated in “1-1-C. Configuration of image file” and performs image processing on the image file. The processed image file is transmitted to the color printer 13 which is an image output device. The color printer 13 is a printer that outputs a color image. For example, cyan (C), light cyan (light cyan, LC), magenta (M), light magenta (light magenta, LM), yellow (Y), dark yellow. This is an inkjet printer that forms an image by ejecting seven color inks (dark yellow, DY) and black (K) onto a print medium to form a dot pattern. In addition, various printers such as an electrophotographic printer that forms an image by transferring color toner onto a printing medium and heat-fixing the so-called color laser printer can be applied.
The outline of the operation of the image processing method applied in the image processing apparatus will be described below based on the flowchart of FIG.

  FIG. 9 is a flowchart showing an image processing method according to the present invention. In particular, the image segmentation process (step S240), the identifier assignment process (step S250), and the smoothing process (step S280), which are features of the present invention, will be described in detail below.

First, the CPU of the personal computer 12 extracts original image data from the acquired image file 100 (step S200). The electronic camera 11 stores image data as a JPEG file, and the JPEG file stores image data using the YCbCr color space in order to increase the compression rate.
The CPU executes a first matrix operation for converting image data based on the YCbCr color space into image data based on the sRGB color space (step S210). The CPU performs gamma correction (step S220) and second matrix calculation on the image data based on the sRGB color space thus obtained (step S230). When executing gamma correction, the CPU acquires a gamma value on the electronic camera 11 side from the image processing control data, and executes gamma conversion processing on the image data using the acquired gamma value. The second matrix calculation is a calculation process for converting the sRGB color space into the XYZ color space.

The CPU executes region division using the features of the XYZ color space and the actual image space of the original image data in order to execute image quality adjustment based on the shooting conditions (step S240). Next, an identifier is assigned to the divided area (step S250). Then, the third matrix calculation and inverse gamma correction are executed (step S260), and the XYZ data is converted into RGB data. Further, matrix calculation is performed using a matrix corresponding to the conversion to the color space, and the conversion to the CMYK color space corresponding to the RGB color space stored in the ROM is performed (step S270).
The conversion look-up table (LUT) is referred to as processing for changing the color space of the image data from the RGB color space to the CMYK color space, and the image data composed of R, G, and B gradation values is converted to the color printer 13. For example, C, M, Y, K, LC, LM, and DY.
Finally, the CPU executes a smoothing process on the obtained print data (step S280).

The above steps are sequentially executed, the obtained image output data is formed on a recording medium, and an output image is obtained by outputting the data.
In the following, details of each process in the image division process (step S240), the identifier assignment process (step S250), and the smoothing process (step S280) will be described.

(2-2. Region Division Method (Step S240))
Various methods can be adopted as the region dividing method. For example, there are a method of recursively dividing from statistical information of original image data (see Non-Patent Document 2), a method of dividing an area using image statistical information (see Non-Patent Document 3), and the like. In the present embodiment, a description will be given of an area dividing method using the K-average algorithm as an example.

FIG. 10 is a flowchart showing a region dividing method according to the present invention.
This image region dividing method includes initial region generation (step S300), region division (step S310), minute region removal (step S320), numbering (step S330), small region integration (step S340), feature amount calculation (step S350).

The initial region generation step S300 generates an initial region by dividing the entire screen of the input original image data into preset small amount regions, and calculates statistical information for each region. At this time, as the statistical information, in this embodiment, an arithmetic average for each region was used. The initial area data is temporarily held and the process proceeds to an area dividing step. It is desirable that the number of area divisions at this time is appropriately changed according to the characteristics of the image data. The number of small areas in this embodiment is 5 in the vertical direction and 5 in the horizontal direction, and is divided into 5 × 5 = 25 areas (see FIG. 2C).
Next, using the original image data, the initial divided region obtained in the initial region generation step S300, and its statistical information, the region integration and division of the original image data are repeated a predetermined number of times (step S310). The obtained image signal is output to the minute region removing step S320. The minute area removing step S320 is constituted by a majority filter, removes small granular areas included in the image signal supplied from the area dividing step S310, and outputs the area image signal from which the granular areas have been removed to the numbering step S330. To do. The window size of the majority filter at this time is preferably changed as appropriate according to the characteristics of the image data. In this embodiment, the window size is 5 × 5 = 25 pixels.

The number assigning step S330 inputs the image signal from which the granular region is removed, assigns the same identification number to the pixels connected on the screen included in the obtained same region, and assigns the identification number to the region The image signal is output to the small region integration step S340.
The small region integration step S340 is obtained by inputting the region image signal to which the identification number is assigned and executing a process of integrating a region having a small number of pixels equal to or less than a predetermined threshold value into another region. The image signal is output to the feature amount step S350.
In number assignment step S330, an arbitrary number is assigned to each image area. At this time, since it is only necessary to recognize that the areas are different, the numbers given to the areas do not need to be consecutive numbers. By this numbering step S330, the modified divided areas can be individually identified even if they are areas having the same feature amount as long as they are not connected.

  Further, the region image signal to which the identification number is assigned in the number assigning step S330 is supplied to the small region integration step S340. In the small region integration step S340, a region having a small number of pixels equal to or smaller than a predetermined threshold value is integrated into another region. The threshold value is set to about 1% of the number of pixels on the previous screen, for example. The integration is performed by starting from the outermost area of the integration target area and causing pixels in contact with another area to belong to the other area. Since the pixels belong to multiple areas in order from the outside, all the pixels always belong to other areas.

The integration in the present embodiment is performed by integrating the target pixel into the adjacent region having the largest number of pixels among the eight neighboring pixels surrounding the target pixel in the range of 3 pixels × 3 pixels centering on the target pixel. It was. When the outer pixel is in contact with a plurality of regions, it belongs to the region with the largest number of pixels in contact. When the number of pixels in the adjacent area was the same, the entire adjacent area was compared and integrated into the maximum area.
Finally, in the feature amount calculating step, statistical information is calculated for each of the divided areas, and is temporarily held. FIG. 3A shows an example of the divided image obtained through the above steps, and Table 2 shows the feature amount. The region dividing step S310, which is a feature of this embodiment, will be described next.

In region division step S310, region division processing is executed in accordance with the following processing order from the original image data and the initial region signal from initial region generation step S300.
First, an average value of all feature values for each area of the initial area is calculated. The feature amount here is a total of five pieces of information including three color information of pixels (tristimulus values (X ′, Y ′, Z ′)) and two pieces of position information (coordinates (I, J)). is there.
Next, for each pixel of the entire screen of the original image data, it is determined by the following evaluation function (5) which region is closest to the average value of the feature amount of the region as a whole. .

Dk_mean
= Kl {(X'-Xm) ²
+ (Y′−Ym) ²
+ (Z′−Zm) ² }
+ Kp {(I-Im) ²
+ (J−Jm) ² } (5)
However,
Xm, Ym, Zm: average value in each area of the color information X, Y, Z,
Im, Jm: average value in each area of the position information I, J,
Kl, Kp: These are coefficient values, and all feature quantities are in the same range,
For example, it is multiplied to be 0 to 1,
It is.

  Next, an evaluation function is obtained for all regions with respect to the feature amount of each pixel. The region where the obtained evaluation function is the smallest is considered as the optimum region, and the target pixel is included in the optimum region, and a region is newly constructed. After performing the minimum value search using the above evaluation function for all the pixels and reconstructing the region, the average values Xm, Ym, Zm, Im, and Jm of each feature amount are recalculated for each newly reconstructed region. To do. The above operation is performed again for all pixels. All processes are repeated until the average values Xm, Ym, Zm, Im, and Jm of each region do not change.

(2-3. Identifier Assignment Method (Step S250))
Here, a processing method applied to the identifier assigning method will be described. Processing, based on the preselected various data, are sequentially performed in the following manner.

First , light source information obtained from shooting information is acquired and set as a shooting light source. Various light sources are assumed in the situation where the image was taken. For example, A light source (tungsten light source), black body radiation light source assuming a sunny day (color temperature of about 7,500 [K] to 10,000 [K]), F1-12 light source assuming indoor fluorescent lamp, etc. It is. In this embodiment, the light source recorded in the tag name of the light source in the shooting information of the image file is read, and the C light source: 6,740 [K] is set as Ddst.

Then select one type of light source which is a target in the images processing device, and sets this as the target source. Examples of target light sources include daylight colors such as D55 (relative color temperature 5,500 [K]) and D65 (6,500 [K]), and C light sources (6,740 [K] as artificially created simulator light sources). ]). In this example, an sRGB space is assumed as the standard signal of the monitor, and the tristimulus value XYZ _Dsrc at 6,500 [K] of the D65 light source as the light source and the luminance of the phosphor: _{80 cd} / m ² is calculated. A calculation example is shown in Table 3. A Macbeth chart was used as a photographing sample, and RGB values of image data photographed with a C light source were converted into tristimulus values with a D65 light source. The formula used for the calculation is shown in the following (6) to (8).

Here, r _src , g _src , b _src , r _dst , g _dst , b _dst are the response values of the cones in each light source (src, dst).
[M _Pitt ] is a chromaticity coordinate of the basic primary color and is represented by the following equation (*).

Subsequently , the output value of the image signal estimated for the area under the photographing light source set first is adapted to the appearance of the color under the photographing light source, and the tristimulus value XYZ _Ddst (see “2-2. Xm, Ym, Zm) of step S240) ”is converted into XYZ ′ _Dsrc (Xm ′, Ym ′, Zm ′ of procedures (i) to (v) described later) of the appearance of color under different light sources. A Von · Kries model is applied to this color appearance calculation method under different illumination. The calculation here follows the procedures (i) to (v) below.
It should be noted that various models (CIECAM97s, FairChild model) can be applied to the appearance of color under different illuminations at this time.

As procedure (i), basic tristimulus values (L ₀ , M ₀ , S ₀ ) under the light sources Dsrc and Ddst are calculated. The symbol described after the basic tristimulus value indicates light source information. The calculation formula at this time is as follows.
L ₀ = + 0.3982 · Xm
+0.7040 ・ Ym
−0.0804 · Zm,
M ₀ = -0.2268 · Xm
+1.1679 ・ Ym
+ 0.0458 · Zm,
S ₀ = + 0.08458 · Zm

As the procedure (ii), the ratio of the calculated values, that is, the von Kries coefficient (k _L , k _M , k _S ) is calculated. The calculation formula at this time is as follows.
k _L = L _{m, Dsrc} / L _{m, Ddst} ,
k _M = M _{m, Dsrc} / M _{m, Ddst} ,
k _S = S _{m, Dsrc} / S _{m, Ddst}

As a procedure (iii), the basic tristimulus values (l ′, m ′, s ′) of the tristimulus values (Xm, Ym, Zm) of each divided region are calculated. The calculation formula at this time is as follows.
l '= + 0.3982 · Xm
+0.7040 ・ Ym
−0.0804 · Zm,
m ′ = − 0.2268 · Xm
+1.1679 ・ Ym
+ 0.0458 · Zm,
s' = +0.08458 ・ Zm

As a procedure (iv), basic tristimulus values (L ′, M ′, S ′) including adaptation are calculated. The calculation formula at this time is as follows.
_{L '= k L · l'} ,
M ′ = k _M · m ′,
S '= k _S · s'

As the procedure (v), the tristimulus value (Xm ′, Ym ′, Zm ′) is calculated. The calculation formula at this time is as follows.
Xm '= +1.86950 ・ L'
− 1.126900 ・ M '
+ 0.23873 · S ',
Ym '= +0.36304 ・ L'
+ 0.63740 ・ M '
-0.00001 · S ',
Zm '= +1.18231 ・ S'

Then , the above-described tristimulus value XYZ _Dsrc value and tristimulus value XYZ ′ _Dsrc are compared. The identification number of the closest XYZ _Ddst value is _{assigned as} an identifier (estimation information No). The comparison at this time is as follows:
ΔD = {(x0−Xm ′) ²
+ (Y0−Ym ′) ²
+ (Z0−Zm ′) ² } ^0.5 (9)
The relative distance ΔD in the color space shown in FIG.
In addition, as an example of color information including chromatic colors including color information of memory colors stored in the storage unit in advance, publicly known information other than the above Munsell color chart can be used. For example, Macbeth Color Checker Spectral reflectance color charts defined by CIE 13.3, spectral reflectance color charts defined by SOCS (JIS-TR X 0012, Standard Object Color Spectroscopy Database for Color Reproduction Evaluation (SOCS) (1998)), etc. It is done.

Also, a physical quantity for each divided area is calculated as a feature quantity of the original image data. As the calculation formula at this time, the following formulas (10) and (11) can be exemplified.
s (t) = {Pix (t) / height · width} (10),
Φ (t) = {l (t)} ² / {s (t) · 4π} (11)
However,
s (t): Area ratio [%] in the entire image of the region (t)
l (t): Perimeter of area (t) [pix]
Pix (t): Number of pixels belonging to region (t) [pix]
height: height of original image data [pix]
width: width of original image data [pix]
It is.
By performing the above processing, it is possible to obtain a table in which the column “identifier” in Table 2 and the feature amount of the region are described.

(2-4. Smoothing processing method (step S280))
A smoothing process is performed in accordance with the region division, identifier assignment information, and shooting information. The details are described below.
In the present embodiment, an example in which smoothing processing is applied to the original image data FIG. 2A using the shooting information and the statistical information data of the image area (Table 2) will be described. The resulting smoothed image is shown in FIG. The divided regions to be subjected to the smoothing process are smoothed for the region (a) having the largest memory area and the largest area in the image in Table 2. Details of the smoothing filter threshold value calculation method will be described below.

The threshold value of the smoothing filter is calculated from the shooting mode of the Exif information and the shutter speed. In this embodiment, a threshold value for the shutter speed is set.
As the image file photographing condition, the threshold THR of the smoothing filter is calculated by the following equation (12) with respect to the shutter speed.
THR = f (B, S. speed) (12)
However,
S. speed: shutter speed B: B data.

Since the value of the B data of the independent variable of equation (12) and the shutter speed can each be in the range of about 0 to 255, 1/10 to 1 / 5,000, the value of the threshold value THR using these values is In the case of simple division, a value on the order of 10 ¹ to 10 ⁴ is taken. In addition, it is also preferable that the calculation formula at this time is calculated using a unit of APEX (Additive System of Photographic Exposure). The relationship between APEX and shutter speed is basically the relationship of the following formula (**).
Ev = Av + S. speed (Tv)
= Bv + Sv (**)
here,
Av: Aperture value,
Bv: luminance value,
Sv: ISO sensitivity.
Note that, as shown by the equation (12), for calculating the threshold value THR using these independent variables, it is possible to use a linear linear equation, a nonlinear polynomial, or the like corresponding to the photographing mode. Furthermore, it is also preferable that this threshold value calculation method stores in advance a fixed value in an image file in a table format.

According to the above processing method, the original image data of the image file 100 generated by the electronic camera described above does not contain noise, can achieve high image quality, and is subjected to processing excellent in reproducing the memory color. It is possible to output an image.
In this embodiment, the personal computer 11 performs image processing and the color printer 13 does not perform image processing. This operation is realized by providing the image processing function described in this embodiment to an image data processing application (program) such as a retouch application or a printer driver installed in a hard disk of a computer.

  As another embodiment, all image processing can be executed by the color printer 13. Alternatively, a part may be executed on a computer or a server via a network. The image file 100 generated by the electronic camera 11 may be in the form of providing data directly to the printer via a memory card.

(Overview of Example 3)
In the third embodiment, in the image processing system configuration having the apparatus configuration of FIG. 1, the original image data captured by the electronic camera 11 is subjected to image processing according to the present invention by the personal computer 12 and output by the color printer 13. It is description about a form.
In the present embodiment, the image processing system, the image processing method, and the original image data dividing method are the same as those in the previous embodiment, and thus the description thereof is omitted.

FIG. 11 is a diagram showing original image data in the present invention.
The original image data applied in the present embodiment is shown in FIG. 11- (a), in which the “night scene mode” that is the shooting mode of the shooting information is extracted, and the area where the image texture level is low in the area that is the background area This is an example in which smoothing is performed on the empty areas (A), (B), (C), and (D) according to the threshold value calculated from the photographing information.
Here, a processing method applied to the identifier assigning method will be described. Processing, based on the preselected various data, are sequentially performed in the following manner.

(3-1. Identifier assignment method)
First , light source information obtained from shooting information is acquired and set as a shooting light source. Various light sources are assumed in the situation where the image was taken. For example, there are an A light source (tungsten light source), a black body radiation light source assuming a shade on a sunny day (color temperature of about 7,500 to 10,000 [K]), an F1 to 12 light source assuming an indoor fluorescent lamp, and the like.
In this embodiment, the light source recorded in the tag name of the light source in the shooting information of the image file is read, and the flash light source: 7,000 [K] is set as Ddst.

Next , one type of target light source in the image processing apparatus is selected and set as a target light source. Examples of target light sources include daylight colors such as D55 (relative color temperature 5,500 [K]) and D65 (6,500 [K]), and C light sources (6,740 [K] as artificially created simulator light sources). ]). In the present embodiment, an sRGB space is assumed as a standard signal of the monitor, the tristimulus value XYZ _Dsrc of the Munsell color chart (Munsell book) at 6,500 [K] of the D65 light source as the light source and the luminance of the phosphor: _{80 cd} / m ² . Was calculated. An example of calculation is shown in Table 3 above.

Subsequently , the output value of the image signal estimated for the area under the photographing light source set first is adapted to the appearance of the color under the photographing light source, and the tristimulus value XYZ _Ddst (see “2-2. Xm, Ym, Zm) of step S240) ”is converted into XYZ ′ _Dsrc (Xm ′, Ym ′, Zm ′ of the procedures (i) to (v)) of color appearance under different light sources. A Von · Kries model is applied to this color appearance calculation method under different illumination. The calculation here follows the procedures (i) to (v).
Note that various models (CIECAM97s, FairChild model) can be applied to the color appearance under different illuminations at this time. Since this calculation method is the same as that of the first embodiment, the description thereof is omitted.

Then , the above-described tristimulus value XYZ _Dsrc value and tristimulus value XYZ ′ _Dsrc are compared. The identification number of the closest XYZ _Ddst value is _{assigned as} an identifier (estimation information No). The comparison at this time includes the relative distance ΔD in the color space shown in the above equation (9).
In addition, as an example of color information including chromatic colors including color information of memory colors stored in the storage unit in advance, publicly known information other than the above Munsell color chart can be used. For example, Macbeth Color Checker Spectral reflectance color charts defined by CIE 13.3, spectral reflectance color charts defined by SOCS (JIS-TR X 0012, Standard Object Color Spectroscopy Database for Color Reproduction Evaluation (SOCS) (1998)), etc. It is done. Also, a physical quantity for each divided area is calculated as a feature quantity of the original image data. Examples of the calculation formula at this time include the above-described formulas (10) and (11).

Here, a method for calculating a so-called texture amount perception amount generated by repeating a certain element pattern as a feature amount of original image data will be described.
First, the Y data of the original image data is differentiated by the Laplacian operator of FIG. 12A, and the obtained data is quantized by the threshold value T. As this threshold value T, the equation (13) can be cited.
Next, the number of partial windows is calculated by applying the 9 × 9 [pixel] nonlinear operator of FIG. 12- (b) to the obtained secondary differential image data. The value calculated in this way is the texture degree ε (t) [%]. The partial window number calculation process is applied to the entire range of the divided area, and the texture amount σ (t) is calculated by Expression (14). Through the above processing, it is possible to obtain a table in which the column “identifier” in Table 4 and the feature amount of the region are described.
T = (Mode value of differential value)
+ (Standard deviation of differential value) × 1.4 (13),
σ (t) = Σ {ε (t)} (14)

(3-2. Smoothing processing method)
A smoothing process is performed in accordance with the region division, identifier assignment information, and shooting information. The details are described below.
In this embodiment, an example will be described in which smoothing processing is applied to original image data FIG. 11- (a) using shooting information and statistical information data of an image area (see Table 4).
Since the smoothing filter is the same as that of the second embodiment, the description thereof will be omitted, and the threshold value THR will be described below.

The threshold value of the smoothing filter is calculated from the shooting mode of the Exif information, the ISO sensitivity, and the feature amount of the image. In this embodiment, threshold values for ISO sensitivity and image texture amount are set.
The threshold value THR is calculated by equation (15).
THR = f (B, I. Sensitivity)
+ G (B, σ (t)) (15)
However,
I. Sensitivity: ISO sensitivity B: B data σ (t): Texture amount [%]
It is.

Since the value of the B data of the independent variable of Formula (15) and the ISO sensitivity can be in the range of about 0 to 255 and 100 to 1,600, respectively, it is assumed that the texture amount can take the range of up to 5%. When the threshold value THR using these is expressed by simple division, the first term is approximately 2 · 10 ⁻¹ to 2 · 10 ³ , and the second term is approximately 2 · 10 ⁻³ to 2 -Take a value of 10 ^-1 order. In addition, it is also preferable that the calculation formula at this time is calculated using a unit of APEX (Additive System of Photographic Exposure). The relationship between APEX and ISO sensitivity basically consists of the following relationship (***).
Ev = Av + Tv
= Bv + I. Sensitivity (Sv) (***)
here,
Av: Aperture value,
Tv: Shutter speed,
Bv: a luminance value.
Note that, as shown in the equation (14), for calculating the threshold value using these independent variables, a linear linear equation, a nonlinear polynomial, or the like corresponding to the photographing mode can be used. Furthermore, it is also preferable that the threshold value calculation method stores in advance a fixed value in an image file in a table format.

According to the above processing method, the original image data of the image file 100 generated by the electronic camera described above does not contain noise, can achieve high image quality, and is subjected to processing that is excellent in reproducing the memory color. Can be output.
In the above embodiment, the personal computer 11 performs image processing and the color printer 13 does not perform image processing. This operation is realized by providing the image processing function described in this embodiment to an image data processing application (program) such as a retouch application or a printer driver installed in a hard disk of a computer.

  As another embodiment, all image processing can be executed by the color printer 13. Alternatively, a part may be executed on a computer or a server via a network. The image file 100 generated by the electronic camera 11 may be in the form of providing data directly to the printer via a memory card.

(Summary of Example 4)
In this embodiment, in the image processing system configuration including the apparatus configuration of FIG. 13, original image data captured by an electronic camera is transmitted to a color printer connected via a wireless LAN and directly output. It is an explanation.
In the present embodiment, the image processing system, the image processing method, the original image data dividing method, and the method for assigning an identifier to each area of the divided original image data are the same as in the previous embodiment. The description is omitted.
Also, the original image data applied as an example is shown in FIG. 14- (a), the shooting mode of the shooting information is “landscape mode”, and the memory color is set according to the sizes of the input image data and the output image data. This is an embodiment in which appropriate smoothing is applied to the empty area (A), which is an area.

Hereinafter, embodiments of the present invention will be described based on examples.
Note that the generation of the image file and the configuration of the image file are the same as those in the first embodiment, and thus description thereof is omitted.

(4-1. Configuration of image processing system)
FIG. 13 is a diagram showing a schematic configuration of an image processing system according to the present invention.
The image processing system 10 of this embodiment includes an electronic camera 11 as an image file generation device, and a color printer 13 as an image processing device and an image output device.
The electronic camera 11 can set various shooting information. Shooting information means acquisition conditions of original image data such as shutter speed, exposure, and aperture. The electronic camera 11 performs shooting (original image data generation) under shooting conditions set by the photographer, and generates an image file that integrally includes image data and shooting information. The generated image file is stored in a recording medium (for example, a memory card).
After obtaining the image file, the color printer 13 performs image region division, identifier assignment, smoothing processing, and finally output processing. The image file can be transmitted from the electronic camera 11 to the color printer 13 via an interface such as a LAN or USB cable, or data can be directly transmitted via a wireless LAN or the like. A known output format can be applied to the output operation of the color printer at this time.

(4-2. Identifier assignment method)
Below, after mentioning the optimization method applied to the identifier assigning method, the identifier assigning method will be described. The optimization process is executed based on various data selected in advance. In the optimization process, various data necessary for optimization are selected in advance.
First , light source information obtained from shooting information is acquired and set as a shooting light source. Various light sources are assumed in the situation where the image was taken. For example, there are an A light source (tungsten light source), a black body radiation light source assuming a shade on a sunny day (color temperature of about 7,500 to 10,000 [K]), an F1 to 12 light source assuming an indoor fluorescent lamp, and the like.
In this embodiment, the light source recorded in the tag name of the light source in the shooting information of the image file is read, and the B light source: 4,870 [K] is set as Ddst.

Next , one type of target light source in the image processing apparatus is selected and set as a target light source. Examples of target light sources include daylight colors such as D55 (relative color temperature 5,500 [K]) and D65 (6,500 [K]), and C light sources (6,740 [K] as artificially created simulator light sources). ]).
In the present embodiment, an sRGB space is assumed as a standard signal of the monitor, the tristimulus value XYZ _Dsrc of the Munsell color chart (Munsell book) at 6,500 [K] of the D65 light source as the light source and the luminance of the phosphor: _{80 cd} / m ² . Was calculated. An example of calculation is shown in Table 3 above.

Subsequently , the output value of the image signal estimated for the area under the photographing light source set first is adapted to the appearance of the color under the photographing light source, and the tristimulus value XYZ _Ddst (see “2-2. Xm, Ym, Zm) of step S240) ”is converted into XYZ ′ _Dsrc (Xm ′, Ym ′, Zm ′ of the procedures (i) to (v)) of color appearance under different light sources. A Von · Kries model is applied to this color appearance calculation method under different illumination. The calculation here follows the procedures (i) to (v).
Note that various models (CIECAM97s, FairChild model) can be applied to the color appearance under different illuminations at this time. Since this calculation method is the same as that of the first embodiment, the description thereof is omitted.

Then , the above-described tristimulus value XYZ _Dsrc value and tristimulus value XYZ ′ _Dsrc are compared. The identification number of the closest XYZ _Ddst value is _{assigned as} an identifier (estimation information No). The comparison at this time includes the relative distance ΔD in the color space shown in the above formula (9).
In addition, as an example of color information including chromatic colors including color information of memory colors stored in the storage unit in advance, publicly known information other than the above Munsell color chart can be used. For example, Macbeth Color Checker Spectral reflectance color charts defined by CIE 13.3, spectral reflectance color charts defined by SOCS (JIS-TR X 0012, Standard Object Color Spectroscopy Database for Color Reproduction Evaluation (SOCS) (1998)), etc. It is done. Also, a physical quantity for each divided area is calculated as a feature quantity of the original image data. As a calculation formula at this time, the above-described formulas (10) and (11), the texture amount in the second embodiment, and the like can be exemplified.
By performing the above processing, it is possible to obtain a table in which the column “identifier” in Table 5 and the feature amount of the region are described.

A smoothing process is performed in accordance with the region division, identifier assignment information, and shooting information. The details are described below. In the present embodiment, an example will be described in which smoothing processing is applied to original image data FIG. 14- (a) using shooting information and statistical information data of image areas.
Since the smoothing filter is the same as that of the first embodiment, the description thereof will be omitted, and the description of the threshold value THR will be described below.

(Threshold acquisition method)
The smoothing filter threshold value is calculated from the Exif information shooting mode, shutter speed, image feature amount, and print settings. As a method for calculating the threshold value THR, the following equation (16) can be exemplified.
THR = f (B, S.speed)
+ H (O.size / I.size) (16)
However,
S. speed: Shutter speed B: B data O.D. size: output size size: Input size.

The value of the B data of the independent variable of the equation (16) and the range that the shutter speed can take are the same as those in the first embodiment, and the description is omitted. Here, when the threshold value THR using these is expressed by simple division, the first term has a value of about 10 ¹ to 10 ⁴ , and the second term has a value of about 10 ⁻² to 10 ^2. take. In addition, it is also preferable that the calculation formula at this time is calculated using a unit of APEX (Additive System of Photographic Exposure). The relationship between APEX and ISO sensitivity basically consists of the following relationship (***).
Ev = Av + Tv
= Bv + I. Sensitivity (Sv) ... (***)
here,
Av: Aperture value,
Tv: Shutter speed,
Bv: a luminance value.
Note that, as shown by the equation (16), for calculating the threshold value using these independent variables, a linear linear equation, a nonlinear polynomial, or the like corresponding to the shooting mode can be used. Furthermore, it is also preferable that the threshold value calculation method stores in advance a fixed value in an image file in a table format.

According to the above processing method, the original image data of the image file 100 generated by the electronic camera 11 described above does not contain noise, can achieve high image quality, and is subjected to processing excellent in reproducing the memory color. An image can be output.
In the above embodiment, the color printer 13 performs image processing. This operation is realized in advance by providing the image data processing application (program) with the image processing function described in this embodiment in a storage area such as the ROM of the printer.
Further, as another embodiment, there is a form similar to that of the first embodiment, that is, a form in which the image enlargement process is appropriately performed by the personal computer 11 and the color printer performs only the output process.

(effect)
As is clear from the above description, an image generation step of generating output image data from an image file composed of original image data and shooting information, and light source information and sensitivity information of the shooting information in the image generation step With the smoothing process for performing the smoothing process based on the above, it is possible to generate a highly sensitive image file while reducing noise.

  Further, the smoothing step includes a calculation step of calculating an image feature amount of the original image data, a calculation step of calculating a threshold value of a noise removal filter based on the feature amount of the image, and the calculation result. By providing the second smoothing step, it is possible to generate a highly sensitive image file while reducing noise.

  In addition, the calculation step includes a division step of dividing the original image data, and statistics such as a centroid value, frequency value, and standard deviation of color information (tristimulus values) of the original image data for the divided region. By having a second calculation step that calculates at least one physical quantity such as the quantity, area ratio, and texture level, it becomes possible to perform an appropriate smoothing process according to the characteristics of the image.

  In addition, the calculation step includes an identifier based on a determination step of determining a priority order based on a statistic amount and a physical amount of the original image data, and photographing information and color information related to a memory color stored in advance for the divided areas. It is possible to perform an appropriate smoothing process in accordance with the characteristics of the image by including an applying step for applying the image and an estimating step for estimating a divided region that is more important than the results of the determining step and the applying step. Become.

  The adding step calculates a relative distance between the color information regarding the memory color and color information (tristimulus values) in the light source information of the photographing information, and the color information in the divided area and the memory color. By including the third calculation step and the determination step of determining the identifier having the smallest relative distance as the identifier of the divided region, smoothing suitable for the color reproduction of the memory color can be performed.

1 is a diagram illustrating a schematic configuration of an image processing system according to the present invention. It is a figure which shows the original image data in this invention (the 1). It is a figure which shows the original image data in this invention (the 2). It is a figure which shows notionally an example of a structure of the image file in this invention. It is a figure which shows an example of the imaging | photography information data in this invention. It is a flowchart which shows the image processing method in this invention. It is a conceptual diagram which shows the method of calculating the characteristic value on xy chromaticity diagram in this invention. It is a figure which shows the concept of the smoothing filter process in this invention. It is a flowchart which shows the image processing method in this invention. It is a flowchart which shows the method of the area | region division in this invention. It is a figure which shows the original image data in this invention. It is a figure which shows the Laplacian calculation of the original image data in this invention, and a nonlinear calculation. 1 is a diagram illustrating a schematic configuration of an image processing system according to the present invention. It is a figure which shows the original image data in this invention. It is a figure which shows the correlation of a sensitivity and the noise amount perceived. It is a flowchart which shows the principal part of an electronic camera. It is a characteristic view which shows the relative sensitivity with respect to three primary colors of an image signal.

Explanation of symbols

11 Electronic camera (digital camera)
12 Personal computer (information processing equipment)
13 Color Printer 100 Image File 101 Original Image Data Storage Area 102 Shooting Information Storage Area

Claims (2)

An image processing method for generating output image data based on an image file composed of original image data and shooting information of the original image data,
An original image data dividing step for dividing the original image data, and a statistic calculating step for calculating at least one of a statistic and a physical quantity of color information (tristimulus values) of the original image data for each of the divided regions; An image feature amount calculating step for calculating an image feature amount of the original image data,
A threshold value calculating step for calculating a threshold value of a noise removal filter based on the sensitivity information included in the photographing information and the calculated image feature amount ;
A priority order determining step for determining a priority order of smoothing based on a statistic or a physical quantity of the original image data, and light source information included in the photographing information for each of the divided areas and a color relating to a memory color stored in advance An identifier providing step for assigning an identifier from information, and an important region estimation step for estimating a divided region to be emphasized based on the results of the priority determination step and the identifier assigning step, and based on the calculated threshold value Smoothing process for smoothing,
An image processing method comprising: The identifier giving step includes
A color information conversion step of converting color information relating to the memory color into color information (tristimulus values) in light source information of photographing information;
A relative distance calculating step of calculating a relative distance of color information in each region divided by the memory color and the dividing step;
An identifier determination step for determining an identifier having the smallest relative distance as an identifier of each region divided by the division step;
The image processing method according to claim 1, further comprising :