JP4752421B2 - Image retouching method, image retouching device, and image retouching program - Google Patents

Description

  The present invention relates to an image retouching method, an image retouching apparatus, and an image retouching program, and more particularly to RAW data processing.

  Conventionally, a RAW data format is known as a data recording format of an image input device such as a digital camera. RAW data having only one color component per pixel is incomplete information as an image representing an object. Therefore, an image representing an object is displayed until a demosaic process (also referred to as a color filter array interpolation process) for interpolating at least the missing color component of each pixel with a color component of a neighboring pixel is performed on the RAW data. I can't print. The advantage of the RAW data format over image data formats such as JPEG and TIFF is that a preferred image representing the object can be generated after the object is recorded without being restricted by the continuous shooting interval, hardware resources, etc. It is. A system for generating an image representing an object from RAW data after the object is recorded generally includes retouch processing functions such as color balance correction, luminance correction, and sharpness correction. In such a system, every time a retouch operation value is input by the user, a preview image representing the result of the retouch process corresponding to the input retouch operation value is displayed. In order to shorten the response time of the preview display with respect to the input of the retouch operation value, it is preferable to execute the demosaic process before the input of the retouch operation value.

  By the way, in the white balance correction process, the most accurate correction result is obtained by multiplying the RAW data, which directly represents the amount of light transmitted through each color filter of the image sensor provided in the digital camera, for each chromatic component. Obtainable. In the brightness correction process for recovering overexposure and underexposure, the most accurate correction result can be obtained by multiplying the raw data value that directly represents the amount of light transmitted through each color filter of the image sensor. Obtainable.

  However, in the demosaic processing, the color component that is lacking in each pixel is interpolated with the color components of neighboring pixels, so false information that does not represent the amount of light transmitted through each color filter of the image sensor is added. Is done. Therefore, when white balance correction processing and luminance correction processing are performed after demosaic processing, there is a problem that an accurate correction result cannot be obtained or a very complicated calculation process is required to obtain an accurate correction result. is there.

JP-A-2005-33468

  The present invention was created to solve the above-described problem, and an image retouching method and an image retouching device capable of performing accurate retouching with simple processing, with a short response time for preview display in response to input of a retouching operation value. And an image retouching program.

(1) An image retouching method for achieving the above object includes obtaining RAW data stored in a storage medium, generating an intermediate image from the RAW data using demosaic processing, and generating the intermediate image after generating the intermediate image. When a retouching operation value is set according to the setting operation, and the color balance operation value that is the retouching operation value is set, display is performed from the intermediate image using a color balance backward correction process according to the color balance operation value. When an execution request is generated after the setting operation is generated using a color balance forward correction process according to the color balance operation value and a demosaic process after the color balance forward correction process, Generating a final image from the RAW data.
By generating an intermediate image from RAW data using demosaic processing before the setting operation, a preview image with the color balance corrected is generated from the intermediate image within a short time after setting the color balance operation value. Can do. When an execution request is generated, the color balance can be accurately corrected by correcting the color balance before the demosaic process and executing the demosaic process after the color balance correction. Further, when the color balance is corrected before the demosaic process, the color balance can be accurately corrected with a simple process compared to the case where the color balance is corrected after the demosaic process. Further, when the color balance is corrected before the demosaic process, the processing target data is 1/3 of that after the demosaic process, so that a high-speed process is possible. In other words, by changing the execution order of demosaic processing and color balance correction processing at the time of preview image generation and final image generation, the response time of preview display to the input of retouching operation values is shortened, and simple processing is accurate. It becomes possible to generate the final image retouched in a high speed.

(2) In the image retouching method, when a luminance operation value that is the retouch operation value is set, the preview image is generated from the intermediate image by using a lightness backward correction process according to the luminance operation value. When an execution request is generated after the setting operation, the final image may be generated from the RAW data by using a luminance forward correction process according to the luminance operation value before the demosaicing process.
By generating an intermediate image from RAW data using demosaic processing before the setting operation, it is possible to generate a preview image whose lightness is corrected within a short time after setting the luminance operation value. When an execution request is generated, the brightness can be corrected accurately by correcting the brightness before the demosaic process and executing the demosaic process after the brightness correction. Further, when the luminance is corrected before the demosaic process, the luminance can be accurately corrected by a simple process compared to the case where the luminance is corrected after the demosaic process. In addition, when the luminance is corrected before the demosaic process, the processing target data is 1/3 of that after the demosaic process, so that a high-speed process is possible. That is, by changing the execution order of demosaic processing and luminance correction processing at the time of preview image generation and final image generation, the response time of the preview display to the input of the retouch operation value is shortened, and accurate processing is performed with simple processing. It becomes possible to generate the retouched final image at high speed.

  (3) An image retouching apparatus for achieving the above object includes a RAW data acquisition unit that acquires RAW data stored in a storage medium, and an intermediate image generation that generates an intermediate image from the RAW data using demosaic processing. Means, setting means for setting a retouch operation value according to the setting operation after the generation of the intermediate image, and setting a color balance operation value as the retouch operation value, a color corresponding to the color balance operation value Preview image generation means for generating a preview image for display from the intermediate image using a balance rearward correction process, and when an execution request occurs after the setting operation, the color balance front according to the color balance operation value Using the correction process and the demosaic process after the color balance forward correction process, Comprising the final image generation means for generating an image.

  (4) An image retouching program for achieving the above object includes a RAW data acquisition unit that acquires RAW data stored in a storage medium, and an intermediate image generation that generates an intermediate image from the RAW data using demosaic processing. Means, setting means for setting a retouch operation value according to the setting operation after the generation of the intermediate image, and setting a color balance operation value as the retouch operation value, a color corresponding to the color balance operation value Preview image generation means for generating a preview image for display from the intermediate image using a balance rearward correction process, and when an execution request occurs after the setting operation, the color balance front according to the color balance operation value Using the correction process and the demosaic process after the color balance forward correction process, the RAW data A final image generation means for generating a Luo final image, it causes a computer to.

  In addition, the order of each operation | movement of the method described in the claim is not limited in order of description, unless there is a technical obstruction factor, Each operation | movement may be performed in what order, and simultaneously May be executed. In addition, each function of the plurality of means provided in the present invention is realized by a hardware resource whose function is specified by the configuration itself, a hardware resource whose function is specified by a program, or a combination thereof. The functions of the plurality of means are not limited to those realized by hardware resources that are physically independent of each other. The present invention can also be specified as an invention of a recording medium on which an image retouching program is recorded.

Hereinafter, embodiments in which the present invention is applied to a personal computer (PC) will be described in the following order.
1. 1. Configuration of image retouching device 2. Operation of image retouching device 2-1. Preview image generation process triggered by target file selection 2-2. Setting of retouch operation value ・ Setting of white balance operation value ・ Setting of brightness operation value 2-3. Preview image generation processing triggered by setting of the retouch operation value 2-4. 2. Final image generation processing Other embodiments

1. Configuration of Image Retouching Device FIG. 2 is a block diagram showing a schematic configuration of the PC 1 and its peripheral devices constituting the image retouching device. The PC 1 is used to acquire RAW data from an image input device such as the digital camera 6 or the image scanner 5 via a communication line or a removable memory, and generate a preferable image from the acquired RAW data.

  The interface unit 16 includes a USB controller, a USB connector, and the like for communicating with external devices such as the digital camera 6, the image scanner 5, and the printer 4. The RAW data generated by the image input device such as the digital camera 6 and the image scanner 5 can be once stored in the memory of the image input device and then directly input to the PC 1 by communication using the interface unit 16. That is, the interface unit 16 can function as a RAW data acquisition unit. The interface unit 16 may employ a configuration corresponding to all communication standards such as IEEE 1394, IrDA (Infrared Data Association), and Bluetooth. Further, the RAW data file stored in the recording medium of the external system may be input to the PC 1 via a removable memory or the Internet.

  The hard disk device (HDD) 18 stores an image retouching program and an OS for causing the PC 1 to function as an image retouching device. The HDD 18 can store a RAW data file to be retouched. The image retouching program may be transferred to the HDD 18 from a computer-readable recording medium such as a removable memory, or may be transferred to the HDD 18 via a communication network from a remote server computer.

  The CPU 14 executes programs stored in the ROM 20 and the HDD 18 to control each part of the PC 1 and peripheral devices. When executing the image retouching program, the CPU 14 functions as an intermediate image generating unit, a setting unit, a preview image generating unit, and a final image generating unit. The CPU 14 operates according to commands input from the mouse 3 and the keyboard 2. The mouse 3 and the keyboard 2 are used, for example, for a setting operation for inputting a retouching operation value setting request or a retouching or printing execution request.

The RAM 12 is a volatile memory that temporarily stores programs and processing target data. Raw data acquired from the HDD 18, a removable memory, or a recording medium of an external system is stored in the RAM 12.
The display control unit 10 includes a graphic controller, a frame memory, and the like, and displays a display target image such as a preview image stored in the frame memory on the display 7.

2. Operation of Image Retouching Device FIG. 1 is a flowchart showing a flow of image retouching processing by the PC 1 during execution of an image retouching program.
FIG. 3 is a diagram showing a configuration of a screen displayed on the display 7 during execution of the image retouching program. Via this screen, a retouch target file selection operation, a preview image generated in response to the retouch target file selection operation, a retouch operation value setting operation, and a preview image generated in accordance with the retouch operation value display An input operation of a retouch execution request is performed.

2-1. Preview image generation process triggered by target file selection When the folder selection button 30 is clicked, a folder reference screen (not shown) is displayed. Thumbnail images 50 and 52 stored in each RAW data file in the folder designated by the user on the folder reference screen are displayed in the selection area 54. A thumbnail image is a low-resolution image of an object generated by an image input device such as a digital camera and recorded by RAW data.

  When file selection occurs, the CPU 14 activates display processing of a preview image generated according to the initial setting value (step S100). Specifically, for example, when one of the thumbnail images 50 and 52 in the selection area 54 is clicked, the CPU 14 sets the white balance operation value to “setting at the time of shooting” and sets the brightness correction operation value to “0EV”. To start the following preview image display process. It should be noted that although the description of how many bits are used in the process of generating the preview image and the final image from the RAW data is omitted, for example, the process of increasing the number of bits to avoid information loss due to overflow, the output A process of reducing the number of bits according to the format can be added as necessary.

  In step S102, the CPU 14 analyzes the target file and obtains information necessary for generating the preview image and the final image. Specifically, for example, digital camera model, area image sensor arrangement, resolution, number of bits per pixel, pre-gain information, optical black information, gamma correction information, white balance shooting settings, etc. The specified attribute information is stored in the RAM 12. These attribute information is generally recorded in a format unique to each manufacturer, but the DNG format is proposed as a general-purpose standard. In any case, even if it is a RAW data file generated by an image input device of any manufacturer, these attribute information can be specified by analyzing the file.

In step S104, the CPU 14 executes pre-gain processing. In the pre-gain process, all pixel values of the RAW data are linearly amplified based on the pre-gain information acquired from the file header.
In step S106, the CPU 14 executes an optical black compensation process based on the optical black information. The optical black information represents the amount of charge accumulated in the light receiving element that is not incident on the light provided in the image sensor of the image input device. In the optical black compensation processing, the value indicated by the optical black information acquired from the file header is regarded as black (for example, R = 0/255, G = 0/255, B = 0/255), and all the pixel values of the RAW data are used. The value indicated by the optical black information is subtracted.

  In step S108, the CPU 14 executes a demosaic process according to the arrangement of the color filters of the area image sensor. The demosaic process is basically a process of compensating for a lack of chromatic color components between neighboring pixels having only one chromatic color component (for example, any one of R, G, and B) per pixel. However, in order to suppress moire patterns, jaggies, suspicious colors, halation, etc. in demosaic processing, complicated matrix operations are required. The demosaic process is a process of adding a value that does not represent the light actually incident on the image sensor of the image input apparatus, and is generally an irreversible conversion. Therefore, if the amount of light incident on the image sensor of the image input device is unfavorable, the correction for recovering it is aimed at simpler processing before demosaic processing than before demosaic processing. You can get street results. However, in this embodiment, white balance correction is performed in order to unify the execution order of each correction process in the preview image generation process triggered by the target file selection and the preview image generation process triggered by the setting of the retouch operation value. Processing and brightness correction processing are executed after demosaic processing.

In step S110, the CPU 14 executes a suspicious color suppression process. In an area image sensor having a Bayer array color filter, the number of light receiving elements for G is twice the number of light receiving elements for R and B, and therefore aliasing is likely to appear in the R and B components. For this reason, the suspicious color generated is suppressed by a flattening process using a median filter or the like.
In step S112, the CPU 14 executes color matching processing. Since the value of each color component of the RAW data depends on the spectral sensitivity characteristic of the image sensor of the image input device, it does not correspond to the tristimulus values defined by general-purpose standards such as sRGB. The color matching process is a process for representing the color light represented by the RAW data that is an input value with a stimulus value defined by a general standard such as sRGB. Specifically, pixel values are converted using a 3 × 3 matrix operation or a three-dimensional LUT (Look Up Table). The matrix and the three-dimensional LUT used for this conversion can be acquired from the file header, or can be applied in accordance with the model information acquired from the file header, which is predefined according to the model.

  In step S114, the CPU 14 performs white balance correction based on the white balance information acquired from the file header. Although the spectral radiant energy distribution of an object is naturally different if the illumination environment is different, humans perceive the same object color as the same color even if the illumination environment is different. This tendency is called color constancy. On the other hand, the image sensor accumulates electric charges corresponding to each chromatic color component (described as RGB in this embodiment) in a balance according to the spectral radiant energy distribution of the object. For this reason, for example, color balance correction for correcting the color balance of each of the RGB components is required so that an object that is perceived as white in the daytime in a sunny day is represented by white pixels. This color balance correction is white balance correction. In the color balance correction, an operation for multiplying each RGB component basically by a constant using different constants for each RGB component is performed. However, in order to suppress gradation loss due to saturation, a partially nonlinear calculation can be performed. Specifically, for example, when white balance information “fluorescent light: white” is stored in the file header, the object should have been illuminated with a white fluorescent light, so that the entire image is slightly bluish A constant for correcting the balance is set.

  In step S116, the CPU 14 performs brightness correction, and stores a copy of the intermediate image generated as a result in another area of the RAM 12 for use in preview image generation processing triggered by setting of the retouch operation value. In this brightness correction, for example, when the physical exposure of the digital camera is inappropriate, an exposure correction operation after shooting is performed by the digital camera in order to recover it, and the exposure correction operation is performed according to the exposure correction operation. This is performed only when exposure correction information is stored in the file header. The luminance may be corrected by multiplying the luminance component by a constant, but the luminance may be corrected by multiplying each of the chromatic color components by a constant. Note that a copy of the intermediate image for use in the preview image generation process triggered by the setting of the retouch operation value may be stored in another area of the RAM 12 at an earlier stage (for example, immediately after the suspicious color suppression process in step S112). .

In step S118, the CPU 14 executes gamma correction. This gamma correction is a non-linear operation for linearly matching the brightness and output signal when the output signal of the image sensor does not correspond linearly to the brightness that is the amount of light incident on the image sensor. is there.
In step S120, the CPU 12 executes gradation reproduction processing. The brightness perceived by humans does not correspond to the brightness that is the photometric quantity of the image sensor. In general, the luminance is converted into a non-linearity based on Weber-Fechner's hypothesis that the sensory amount is proportional to the logarithm of the stimulus amount. Note that even in the gradation reproduction process, in order to avoid gradation loss due to saturation, conversion in which the gradation is gently suppressed in the highlight region can be performed.
Note that the processing of steps S114 to S120 may be executed sequentially, but a conversion formula obtained by combining the conversion formulas of the respective processing or an LUT for conversion can be set and executed at once using these. . As a result of the above processing, a preview image is generated according to attribute information that is an initial setting value acquired from the RAW file header.

In step S122, the PC 1 displays a preview image on the display 7. Specifically, for example, the preview image stored in the frame memory is displayed in the preview area 34 shown in FIG. 3 by the display driver. Incidentally, the number of pixels in the preview area 34 and the number of pixels in the RAW data are generally different. The number of pixels in the preview image matches the number of pixels in the preview area 34. Therefore, in the process of generating a preview image from RAW data, a process for converting the resolution is required. This resolution conversion process may be executed at this stage immediately before display, or may be executed at an earlier stage to reduce the number of pixels to be processed thereafter and increase the speed. Also, resolution conversion may not be performed and a part of the preview image may be displayed in the preview area 34.
When all the above processes are completed, a preview image is displayed on the display 7 as a response to the operation of selecting the retouch target file.

2-2. Retouch Operation Value Setting The PC 1 accepts a retouch operation value setting operation as follows, for example.
White Balance Operation Value Setting The white balance correction area 56 shown in FIG. 3 includes a drop-down menu 48, a color temperature slider 76, a color tone slider 70, a color temperature up / down button 72, a color tone up / down button 66, and a dropper activation check box 46. Has been placed. In the drop-down menu 48, an illumination light source at the time of shooting is displayed as a selection item. The initial value is “shooting setting”. “Shooting setting” is an item to which attribute information acquired from a file header is applied as a retouch operation value. Therefore, when “shooting setting” is selected, a constant for multiplying each chromatic color component by a constant is determined according to attribute information acquired from the file header. “Clear sky”, “Cloudy sky”, “Fluorescent lamp: White”, “Incandescent lamp”, “Flash”, etc., each chromatic color component (for example, R, G, etc.) according to the spectral radiant energy distribution of the light source corresponding to each item. This is an item in which a constant for multiplying B) by a constant is set. “Auto” is an item for designating that an optimum constant for multiplying each chromatic color component by a constant is obtained by image analysis.

  When “Custom” is selected, the color temperature slider 76, the color tone slider 70, the color temperature up / down button 72, the color tone up / down button 66 and the dropper activation check box 46 are enabled, and operation values can be set using these. Transition to mode. The color temperature slider 76 and the color temperature up / down button 72 are for setting the color temperature of the light source at the time of shooting as an operation value. When the color temperature of the light source is set as the operation value, a constant determined in advance according to the color temperature is set for each chromatic color component. The set color temperature is displayed in the text box 74. The color tone slider 70 and the color tone up / down button 66 are for setting a ratio between a constant multiplied by the B component and a constant multiplied by the G component as an operation value. The ratio is displayed step by step (for example, ± 10 steps) in the text box 68. When the dropper activation check box 46 is turned on, the pointer becomes a dropper shape, and a constant that sets the saturation of the clicked pixel to 0 is set for each chromatic color component. The white balance operation value in this case is the position of the clicked pixel in the preview image.

Setting of Brightness Operation Value In the exposure correction area 58 shown in FIG. 3, an EV slider 62 and an EV up / down button 64 are arranged. The EV slider 62 and the EV up / down button 64 are for setting EV (Exposure Value) as a luminance operation value. The brightness operation value is displayed in the text box 60. Every time EV increases by 1, a constant is set that doubles the brightness in terms of luminance, and every time EV decreases by 1, a constant that sets the brightness by half in terms of luminance is set. The EV is a concept of correcting the amount of light actually incident on the image sensor of the image input device, since the ratio of the actual measurement value and the reference value is expressed as a logarithm with a base of 2 when the exposure time is considered alone. The parameter used. In addition, the F value and shutter speed of the digital camera can be set in stages so that EV changes as an integer. Therefore, by using EV as an operation value for luminance correction for recovering overexposure and underexposure, it becomes easy for the user to understand how much luminance is corrected. When EV is set as the luminance operation value, the user's intention is to change the brightness to 1/4, 1/2, 2, 4 or 4 in terms of luminance. However, after performing non-linear gradation reproduction processing and gamma correction processing in this way, changing the brightness accurately in this way requires execution of these inverse transformation processing, so at least unless the response time is significantly sacrificed. Impossible.

2-3. Preview image generation process triggered by setting of the retouch operation value When a setting operation occurs after the retouch target file is selected, the PC 1 generates a new preview image from the preview image as follows (step S130).
In this embodiment, in order to shorten the image, a new preview image is generated from the intermediate image that has not been subjected to the gamma correction process and the gradation reproduction process. However, the preview image is generated only after the retouch target file is selected. Alternatively, white balance correction processing and luminance correction processing may be performed on the complete preview image. In any case, by executing white balance correction processing or luminance correction processing on an image that has been subjected to demosaic processing with a large amount of processing, a shorter time is required than when a preview image is generated from RAW data. The preview image can be updated within a short time.

In step S132, the CPU 14 executes white balance correction processing on the intermediate image copy stored in step S110. In this embodiment, since the white balance of the intermediate image whose white balance has already been corrected is corrected, the constant multiplied by RGB is affected by the constant used in the white balance correction applied to the intermediate image. . Specifically, for example, a constant uniquely determined by the operation value used for the R component according to the operation value “sunny” is k ₁ , and for the R component according to the operation value “incandescent lamp”. Let k ₂ be a constant that is uniquely determined by the operation value. When the constant k ₁ corresponding to “clear sky” has been applied to the intermediate image and the latest retouching operation value is “incandescent lamp”, the latest constant applied to the R component of the preview image is k _2. / K ₁ Note that the setting of the retouch operation value “incandescent lamp” is triggered for an intermediate image that has not been subjected to the white balance correction process (step S114) in the preview image generation process that is activated by the target file selection as a trigger. and in the case of executing the white balance correction processing, which is made constant to be applied to the R component of the intermediate image may be a k _2.

  In step S134, the CPU 14 executes a brightness correction process. In this embodiment, the brightness of the intermediate image whose brightness has already been corrected is corrected, and since EV is an operation value recognized based on the RAW data, the constants used for brightness correction are multiplied by the brightness correction of the intermediate image. Is affected by constants. Specifically, a constant obtained by multiplying the reciprocal of the constant multiplied with the intermediate image and a constant determined by the latest operation value for luminance correction is multiplied with the luminance component of the intermediate image. The luminance component may be multiplied by a constant, but each chromatic color component may be multiplied by a constant. Of course, when the brightness correction process is executed on the intermediate image that has not been subjected to the brightness correction process (step S116) in the preview image generation process activated by the target file selection as a trigger, the latest brightness correction process is performed. The luminance may be corrected using a constant that is uniquely determined by the operation value. In addition, if the user recognizes that the luminance operation value is an operation value based on the luminance of the currently displayed preview image, the intermediate image generated according to the luminance operation value set until immediately before It is sufficient to multiply the brightness component by a constant that is uniquely determined by the brightness correction operation value.

In step S136, the gamma correction process is executed as in step S116.
In step S138, the gradation reproduction process is executed in the same manner as in step S118.
As a result of the above processing, a preview image is generated (updated) according to the setting of the retouch operation value. Note that the processing of steps S132, S134, S136, and S138 may be executed sequentially, but a conversion formula obtained by combining the conversion formulas of each process or an LUT for conversion is set and executed at once using these. You can also

  In step S140, the PC 1 displays the updated preview image on the display 7. Specifically, for example, the preview image stored in the frame memory is displayed in the preview area 34 shown in FIG. 3 by the display driver. The response time from when the retouch operation value setting operation is performed until the preview image display is updated includes pre-gain processing (step S104), optical black compensation processing (step S106), color matching processing (step S108), and demosaic. Since the process (step S110) and the suspicious color suppression process (step S112) are performed on the intermediate image that has already been performed, the preview image corresponding to the set retouch operation value is generated from RAW data. Short.

2-4. Final image generation process The development start request is generated, for example, as follows. When the determination button 36 shown in FIG. 3 is clicked, the retouching operation value applied to the preview image displayed in the preview area 34 is saved in another area of the RAM 12 and reduced to indicate the result of the retouching process. A preview image 38 is displayed in the candidate area 40. When the development start button 42 is clicked, a dialog window (not shown) for setting the output format is displayed. In this dialog window, the final image size, color space, file format (JPEG, TIFF, etc.), compression rate (high quality, medium quality, low quality, etc.), folder for saving the final image, and the like are set. When the “OK” button is clicked in this dialog window, a development start request that is an execution request is generated, and the following processing for the RAW data corresponding to the reduced preview image 38 displayed in the candidate area 40 is activated. (Step S150).

In steps S152 and S154, the CPU 14 executes pre-gain processing and optical black compensation processing similar to those in steps S104 and S106.
In step S156, white balance correction processing is performed in which each component of R, G, and B is multiplied by a constant, as in step S114, using constants that are uniquely determined for each of R, G, and B depending on the retouching operation value. . However, since this process is performed before demosaic processing, a constant multiplication operation is performed on intermediate data in which only one component of R, G, or B exists for one pixel. Therefore, the constant applied to each pixel is only one of the three constants for each color component. Further, since the processing target data amount is 1/3 compared with after demosaic processing, the processing time is shortened. Since this processing is performed on intermediate data that does not represent the amount of light transmitted through each color filter of the image sensor, that is, so-called false information is not mixed, simply R, G, B By simply multiplying each component by a constant, the white balance is corrected exactly as intended.

  In step S158, the luminance correction process similar to that in step S116 is executed using a constant uniquely determined by the retouch operation value. Since this luminance correction processing is also performed before demosaicing processing, a constant multiplication operation is performed on intermediate data in which only one component of R, G, or B exists for one pixel. Further, since the processing target data amount is 1/3 compared with after demosaic processing, the processing time is shortened. Since this processing is performed on intermediate data that does not represent the amount of light transmitted through each color filter of the image sensor, that is, so-called false information is not mixed, simply R, G, B By simply multiplying each component by a constant, the brightness is accurately corrected as intended.

In steps S160, S162, S164, S166, and S168, the demosaic process, the suspicious color suppression process, the color matching process, the gamma correction process, and the gradation reproduction process similar to those in steps S108, S110, S112, S118, and S120 are executed. However, in the color matching process, color space conversion to the color space set for the final image is executed.
In step S170, the final image is generated by executing resolution conversion and encoding according to the size, file format, and compression rate set for the final image, and the final image is stored in the set folder.

3. Other Embodiments Although the example in which the present invention is applied to the PC 1 has been described above, the present invention can be applied to an MFP (Multi Function Printer) having a RAW data printing function and a stand-alone printer, The present invention can also be applied to a digital camera having a developing function after photographing processing. Further, although an example in which a preview image is displayed on the screen for each setting of the retouch operation value has been described, a preview image may be printed for each setting of the retouch operation value. Further, the example in which the execution order other than the white balance correction process and the brightness correction process is the same at the time of preview image generation and the final image generation has been described. The image generation and the final image generation may be interchanged. In addition, the execution order of the white balance correction process and the brightness correction process at the time of preview image generation may be switched, and the execution order of the white balance correction process and the brightness correction process at the time of final image generation may be switched. In addition, the execution order of processes other than those may be changed to an order different from that in the above embodiment at the time of preview image generation and at the time of final image generation.

The flowchart which concerns on one Embodiment of this invention. The block diagram which concerns on one Embodiment of this invention. The figure which shows the structure of the screen which concerns on one Embodiment of this invention.

Explanation of symbols

1: PC (image retouching device), 14: CPU (intermediate image generation means, setting means, preview image generation means, and final image generation means), 16: interface unit (RAW data acquisition means),

Claims (4)

RAW data acquisition means acquires RAW data ,
The intermediate image generation means performs demosaic processing on the RAW data, generates and stores an intermediate image,
The setting means sets the retouch operation value according to the operator 's setting operation,
A preview image display means corrects the demosaiced intermediate image based on the retouch operation value and displays it on the display screen;
A final image generation unit corrects the RAW data based on the retouch operation value, and performs a demosaic process on the corrected RAW data to generate a final image;
An image retouching method including: The image retouching method according to claim 1, wherein the retouch operation value is a color balance operation value and / or a luminance operation value. RAW data acquisition means for acquiring RAW data;
Intermediate image generating means for performing demosaic processing on the RAW data, generating and storing an intermediate image;
Setting means for setting a retouch operation value in accordance with a setting operation by an operator;
A preview image display means for correcting the demosaiced intermediate image based on the retouch operation value and displaying the corrected image on a display screen;
Final image generation means for correcting the RAW data based on the retouch operation value, and performing a demosaic process on the corrected RAW data to generate a final image;
An image retouching device comprising: RAW data acquisition means for acquiring RAW data;
Intermediate image generating means for performing demosaic processing on the RAW data, generating and storing an intermediate image;
Setting means for setting a retouch operation value in accordance with a setting operation by an operator;
A preview image display means for correcting the demosaiced intermediate image based on the retouch operation value and displaying the corrected image on a display screen;
Final image generation means for correcting the RAW data based on the retouch operation value, and performing a demosaic process on the corrected RAW data to generate a final image;
Image retouching program that causes a computer to function as a.

Images