JP2007249802A - Image processor and image processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processor for correcting an image to an appropriate image highly precisely by emphasizing a characteristic value of a face area which is highly likely to be a major subject even when a plurality of face areas can be detected in an image. <P>SOLUTION: The image processor is provided with: a face area characteristic value operating means 102 for calculating a face area average characteristic value from a face area included in inputted original image data; and a face area characteristic value correction amount calculating means 111 for calculating a face area characteristic value correction amount on the basis of a deviation between the calculated face area average characteristic value and preset target characteristic value. The face area characteristic value operating means 102 calculates as a face area average characteristic value a weighted average characteristic value obtained by weighting each average characteristic value of the plurality of face areas included in the original image data by each occupation area. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention provides a face area characteristic value calculating means for calculating a face area average characteristic value from a face area included in input original image data, and the calculated face area average characteristic value and a preset target characteristic value. The present invention relates to an image processing apparatus and an image processing method including a face area characteristic value correction amount calculating unit that calculates a face area characteristic value correction amount based on a deviation.

  Generally, a photographic image processing apparatus captures a photographed image formed on a photographic film as digital image data by a film scanner, or captures image data photographed by a digital photographing device such as a digital still camera via various media drive devices. Then, various image characteristics such as color characteristics and gradation characteristics of the captured original image data are corrected, and as a result, the photosensitive material (printing paper) is exposed based on the obtained output image data to generate a photographic print. It is configured as follows.

  In the above correction processing, in order to obtain a photographic print with good image quality, the difference between the gradation characteristics and color characteristics of the input original image data and the gradation characteristics and coloring characteristics of the photosensitive material, and human visual characteristics are taken into account. Then, density correction processing and color correction processing are performed so that appropriate gradation characteristics and color characteristics can be obtained.

  As such density correction processing, correction processing is generally performed so that the input image data and the output image data have a density characteristic curve showing a predetermined gamma characteristic. Proposes a technique for obtaining high-quality photo prints by correcting the gradation so that the face area of a person has appropriate brightness and color correction so that the face area has a preferable skin area. Has been.

  For example, Patent Document 1 proposes a technique for correcting a correction value obtained from a face average density value of a face area detected from captured image data and a density value of the entire image.

JP 2005-159387 A

  However, when a plurality of face images are detected in the original image, according to the above-described conventional technique, the average density (average characteristic value) is uniformly obtained for the plurality of face regions and reflected in the correction. As a result, the characteristic values of the person who is not the main subject and the person who is the main subject are processed at the same level, resulting in a problem that correction cannot be made as expected.

  Even when the correction value obtained from the characteristic value of the entire image and the correction value obtained from the face area are merged to obtain an appropriate correction value, the fusion processing is performed based on a predetermined parameter. Therefore, there is a problem that the correction is not necessarily performed properly.

  In view of the above-described conventional problems, the present invention performs correction with an emphasis on the characteristic value of a face area that is likely to be a main subject even when a plurality of face areas are detected in an image. Another object of the present invention is to provide an image processing apparatus and an image processing method capable of correcting an appropriate image with high accuracy.

  In order to achieve the above-mentioned object, the first characteristic configuration of the image processing apparatus according to the present invention is that, as described in claim 1 of the claims, the face region is included in the face area included in the input original image data. Face area characteristic value calculating means for calculating an area average characteristic value, and a face area characteristic value for calculating a face area characteristic value correction amount based on a deviation between the calculated face area average characteristic value and a preset target characteristic value An image processing apparatus comprising a correction amount calculating means, wherein the face area characteristic value calculating means weights each weighted average characteristic value of a plurality of face areas included in the original image data with each occupied area The average characteristic value is calculated as the face area average characteristic value.

  According to the above-described configuration, the face area average characteristic emphasizing a large face area among the plurality of face areas by weighting the average characteristic values of the plurality of face areas included in the original image data by the respective occupied areas. Calculate the value. If a photographic image contains multiple face areas, the large face is generally likely to be the main subject, so correction that emphasizes the characteristic values of the face area that is likely to be the main subject is important. It can be done.

  In the second feature configuration, as described in claim 2, in addition to the first feature configuration described above, a reference characteristic value for calculating a reference characteristic value correction amount based on all image characteristic values of the original image data. A correction amount calculation unit, a reference characteristic value correction amount calculated by the reference characteristic value correction amount calculation unit, and a face region characteristic value correction amount calculated by the face region characteristic value correction amount calculation unit at a predetermined fusion rate. There is an appropriate characteristic value correction amount calculating means for calculating an appropriate characteristic value correction amount by performing the fusion processing.

  According to the above configuration, the reference characteristic value correction amount obtained from the characteristics of the entire original image data and the face area characteristic value correction amount obtained from the characteristics of the entire face area data are fused at a predetermined fusion rate. It is possible to obtain an appropriate characteristic value correction amount that takes into account not only the entire original image data but also the face area of the person who is the subject.

  In the third feature configuration, as described in claim 3, in addition to the first feature configuration described above, the face area characteristic value calculating means calculates an average density of each face area for each RGB in each face area. The weighted average density for each RGB weighted by the occupied area is calculated as a face area average characteristic value, and the face area characteristic value correction amount calculating means calculates the hue in the hue saturation coordinate system obtained by converting the face area average characteristic value. The hue correction amount is calculated based on a deviation from the target hue in the hue saturation coordinate system corresponding to the target characteristic value, and the color correction amount for each RGB obtained by inversely converting the calculated hue correction amount is used as the face correction value. This is to calculate the area characteristic value correction amount.

  For example, when a target value is set for each RGB value, the target value is uniquely determined and the correction amount is determined regardless of the saturation, but according to the above configuration, the average density for each RGB is set to hue saturation. Conversion process, the optimum hue correction amount according to the saturation of the original image is apparent in the relationship between the hue that has been converted and placed on the hue saturation coordinate system and the preset target hue. Therefore, the hue correction amount can be calculated accurately and simply. Therefore, the color correction amount obtained by inversely converting the hue correction amount is also calculated as a face region characteristic value correction amount that can appropriately correct the face region image.

  Further, according to the above-described configuration, when each average characteristic value of the face area is an average density, it is possible to calculate a face area average characteristic value that places importance on a large face area among a plurality of face areas. This is a preferred embodiment of the invention.

  As described in claim 4, the fourth feature configuration is that, in addition to the second feature configuration described above, the fusion rate is set based on an area ratio of the entire face region to the entire original image. is there.

  According to the above configuration, when the area ratio of the entire face area in the entire original image is large, the fusion rate is adjusted so that the face area is emphasized, and when the area ratio of the entire face area in the entire original image is small Since the fusion rate can be adjusted so that areas other than the face area are emphasized, it is possible to calculate an appropriate correction amount according to the importance of the face area in the entire original image.

  In order to achieve the above object, the first characteristic configuration of the image processing method according to the present invention is that a face region is included in the input original image data as described in claim 5 of the claims. A face area characteristic value calculating step for calculating a face area characteristic value based on a deviation between the calculated face area characteristic value and a preset target characteristic value. An image processing method comprising a correction amount calculation step, further comprising a reference characteristic value correction amount calculation step for calculating a reference characteristic value correction amount based on all image characteristic values of the original image data, The value calculating step calculates a weighted average characteristic value obtained by weighting the average characteristic value of each of the plurality of face areas included in the original image data by the occupied area as the face area average characteristic value, and compensates the reference characteristic value. The reference characteristic value correction amount calculated in the amount calculation step and the face region characteristic value correction amount calculated in the face region characteristic value correction amount calculation process are merged at a predetermined fusion rate to calculate an appropriate characteristic value correction amount. And a proper characteristic value correction amount calculating step.

  As described above, according to the present invention, even when a plurality of face areas are detected in an image, the correction is performed by emphasizing the characteristic value of the face area that is likely to be the main subject. An image processing apparatus and an image processing method that can correct an appropriate image with high accuracy can be provided.

  Embodiments of a photographic image processing apparatus incorporating an image processing apparatus according to the present invention will be described below. As shown in FIG. 2, the photographic image processing apparatus 1 includes a photographic printer 2 that performs an exposure process on the photographic paper P based on output image data and develops the exposed photographic paper, and a developed photographic film F. A media driver 32 that reads image data from an image data storage medium M such as a memory card that stores image data taken by a film scanner 31 or a digital still camera that reads an image from the camera, a general-purpose computer as a controller 33, and the like The print data is set and inputted with the print order information for the inputted photographic image as the original image, and includes an operation station 3 for performing various image correction processes, and the print data edited from the original image by the operation station 3 Is output to the photographic printer 2 to produce a desired photographic print. That.

  As shown in FIGS. 2 and 3, the photographic printer 2 has two systems of photographic paper magazines 21 containing roll-shaped photographic paper P, and photographic paper P drawn from the photographic paper magazine 21 with a predetermined print. A sheet cutter 22 that cuts into a size; a back print unit 23 that prints print information such as a frame number on the back of the cut photographic paper P; an exposure unit 24 that exposes the photographic paper P based on the print data; A development processing unit 25 including a plurality of processing tanks 25a, 25b, and 25c filled with processing solutions for developing, bleaching, and fixing the exposed photographic paper P is disposed along the transport path of the photographic paper P. The laterally-feeding conveyor 26 that discharges the photographic paper P that has been dried after the development process, and the sheet-paper (photo print) P that is stacked on the laterally-feeding conveyor 26 is sorted in order units. Configured to include the data 27.

  The exposure unit 24 receives a laser beam bundle of RGB three colors modulated based on the print data in the main scanning direction orthogonal to the conveyance direction with respect to the photographic paper P conveyed in the sub-scanning direction by the conveyance mechanism 28. An exposure head 24a for outputting and exposing is accommodated.

  A transport mechanism 28 including a plurality of roller pairs that transport the photographic printing paper P at a process speed corresponding to the exposure unit 24 and the development processing unit 25 disposed along the transport path is disposed before and after the exposure unit 24. Is provided with a chucker-type transport mechanism 28a capable of transporting photographic paper P in multiple rows.

  The controller 33 provided in the operation station 3 is installed with an application program that operates under the control of a general-purpose operating system and executes various controls of the photographic processing apparatus 1, and a monitor 34 as an operation interface with the operator. A keyboard 35, a mouse 36, and the like are connected.

  The photographic processing process executed by the cooperation of the hardware and software of the controller 33 will be described in functional blocks. As shown in FIG. 4, photographic image data read by the film scanner 31 and the media driver 32 is received. An image input unit 40 that performs predetermined preprocessing and transfers it to a memory 41, which will be described later, and print order information and image editing information are displayed on the screen of the monitor 34, and an operation for inputting necessary data for them. A graphic user interface unit 42 that generates a graphic operation screen for displaying an icon or generates various control commands based on an input operation from the keyboard 35 or mouse 36 to the displayed graphic operation screen, and the image input Photos transferred from the section 40 A memory 41 in which image data and photographic image data after correction processing by an image processing unit 47 (to be described later), correction parameters at that time, and set print order information are stored in a predetermined area, and print order information An order processing unit 43 for generating image data, an image processing unit 47 for performing density correction processing and contrast correction processing for each photographic image data stored in the memory 41, and display commands from the graphic user interface unit 42. A display control unit 46 having a video RAM for displaying the image data and various input / output graphic data developed in the memory 41 on the monitor 34 and the final correction after the various correction processes are completed. Print data for generating print data for outputting an image to the photo printer 2 A generating unit 44, and the like formatter 45 which converts the final corrected image according to customer orders to a file format for writing in a storage medium such as a CD-R.

  The film scanner 31 is configured to operate in two modes: a pre-scan mode that reads an image recorded on the film F at a high speed although it is a low resolution, and a main scan mode that reads a high resolution at a low speed. Various correction processes are performed on the low-resolution image read in the can mode, and the final resolution for the high-resolution image read in the main scan mode based on the correction parameters stored in the memory 41 at that time. Correction processing is executed and output to the printer 2.

  Similarly, the image file read from the media driver 32 includes a high-resolution captured image and its thumbnail image, and various correction processes described later are performed on the thumbnail image. Based on the stored correction parameters, a final correction process is performed on the high-resolution captured image. When the image file does not include a thumbnail image, the image input unit 40 generates a thumbnail image from the high-resolution captured image and transfers it to the memory 41. In this way, the calculation load of the controller 33 is reduced by executing various editing processes that are frequently trial and error on low-resolution images.

  As shown in FIG. 1, the image processing unit 47 performs a color correction process on the original image data, and a face region detection unit 10 that detects a face region of a subject included in the input original image data. A color correction processing unit 11, a density correction processing unit 12 for performing density correction processing on the original image data, and a contrast correction for the density correction curve generated for density correction by the density correction processing unit 12 A contrast correction unit 13 that performs image correction, a distortion correction unit 14 that corrects distortion caused by the photographing lens, a sharpening processing unit 15 that enhances an edge of an image and reduces noise, and an image size suitable for the size of a photographic print. An enlargement / reduction processing unit 16 or the like for conversion is provided.

  The face area detector 10 extracts a face area of a person as a subject from the original image data, and a face from the face area included in the original image data detected by the face area extractor 101. A face area characteristic value calculating means 102 for calculating the area average characteristic value is provided.

  For example, the face area extraction unit 101 determines whether or not the outline based on the density edge and the color edge extracted from the original image data is a face area. It is realized using a known technique such as detection based on a pattern recognition technique by evaluating the degree of coincidence with a plurality of element arrangement patterns such as ears.

  However, the face area detected based on the above-described pattern recognition technology is a quadrangular area that combines the face and the peripheral area and includes elements other than the face, so that more accurate image processing is performed. Alternatively, after removing regions other than the skin region from the detected face region, the average characteristic value of the skin region may be obtained as the average characteristic value of the face region.

  For example, for the detected face area, the coordinates of the center pixel of the face area (intersection of square diagonal lines, etc.) and hue, saturation, and density are calculated, and the peripheral pixels are scanned radially from the coordinates of the center pixel. I will do it. Scanning is performed based on whether or not the difference between the hue / saturation / density of the scanned pixel and the hue / saturation / density of the central pixel is within a predetermined value set in advance. The scanning pixel is regarded as constituting a face area, and if the scanning pixel is larger than a predetermined value, the scanned pixel is regarded as not constituting a face area. If there is a pixel that is not considered to constitute a face area, pixels outside the scanned pixel as viewed from the center pixel are regarded as not constituting a face area, and scanning is not performed. That is, in the above method, a series of integrated areas in which the difference between the hue, saturation, and density of the coordinates of the center pixel is within a predetermined value is detected, and the detected area is regarded as a face area.

  It should be noted that the extraction process in the face area extraction unit 101 is not only configured to be automatically performed as described above, but the operator manually operates the original image displayed on the screen of the monitor 34 with the mouse 36 or the like. It may be configured to extract the area designated by the input as a face area.

  The face area characteristic value calculating means 102 is configured to calculate a face area average characteristic value from the face area included in the input original image data. A weighted average characteristic value obtained by weighting each average characteristic value of the face area with each occupied area is calculated as the face area average characteristic value.

In the present embodiment, the face area characteristic value calculating means 102 is a weighted average for each RGB in which the average characteristic value for each RGB of each face area detected by the face area extracting unit 101 is weighted by the occupied area of each face area. The density is calculated as the face area average characteristic value. For example, the face area detected by the face area extraction unit 101 is two face areas A and B, and the size of the face area is S. _A , S _B , the average characteristic value of red (R) of the face area A is R _A , the average characteristic value of green (G) of the face area A is G _A , and the blue (B) of the face area A Is the average characteristic value of B _A , the average characteristic value of red (R) of the face area B is R _B , the average characteristic value of green (G) of the face area B is G _B , and the blue (B in case the average characteristic value of) is B _B, is calculated as the facial area average characteristic value Husband weighted average concentration for each of RGB s _{D R,} D G, When _{D B,} the weighted average density _{D R, D G, D B} are each equation (1), expression (2), such as Formula 3 It is calculated by performing an operation.

  The color correction processing unit 11 calculates a face area characteristic value correction amount based on a deviation between the face area average characteristic value calculated by the face area characteristic value calculating means 102 and a preset target characteristic value. Calculated by a characteristic value correction amount calculation unit 111, a reference characteristic value correction amount calculation unit 112 that calculates a reference characteristic value correction amount based on all image characteristic values of the original image data, and the reference characteristic value correction amount calculation unit 112 The appropriate characteristic value for calculating the appropriate characteristic value correction amount by fusing the reference characteristic value correction amount and the face area characteristic value correction amount calculated by the face area characteristic value correction amount calculating unit 111 with a predetermined fusion rate The correction amount calculation means 113 is provided.

  The face area characteristic value correction amount calculation unit 111 is configured to calculate a face area characteristic value correction amount based on a deviation between the calculated face area average characteristic value and a preset target characteristic value. Specifically, it is calculated by the following procedure. That is, as the first step, the face area average characteristic value is converted into a hue coordinate system, and as the second step, the face area average characteristic value is converted into the hue saturation coordinate system. Obtained by calculating the hue correction amount based on the deviation between the corresponding hue and the target hue corresponding to the target characteristic value in the hue saturation coordinate system, and as a third step, obtained by inversely converting the calculated hue correction amount A color correction amount for each RGB is calculated as the face area characteristic value correction amount.

Hereinafter, each step will be described in detail. In the first step, the face area average characteristic values calculated as weighted average densities D _R , D _G , and D _B for each RGB are converted into a hue saturation coordinate system based on a matrix arithmetic expression shown in [Equation 4]. .

  In the hue / saturation coordinate system, as shown in FIG. 5, the horizontal axis is indicated by P and the vertical axis is indicated by Q. The saturation is indicated by the distance from the origin, and the hue is indicated by the direction from the origin. ing. Specifically, when the origin is an achromatic color and the straight line from the origin in the P coordinate is set to an angle of zero, red is −5π / 6, blue is π / 2, green is −π / 6, yellow Is −π / 2, cyan is π / 6, and magenta is 5π / 6. The face area average characteristic value converted by [Equation 4] occupies one point of the hue / saturation coordinate system as coordinates D (p, q).

  In the second step, the hue in the hue saturation coordinate system of the face area average characteristic value converted in the first step, that is, the hue saturation coordinate corresponding to the coordinate D (p, q) and the target characteristic value. A hue correction amount is calculated based on the deviation from the target hue in the system.

  More specifically, as shown in FIG. 5, a target hue TH as a target characteristic value is set in the hue saturation coordinate system, and the hue saturation of the face area average characteristic value converted in the first step is set. A deviation from the coordinate D (p, q) in the coordinate system, that is, a hue correction amount RH (Δp, Δq) is calculated. The target hue TH is obtained by statistically processing a large number of sample images including a face area photographed under various photographing conditions, and shows an ideal hue as a human skin color, for example. .

  The hue correction amount RH (Δp, Δq) is a distance between the coordinates D (p, q) and the target hue TH, that is, a perpendicular distance drawn from the coordinates D (p, q) to the target hue TH. Indicated.

  In a third step, a color correction amount for each RGB obtained by inversely converting the calculated hue correction amount RH (Δp, Δq) is calculated as the face region characteristic value correction amount. That is, the hue correction amount RH (Δp, Δq) is inversely converted to the face region characteristic correction amount RF (ΔR, ΔG, ΔB).

  The reference characteristic value correction amount calculation unit 112 is configured to calculate a reference characteristic value correction amount based on all image characteristic values of the original image data. For example, based on Evans' theorem that an average outdoor subject becomes close to gray when the colors of the entire negative are mixed, the pixels indicated by the average density for each RGB of the entire original image data are gray. Employing a LATD (Large Area Transmittance Density) method that calculates a color correction amount for each color as a reference characteristic value correction amount, or a modified LATD method that removes high saturation pixels or obtains a conditional average value weighted by saturation. Can do. A reference characteristic value that can appropriately correct the entire original image by calculating the color correction amount for each RGB in which the pixels indicated by the average density for each RGB of the entire original image data are gray as the reference characteristic value correction amount The correction amount can be calculated.

Specifically, if the average density for each RGB of the entire original image data is R _AVR , G _AVR and B _AVR , and the density for each RGB constituting the gray is R _GR , G _GR and B _GR , respectively, reference characteristic value correction amount RS is calculated for each _{RGB, R GR -R AVR, G} GR -G AVR, the B _{GR -B AVR.} Usually, the correction amount of the other component is set as the reference characteristic value correction amount so that any component of R _AVR , G _AVR , and B _AVR has the same value.

  The appropriate characteristic value correction amount calculation unit 113 includes a reference characteristic value correction amount RS calculated by the reference characteristic value correction amount calculation unit 112 and a face region characteristic value correction calculated by the face region characteristic value correction amount calculation unit 111. A proper characteristic value correction amount RJ is calculated by fusing the amount RF with a predetermined fusion rate. In the present embodiment, the appropriate characteristic value correction amount RJ is calculated based on the area ratio of the entire face area with respect to the entire original image. Hereinafter, after calculating the fusion rate, calculation of the appropriate characteristic value correction amount RJ will be described.

The calculation of the fusion rate is, for example, in the original image, the detected face areas are two face areas A and B, and the sizes of the face areas are S _A and S _B , respectively. When the entire size is S _sum , the ratio P occupied by the face area is calculated by performing an operation such as [Equation 5].

Then, the fusion rate r _f is calculated by performing the calculation of [Formula 6] on the ratio P occupied by the face area.

Here, the coefficients C and D are values that can be set as appropriate within the effective range of the present embodiment, and a predetermined value is stored as a result of trial and error. It is a value that is set as appropriate based on populations with different imaging conditions (for example, the type, intensity, direction, exposure amount, etc.) Specific values of the coefficients C and D include, for example, C = 0.05, D = 0... Under the condition that the weight coefficient r _f based on the average density value and the face area size exceeds 1, When 5 is set, the fusion rate r _f increases from 0.5 to 1 as the ratio P occupied by the face region increases from 0 (%) to 100 (%). In this example, the fusion rate r _f is r _f = 1 when the ratio P occupied by the face region is 10% or more. This is a result of the coefficients C and D being derived from statistics that the ratio of the face area to the entire area is 10% or less in most photographic images taken in a normal situation.

The appropriate characteristic value correction amount RJ is obtained, for example, by performing a weighting calculation as shown in [Equation 7] based on the reference characteristic value correction amount RS, the face region characteristic value correction amount RF, and the fusion rate r _f. Calculated.

When the ratio of the face area in the entire original image is large by using the fusion rate r _f , when the reference characteristic value correction amount RS and the face area characteristic value correction amount RF are fused, It is possible to place a weight on the face area characteristic value correction amount RF.

  Then, color correction is performed based on the appropriate characteristic correction amount RJ obtained by the appropriate characteristic value correction amount calculating means 113.

  In the above-described embodiment, when a face area is not detected, color correction is performed using the reference characteristic value correction amount RS as the appropriate characteristic value correction amount RJ.

  The image processing apparatus according to the present invention described above calculates, as the face area average characteristic value, the weighted average characteristic value obtained by weighting the average characteristic value of each of the plurality of face areas included in the input original image data by the occupied area. A face area characteristic value calculating means, a face area characteristic value correction amount calculating means for calculating a face area characteristic value correction amount based on a deviation between the calculated face area average characteristic value and a preset target characteristic value; Reference characteristic value correction amount calculating means for calculating a reference characteristic value correction amount based on all image characteristic values of the original image data, the reference characteristic value correction amount calculated by the reference characteristic value correction amount calculating means, and the face area The computer is caused to function as an appropriate characteristic value correction amount calculating means for calculating an appropriate characteristic value correction amount by fusing the face area characteristic value correction amount calculated by the characteristic value correction amount calculating means with a predetermined fusion rate. Image processing program can be easily realized by being installed.

  Hereinafter, another embodiment will be described. In the above-described embodiment, the calculation of the hue correction amount RH (Δp, Δq) in the process in the face region characteristic value correction amount calculation unit 111 is based on the coordinate D (p) obtained by converting the face region average characteristic value into the hue saturation coordinate system. , Q) and the distance between the hue saturation coordinate system and the target hue TH, that is, the configuration indicated by the distance of the perpendicular drawn from the coordinates D (p, q) to the target hue TH. The quantity RH (Δp, Δq) may not be a perpendicular distance.

For example, when the hue correction amount RH (Δp, Δq) is a distance of a perpendicular line drawn from the coordinates D (p, q) to the target hue TH, the red component is more than that normally required for human skin color. It tends to decrease. In order to reduce the decrease of the red component as described above, as shown in FIG. 6, the value Delta] p · r _p multiplied by predetermined weighting calculation value r _p in Delta] p the hue correction amount RH (Δp, Δq) In other words, by setting the hue correction amount RH to RH _r (Δp · r _p , Δq), the red component is increased as indicated by a double arrow. The weighting calculation value r _p, for example, as shown in FIG. 7, the saturation of the coordinate D (p, q), that is, determined by the distance from the origin.

Incidentally, the weighting calculation value of the above, the value [Delta] q · r _q obtained by multiplying a predetermined weighting operation value r _q in [Delta] q hue correction amount RH (Δp, Δq) may be a Q component of, also, the Delta] p and [Delta] q wherein both the weighting calculation value r _p and r _q respectively multiplied values hue correction amount RH (Δp, Δq) may be P and Q components of the.

Further, instead of the weighting calculation value determined in accordance with saturation, as shown in FIG. 8, the coordinates D (p, q) correction shift amount is determined according to the hue of s _p and s _q added The values Δp + s _p and Δq + s _q are used as the P and Q components of the hue correction amount RH (Δp, Δq), that is, the hue correction amount RH is set to RH _s (Δp + s _p , Δq + s _q ). There is also a configuration that increases. The horizontal axis (direction from the origin) of the graph shown in FIG. 8 indicates the angle in radians when the straight line from the origin of the P coordinate in the hue saturation coordinate system is zero.

Furthermore, the weighting operation value _r p, using both _{r q} and the correction shift amount _s p, _{s q,} the hue correction amount RH to _{RH rs (Δp · r p +} s p, Δq · r q + s q) And a configuration that increases the red component.

  In the above-described embodiment, the reference characteristic value correction amount calculating unit 112 calculates the color correction amount for each RGB in which the pixels indicated by the average density for each RGB of the entire original image data are gray as the reference characteristic value correction amount. Although the method has been described as adopting the method, the reference characteristic value correction amount calculation unit 112 may use other methods as long as the reference characteristic value correction amount is calculated based on all image characteristic values of the original image data. You may employ | adopt.

  For example, LATD calculation means for obtaining an average pixel value for each color component with respect to all constituent pixels for each inputted frame image, and for each pixel of the frame image, saturation is defined by a distance from the center and an angle around the center. Coordinate conversion means for converting into pixels in a hue saturation coordinate system in which hue is defined, and a plurality of concentric circular regions centering on the average pixel value in the hue saturation coordinate system are defined, and each circular region is defined in each circle region A circular area characteristic value calculating means for obtaining a circular area characteristic value indicating a distribution characteristic of the included pixels for each area; and a correction amount calculating means for obtaining a color correction amount based on the circular area characteristic value. There may be.

  Here, the circular area characteristic value includes the number of pixels included in each circular area and the difference between the average pixel value and the average pixel value that is an average value for each color component of the pixels included in each circular area. The correction amount calculation means can be constituted by a neural network learned by an error back propagation learning method or multiple regression analysis means.

  In the above-described embodiment, the color correction processing unit 11 or the density correction processing unit 12 determines the reference characteristic value correction amount calculated based on the entire image characteristic value and the face area calculated based on the image characteristic value of the face area. The configuration for calculating the appropriate characteristic value correction amount by performing the fusion processing with the characteristic value correction amount has been described. However, the color correction processing unit 11 or the density correction processing unit 12 performs the face area characteristic value correction amount calculation unit 111 in the face area. The configuration may be such that the face area characteristic value correction amount calculated based on the image characteristic value of the area is applied only to the face area without merging.

  In other words, the face area characteristic value correction amount calculation unit 111 may be configured to be used independently as a correction amount calculation unit dedicated to a face area. With this configuration, the face area characteristic value correction amount is used for the face area detected from the original image data, and the reference characteristic value correction amount is used for the area other than the face area. It becomes possible. In addition, it is possible to adopt a configuration suitable for processing an image of only a face area extracted or cut out in advance in another apparatus.

  In the above-described embodiment, the configuration in which various characteristic values and correction amounts are calculated in the RGB color system has been described. However, the present invention is not limited to the RGB color system, and is calculated in another color system. May be. For example, in the case of color correction, it is possible to adopt a CYM color system or the like, and it is also possible to adopt a YCC color system in density correction processing described later.

  In the above-described embodiment, the mode in which the appropriate characteristic value correction amount RJ for use in color correction is calculated in the color correction processing unit 11 and used in color correction has been described. However, the appropriate characteristic value correction amount RJ is used in the density correction processing unit. 12 may be calculated and used for density correction. In the description of the present embodiment, the elements provided in the density correction processing unit 12, that is, the face area characteristic value correction amount calculating means, the reference characteristic value correction amount calculating means, and the appropriate characteristic value correction amount calculating means are described above. The elements and numbers in the embodiment will be described.

  In the present embodiment, for example, as shown in FIG. 9, the face area characteristic value calculation unit 102 replaces the average density value for each RGB of each face area with an RGB average density (for each constituent pixel of each face area). The weighted average density obtained by weighting the pixel average value of (R + G + B) / 3) by the area occupied by each face area is calculated as the face area average characteristic value, and the density correction processing unit 12 is configured to calculate the face area characteristic. A value correction amount calculation unit 121, a reference characteristic value correction amount calculation unit 122, and an appropriate characteristic value correction amount calculation unit 123 are provided.

  The reference characteristic value correction amount calculation unit 122 generates a density histogram in the original image data, and corrects the density correction value Gamma of the original image data according to a function shown in [Formula 8] stored in advance. It is configured to calculate as a quantity.

  Here, as shown in FIG. 10, Maxg is the maximum density value of the original image data, that is, the density value of the brightest pixel, and Ming is the density value of the original image data. The minimum value, that is, the darkest pixel density value, and the Mass is a density value that is an average density of the original image data. The Maxk is the maximum value of the density setting value in the original image data. For example, when the density value is set between 0 gradation and 255 gradation in 8 bits, Maxk is set to 255. The Mink is the minimum density setting value in the original image data. For example, the Mink is set to 0 when the density value is set between 0 gradation and 255 gradation in 8 bits.

The face area characteristic value correction amount calculating means 121 is a face-dependent density correction value Gamma (F) based on a face-dependent density correction value model expression using the face area average characteristic value D _Favr obtained by the face area detection unit 10 as a variable. ) As a face area characteristic value correction amount. The model formula visually evaluated a correction image based on a density correction curve based on a density correction value obtained by changing the reference density correction value for a large number of sample images obtained by shooting various subjects under various shooting conditions. As a result, the correction function F1 is obtained by statistically processing the distribution of density correction values determined to have corrected the average density of the face area and the face area at that time to an appropriate density.

FIG. 11 shows an example of the correction function F1, and the face-dependent density correction value Gamma (F) is obtained by Gamma (F) = F1 (D _Favr ) with respect to the face area average characteristic value D _Favr . It is what Here, the average density of the face region is shown as a value normalized with an 8-bit image.

The appropriate characteristic value correction amount calculation unit 123 includes a density correction value Gamma calculated by the reference characteristic value correction amount calculation unit 122 and a face-dependent density correction value Gamma (calculated by the face area characteristic value correction amount calculation unit 121. F) and the appropriate density correction value Gamma (R) calculated by performing the fusion process at a predetermined fusion rate is calculated as the appropriate characteristic value correction amount. It is configured to use the fusion rate r _f calculated by the same method as that calculated in the form, that is, the method of performing calculations such as [Equation 5] and [Equation 6].

Regarding the calculation of the appropriate density correction value Gamma (R), specifically, the appropriate characteristic value correction amount calculation means 123 includes the reference density correction value Gamma, the face-dependent density correction value Gamma (F), and the fusion. Based on the rate r _f , the density correction value Gamma is corrected to the appropriate density correction value Gamma (R) by performing a weighting calculation as shown in [Equation 9].

  Further, the appropriate characteristic value correction amount calculation means 123 calculates the input pixel density value Nin (i) based on the appropriate density correction value Gamma (R) according to the function shown in [Formula 10] stored in advance. A density correction curve embodied as a look-up table for converting the density value into an output pixel density value Nout (i) corresponding to the density correction value Gamma is generated (i = 0, 1, 2,..., 255 (8 Bit data)).

  Here, the constant L is a target gradation setting width after density correction processing of the original image data. For example, when the output image data is set by 8-bit data, the target gradation setting width is 255. The output pixel density value Nout (i) is output as a value from 0 to 255.

  That is, as shown in FIG. 12, the appropriate characteristic value correction amount calculation unit 123 sets the density so that the original image data becomes bright when the appropriate density correction value Gamma (R) is set to a value larger than 1. A density correction curve for correction processing is generated, and when the appropriate density correction value Gamma (R) is set to a value smaller than 1, a density correction curve for density correction processing is generated so that the original image data becomes dark. It is configured.

In the above description, the case where the face area detection unit 10 determines that a human face area is detected in the subject has been described. However, it is determined that no human face area is detected in the subject. In this case, the appropriate characteristic value correction amount calculation unit 123 is configured to generate a density correction curve based on the density correction value Gamma instead of the appropriate density correction value Gamma (R). Alternatively, the fusion rate r _f = D in which the ratio P occupied by the face area in [Equation 6] is P = 0 may be applied.

  As described above, the original image is corrected by the density correction curve generated by the appropriate characteristic value correction amount calculation unit 123 based on the appropriate density correction value Gamma (R). Is corrected to show a preferable gradation.

  Then, the contrast correction processing unit 13 performs contrast correction on the density correction curve generated by the appropriate characteristic value correction amount calculation unit 123.

  Note that the above-described embodiment is merely an example of the present invention, and it is needless to say that the specific configuration and the like of each block can be changed and designed as appropriate within the scope of the effects of the present invention. Moreover, you may combine the above-mentioned several embodiment suitably as needed.

Explanatory drawing of functional block configuration in image processing unit Explanatory drawing of the external configuration of the photographic image processing apparatus Illustration of photo printer Explanatory drawing of functional block configuration of photographic image processing apparatus Illustration of hue saturation coordinate system Explanatory drawing when weighting calculation is performed on the hue correction amount in the hue saturation coordinate system Explanatory drawing of the graph showing the relationship between saturation and weighted value Explanatory drawing of the graph showing the relationship between hue and correction shift amount Explanatory drawing of a functional block configuration in an image processing unit in which image processing of the present invention is applied to density correction Explanatory diagram of density histogram related to density correction value calculation Explanatory drawing of the graph which is an example of the model formula for deriving the face dependent density correction value Explanatory drawing of density value transition in density correction processing

Explanation of symbols

1: image processing apparatus 10: face area detection unit 101: face area extraction unit 102: face area characteristic value calculation unit 11: color correction processing unit 111: face area characteristic value correction amount calculation unit 112: reference characteristic value correction amount calculation unit 113: Appropriate characteristic value correction amount calculation means 12: Density correction processing unit 47: Image processing unit

Claims (5)

Based on a deviation between the calculated face area average characteristic value and a preset target characteristic value, a face area characteristic value calculating means for calculating the face area average characteristic value from the face area included in the input original image data An image processing apparatus comprising a face area characteristic value correction amount calculating means for calculating a face area characteristic value correction amount,
The face area characteristic value calculation means calculates an average weighted characteristic value obtained by weighting an average characteristic value of each of a plurality of face areas included in the original image data with a respective occupied area as a face area average characteristic value. .   Reference characteristic value correction amount calculating means for calculating a reference characteristic value correction amount based on all image characteristic values of the original image data, the reference characteristic value correction amount calculated by the reference characteristic value correction amount calculating means, and the face area 2. An appropriate characteristic value correction amount calculating means for calculating an appropriate characteristic value correction amount by subjecting the face area characteristic value correction amount calculated by the characteristic value correction amount calculating means to fusion processing at a predetermined fusion rate. Image processing apparatus.   The face area characteristic value calculation means calculates a weighted average density for each RGB obtained by weighting an average density for each RGB of each face area by an occupied area of each face area, and corrects the face area characteristic value The amount calculation means calculates a hue correction amount based on a deviation between a hue in the hue saturation coordinate system obtained by converting the face area average characteristic value and a target hue in the hue saturation coordinate system corresponding to the target characteristic value. The image processing apparatus according to claim 1, wherein a color correction amount for each RGB obtained by inversely converting the calculated hue correction amount is calculated as the face region characteristic value correction amount.   The image processing apparatus according to claim 2, wherein the fusion rate is set based on an area ratio of the entire face area to the entire original image. A face area characteristic value calculation step for calculating a face area average characteristic value from a face area included in the input original image data, and based on a deviation between the calculated face area average characteristic value and a preset target characteristic value An image processing method comprising a face area characteristic value correction amount calculating step for calculating a face area characteristic value correction amount,
The method further comprises a reference characteristic value correction amount calculating step for calculating a reference characteristic value correction amount based on all image characteristic values of the original image data, wherein the face area characteristic value calculating step includes a plurality of faces included in the original image data. A weighted average characteristic value obtained by weighting an average characteristic value of each area with each occupied area is calculated as a face area average characteristic value, and the reference characteristic value correction amount calculated by the reference characteristic value correction amount calculating step and the face area An image processing method comprising an appropriate characteristic value correction amount calculating step of calculating an appropriate characteristic value correction amount by performing a fusion process with the face region characteristic value correction amount calculated in the characteristic value correction amount calculating step at a predetermined fusion rate.