JP2008166963A - Image density correction method and image processing unit executing its method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a technology which automatically discriminates and extracts a photographic image whose object is a bride with a bridal costume from an input photographic image, and corrects a color tone of the bridal costume in proper color. <P>SOLUTION: The image density correction method selects one image which has bridal photographic image characteristics out of input photographic images as a bridal photographic image, and carries out density correction of the bridal costume of the photographed image determined as a bridal costume photographic image based on a discriminated result whether the bridal costume exists or not so as to be a proper color based on face detection information and color density information in the bridal photographic image. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an image density correction technique for a bridal photographed image obtained by photographing a bridal scene, particularly a bridal costume photographed image.

  At present, in the photographic printing industry, it has been obtained by digitizing a photographed image data obtained by digitizing a photographed image formed on a photographic film using a film scanner or a digital photographing device such as a digital camera. The captured image data (hereinafter simply abbreviated as the captured image) is subjected to image processing such as density correction and color correction, and then converted into print data, and the print exposure unit is driven based on the print data to capture the captured image. The mainstream is digital photographic processing technology for printing photographic materials on photographic materials (photographic paper) to produce photographic prints.

  In the field of such digital photo processing technology, in order to produce an appropriate photo print, for each color component (for example, R: red, G: green, B: blue) of the pixels constituting the input photographed image. Processing for correcting the density value is performed. As such a density correction method, density correction by a density correction curve (gamma correction) is generally used. In other words, since the density gradation that the photographic paper develops does not match the density gradation of the input photographed image, the density gradation that the photographic paper develops is set to be suitable for human visual characteristics. The input image data is corrected using the correction curve, and the density gradation that is colored on the photographic print (printing paper) finally obtained based on the corrected output image data corresponds to the human visual characteristics. I try to be something. As the density correction curve, a basic density correction curve created based on theoretical and empirical knowledge is well known and used. However, the subject situation and the shooting environment situation of the shot image are various, and the simple basic density correction curve cannot obtain appropriate output image data depending on the shooting scene of the input shot image data. Has occurred. For example, when correcting photographic image data captured from photographic film taken with over / under exposure (overexposure / underexposure), the brightness of each pixel constituting the photographic image data is low (shadow) or high. The output density is very weak with respect to the input density because the slope of the shadow part and highlight part in the basic density correction curve is moderate, while the brightness (highlight part) is too biased. Will be corrected.

  Therefore, in order to perform an appropriate density correction for a human photographed image including a face area, an area that is likely to include a feature amount of RGB data of a pixel having a hue corresponding to the skin color of the person and a person. There has been known a technique devised so that the density of a person is appropriately corrected by correcting the density correction curve based on the feature value of the RGB data of each pixel (see Patent Document 1).

  In addition, judgment using pattern matching, neural network, etc. in order to create appropriate photo prints from captured images where the main subject such as a person is crushed black or white A determination process for determining whether a main subject such as a person's face or person extracted by the method has backlight or high contrast, and a shadow of scene reference image data when the determination process determines backlight or high contrast. There is known an image processing method including a specifying step for specifying a side saturation point and a highlight side saturation point, and a gradation correction step for changing the brightness of the main subject (see, for example, Patent Document 2). At that time, the average values of the luminance (density values) at which the cumulative values on the shadow side and highlight side of the cumulative luminance histogram are 0.1 to 0.4% of the total are the shadow side saturation point and the highlight side saturation, respectively. After being specified as a point, gradation correction (density correction) is performed.

  In such a density correction technique, when it is clear that the captured image to be processed is a backlight scene or a high contrast scene, an effect as expected can be obtained. However, there are cases in which captured images are called composite imaging conditions in which various imaging conditions (backlight scenes, high contrast scenes, tungsten scenes, etc.) overlap, and in such cases, good results cannot be obtained. In particular, as a photographed image in which such a composite photographing scene is likely to occur, there is a bridal photographed image in which a bride and groom at a wedding ceremony (especially a bride wearing a pure white wedding dress or a white strip) is a subject. In such a bridal photographed image, first, a photo print in which the pure white costume looks pure white is expected regardless of the photographing conditions, whereas the above-described conventional density correction technique has not provided sufficient results. .

  In addition, when the captured image to be processed has a very extreme density distribution or color distribution such as a backlight scene or a high contrast scene, a histogram analysis of the captured image or a pixel value evaluation for each divided section of the captured image, etc. A technique for automatically detecting such a scene is also proposed (for example, see Patent Document 3). However, with this technique, it is impossible to automatically perform image active correction that focuses on a specific region, for example, on a bridal gown and makes its color appropriate.

JP 2002-369020 A (paragraph number 0004-0008, FIG. 5) JP 2005-209012 (paragraph number 0006-0015, FIG. 4) Japanese Patent Laying-Open No. 2006-324987 (paragraph number 0016, FIG. 1)

  In view of the above situation, an object of the present invention is to provide a bride in a bridal photographed image obtained by photographing a bridal scene that tends to be a complex photographing adverse condition such as a backlight photographing condition, a tungsten illumination photographing condition, and a strobe photographing condition. To respond to the desire to easily output a photo print that appropriately represents the costume, the bride automatically wears the bride's costume from the input photographed image, and automatically determines and extracts the photographed image. It is to provide a technique for correcting the color of a costume to an appropriate color.

  In order to solve the above-described problem, in one image density correction method according to the present invention, an input photographed image having a bride photographed image characteristic is selected as a bride photographed image, and face detection information and color are detected in the bride photographed image. It is determined whether or not the bridal gown exists based on the density information, and the density is such that the bridal gown photographed image determined based on the determination result and the bridal gown of the photographed image have an appropriate color (generally white). Make corrections.

  In order to solve the above-mentioned problem, in another image density correction method according to the present invention, an input photographed image having a bride photographed image characteristic is selected as a bride photographed image, and face detection information in the bride photographed image is selected. And determining whether or not a bridal gown exists based on the color density information, specifying a bridal gown area to be corrected in the bridal photographed image determined as a bridal gown photographed image based on the determination result, and the bride Density correction is performed so that the bridal gown in the costume photographed image has an appropriate color (generally white).

  In the bridal photographed image, the photographed image that may be the subject of the bride is the most important. Therefore, in the method of the present invention, first, the bride photographed image is selected based on the preset bride photographed image characteristics. To do. The bride photographed image characteristic can be obtained by statistically analyzing a large number of bridal photographed images including a bridal photographed image in which the bride is a subject. It is determined whether or not a bridal gown exists based on the face position and face size obtained from the face detection information of the selected bride photographed image, the color density information of the area where the person is wearing, and the like. If the photographed image of the bride is regarded as a photographed image of the bridal gown in this determination, the photographed image is subjected to density correction so that the bridal gown becomes an appropriate color.

  When it is desired to intensively correct the density of the bridal gown area, the bridal gown area to be intensively corrected is specified, and density correction is performed so that the bridal gown in the bridal gown photographed image has an appropriate color. This makes it easy to create a photo print that properly represents the bridal gown in a bridal image obtained by photographing a bridal scene that tends to be a complex shooting adverse condition such as tungsten illumination shooting conditions and strobe shooting conditions. Can output automatically.

  As specific bridal shot image characteristics used to estimate the bridal shot image that the bride is the subject, the frequency pattern of the shadow area and highlight area in the histogram of the shot image and the pixel value of the divided section unit of the shot image Distribution. For example, considering that bridal gowns are usually pure white, the other dresses are black, and there are many indoor shots, the histogram of the bridal shot image where the bride is the subject has shadow and highlight areas. Each frequency ratio and the ratio to the overall frequency are quite characteristic and can be used. In addition, since the standing positions of the bride and the groom have regularity, it is also effective to divide the photographed image and compare the luminance, hue, and saturation in the divided section units.

  Furthermore, by matching the face area obtained through the face detection algorithm with a preset bride's face, it is possible to identify the bride in the bridal image, and as a result, a photo print that expresses the bridal gown more appropriately Can contribute to output. Further, the use of a bridal gown reference shape set in advance when determining whether or not a bridal gown exists from a bridal photographed image can improve the accuracy of costume region detection.

  When correcting the bridal gown to the appropriate color, if the shooting site is dominated by a certain color and the bride's face and costume are also under that color, forcing the gown only to the proper color will Since the color balance, particularly the color balance with the bride's face, may be deteriorated, the appropriate color set may be adjusted by the hue of the face area obtained based on the face detection information.

  One of the image processing units that perform the above-described image density correction method includes a bride shooting scene determination unit that selects, as a bride shot image, an image having a bride shot image characteristic among input shot images; A bridal gown determination unit that determines whether a bridal gown exists based on the face detection information and the color density information, and the bridal gown of the photographed image determined as the bridal gown photographed image based on the determination result is the appropriate color. A density correction module.

  Another one of the image processing units that implement the above-described image density correction method includes a bride photographing scene determination unit that selects, as a bride photographed image, a bride photographed image characteristic among the input photographed images, and the bride photographing A bridal gown determination unit that determines whether a bridal gown exists based on face detection information and color density information in the image, and corrects the bridal gown image determined as a bridal gown photographed image based on the determination result. A bridal gown area specifying unit for specifying a bridal gown area and a density correction module so that the bridal gown of the bridal gown photographed image has an appropriate color.

  Any of the image processing units is accompanied by the operational effects described in the image density correction method according to the present invention, and can be provided with various additional characteristic configurations described above.

  The present invention also includes an image density correction program for causing a computer to execute the image density correction method. Other features and advantages of the present invention will become apparent from the following description of embodiments using the drawings.

  An external view of a photographic printing apparatus incorporating an image density correction technique according to the present invention is shown in FIG. This photographic printing apparatus is a printing station 1B as a photographic printer that performs an exposure process and a development process on photographic paper P, and a photograph taken from an image input medium such as a developed photographic film 2a or a memory card 2b for a digital camera. Images (here, unless otherwise specifically required, the captured image data as digital data, and the captured image for printing and the captured image for monitor display based on the captured image data are simply referred to as a captured image) And an operation station 1A for generating / transferring print data used in the print station 1B.

  This photo printing apparatus is also called a digital minilab. As can be understood from FIG. 2, the printing station 1B pulls out the roll-shaped printing paper P stored in the two printing paper magazines 11 and prints it with the sheet cutter 12. The back print unit 13 prints print processing information such as color correction information and frame number on the back side of the photographic paper P, and the print exposure unit 14 cuts the print paper P into the size. A photographed image is exposed on the surface of the photographic paper P, and the exposed photographic paper P is sent to a processing tank unit 15 having a plurality of development processing tanks for development processing. After drying, the photographic paper P, that is, the photographic prints P, sent to the sorter 17 from the transverse feed conveyor 16 at the upper part of the apparatus is collected in a plurality of trays of the sorter 17 in a state of being sorted in order units (see FIG. 1). ).

  A photographic paper transport mechanism 18 is laid to transport the photographic paper P at a transport speed in accordance with various processes for the photographic paper P described above. The photographic paper transport mechanism 18 is composed of a plurality of nipping and transporting roller pairs including a chucker type photographic paper transport unit 18a disposed before and after the print exposure unit 14 in the photographic paper transport direction.

  The print exposure unit 14 applies R (red), G (green), and B (blue) to the printing paper P conveyed in the sub-scanning direction based on print data from the operation station 1A along the main scanning direction. A line exposure head for irradiating laser beams of the three primary colors (1) is provided. The processing tank unit 15 includes a color developing tank 15a for storing a color developing processing liquid, a bleach-fixing tank 15b for storing a bleach-fixing processing liquid, and a stabilizing tank 15c for storing a stable processing liquid.

  A film scanner 20 for obtaining photographed image data from photographed image frames of the photographic film 2a is arranged at the upper position of the desk-like console of the operation station 1A, and is used as a photographed image recording medium 2b mounted on a digital camera or the like. A media reader 21 that acquires captured images as image data from various semiconductor memories and CD-Rs that are used is incorporated in a general-purpose personal computer that functions as the controller 3 of the photographic printing apparatus. The general-purpose personal computer is also connected with a monitor 23 for displaying various types of information, and a keyboard 24 and a mouse 25 as pointing devices used as operation input units used for various settings and adjustments.

  The controller 3 of this photographic printing apparatus uses a CPU as a core member and constructs a functional unit for performing various operations of the photographic printing apparatus by hardware and / or software, as shown in FIG. As shown in the figure, an image input unit 31 that captures a captured image read by the scanner 20 or the media reader 21 and performs preprocessing necessary for the next processing, and a graphic operation screen including various windows and various operation buttons. From the GUI unit 33 and the GUI unit 33 that constructs a graphic user interface (hereinafter abbreviated as GUI) that generates a control command from a user operation input (via the keyboard 24, the mouse 25, etc.) through such a graphic operation screen Input from control command sent or keyboard 24 etc. An image management unit 32 that performs density correction, color correction, photo retouching processing, and the like on the captured image transferred from the image input unit 31 to the memory 30 in order to generate desired print data based on the received operation command; A video control unit 35 that generates a video signal for causing the monitor 23 to display graphic data sent from the GUI unit 33, or a simulated image as a print source image, an expected finished print image, or the like during pre-judge printing work such as correction. A print data generation unit 36 that generates print data suitable for the print exposure unit 14 installed in the print station 1B based on the processed image data for which image processing has been completed, and an original shooting according to a customer's request Images and captured images that have undergone image processing are written as image data on a CD-R. Etc. formatter 37 for formatting into a form for writing and the like.

  The image management unit 32 outputs a face detection module 40 that outputs face detection information including the position and size of the face area detected from the captured image expanded in the memory 30 using a face detection algorithm, and the captured image expanded in the memory 30. A density correction module 80 that performs density correction on the image and an image processing module 90 that performs image processing such as color correction and filtering (such as blurring and sharpness) on the density-corrected captured image are provided. Further, in order to properly generate the density correction curve used by the density correction module 80 in the density correction processing based on the type of the captured image to be corrected, the bridal for generating the density correction curve for the bridal captured image is used. A density correction curve generation module 60 and a normal density correction curve generation module 70 that generates a density correction curve for a general captured image using a known algorithm are provided.

The bridal density correction curve generation module 60 generates a density correction curve so that the bridal gown is appropriate white for a bridal photographed image in which the bride is the subject. When the area of the pure white costume can be clearly determined, it also has a function of creating a correction curve for making the costume area an appropriate white color with priority. For example, a bridal density correction curve is generated from the entire photographed image, and the final bridal density correction curve is created by fusing this bridal density correction curve with a correction curve that limits this area to the appropriate white color. Good. Accordingly, a bride scene detection module 50 is provided to select a bride scene whose main subject is the bride costume from the input photographed image. The bride scene detection module 50 includes information on the identified bride costume area. Is also provided to the bridal density correction curve generation module 60.

  The face detection module 40 can use a general-purpose one. Here, based on a face detection algorithm, an area considered as a face in a photographed image is detected, and the face position and size (based on the face position) are detected. That output face detection information (such as the vertical and horizontal sizes of the rectangular pixel area).

  As shown in FIG. 4, the bride scene detection module 50 generates a histogram of the input captured image and calculates the histogram characteristic, and divides the input captured image into meshes. Then, a divided image feature amount calculation unit 52 that calculates a divided image feature amount from the image characteristic distribution of the divided section, and the histogram characteristic and the divided image feature amount satisfy the determination condition defined as the bride photographed image characteristic. Based on the determination results of the bride photographic scene determination unit 53 that selects a photographic image as a bride photographic image, the bridal gown determination unit 54 that determines whether a bridal gown exists in the bride photographic image, and the bridal gown determination unit 54. A bridal gown area specifying unit 56 for specifying a bridal gown area to be corrected in the determined bridal gown photographed image. It is provided. In this embodiment, the bridal gown determination unit 54 uses the face area included in the face detection information obtained from the face detection module 40 as a reference position, estimates the costume area in consideration of the shape of the bridal gown, and the estimation. Since a bridal gown photographed image showing a pure white bridal gown is determined from the color density information of the area, a bridal gown reference table 55 storing a typical bridal gown shape and color density pattern is attached.

  When the bridal gown area is specified by the bridal gown area specifying unit 56, a bridal density correction curve is generated by the bridal density correction curve generating module 60 in order to correct the bridal density of the photographed image. The bridal density correction curve generation module 60 of this embodiment has a function of correcting a bridal irrespective of the specified bridal gown area and a function of correcting the bridal in consideration of the specified bridal gown area. In any case, the bridal density correction curve generation module 60 generates a bridal density correction curve so that the bridal gown area is appropriate white.

  A basic algorithm for generating a bridal density correction curve so that the bridal gown area becomes appropriate white color by the bridal density correction curve generation module 60 will be described with reference to the schematic diagram of FIG.

  First, an R density histogram (# 01r), a G density histogram (# 01g), and a B density histogram (# 01b) are generated using the density values of all pixels of the captured image to be processed. The cumulative number of pixels from the highlight side of this R density histogram (frequency number): the total number of pixels of Nh # r: the density value where the ratio (%) to Z is the red component first predetermined value: σ1 # r, This density value is set as a red component highlight side saturated density value: Hr (# 02r). The red component first predetermined value: σ1 # r is selected from 1 to 20%. Similarly, the green component highlight side saturation density value: Hg is obtained using the G density histogram and the green component first predetermined value: σ1 # g (# 02g), and the B density histogram and the blue component first predetermined value: σ1. The blue component highlight side saturation density value: H # b is obtained using #b (# 02b). The red component first predetermined value: σ1 # r, the green component first predetermined value: σ1 # g, and the blue component first predetermined value: σ1 # b may be the same or different. Next, the accumulated pixel number (frequency number) from the shadow side of the R density histogram: the total pixel number of Ns # r: the density value at which the ratio (%) to Z is the red component second predetermined value: σ2 # r is obtained. The density value is set as a red component shadow side saturation density value: Sr (# 03r). The red component second predetermined value: σ2 # r is selected in the range of 0.01 to 1% and smaller than the red component first predetermined value: σ1 # r. Further, the number of accumulated pixels (frequency number) from the shadow side of the G density histogram: the total number of pixels of Ns # g: a density value at which the ratio (%) to Z is the green component second predetermined value: σ2 # g is obtained, This density value is set as the green component shadow side saturation density value: Sg (# 03 g). The green component second predetermined value: σ2 # g is selected in the range of 0.01 to 1% and smaller than the green component first predetermined value: σ1 # g. The green component second predetermined value: σ2 # g is selected in the range of 0.01 to 1% and smaller than the green component first predetermined value: σ1 # g. Further, the number of accumulated pixels (frequency number) from the shadow side of the B density histogram: the total number of pixels of Ns # b: a density value at which the ratio (%) to Z is the blue component second predetermined value: σ2 # b is obtained. This density value is set as the blue component shadow side saturation density value: Sb (# 03b). The blue component second predetermined value: σ2 # b is selected in the range of 0.01 to 1% and smaller than the blue component first predetermined value: σ1 # b.

  A red component density correction curve is generated using the red component highlight side saturation density value Hr and the red component shadow side saturation density value Sr thus obtained (# 04r), and the green component highlight side saturation density is obtained. Value: Hg and green component shadow side saturation density value: Sg is used to generate a green component density correction curve (# 04g), blue component highlight side saturation density value: Hb and blue component shadow side saturation density value: Sb The blue component density correction curve is generated by using (# 04b). In each of these density correction curves, the position of the shadow-side saturation density value is the shadow-side saturation point, so any input density value smaller than this becomes the minimum output density value, and the highlight-side saturation density value: H position is highlighted. Since this becomes the side saturation point, all input density values larger than this become the maximum output density value.

  As shown in FIG. 6, the bridal density correction curve generation module 60 that realizes the above-described function includes a density histogram generation unit 61, a highlight side saturation density value calculation unit 62, a shadow side saturation density value calculation unit 63, and a bridal density correction. A curve generation unit (hereinafter simply referred to as a density correction curve generation unit) 64 and a basic density correction curve generation unit 65 are provided.

  The density histogram generation unit 61 obtains a density histogram for each color component (R, G, B) from the bridal photographed image developed in the memory 30, and the density value (pixel value, 0 to 255 in the case of 8-bit color). The number of pixels (frequency number) is obtained and temporarily stored.

  The highlight side saturated density value calculation unit 62 has a density value at which the ratio of the cumulative number of pixels from the highlight side of the density histogram for each color component obtained by the density histogram generation unit 61 to the total number of pixels is a first predetermined value. Is calculated as the highlight-side saturation density value. The first predetermined value: σ1 is set for each color component (red component first predetermined value: σ1 # r, green component first predetermined value: σ1 # g, blue component first predetermined value: σ1 # b). , Highlight side saturation density value: H is also calculated for each color component (red component highlight side saturation density value: Hr, green component highlight side saturation density value: Hg, green component highlight side saturation density value: Hb). The Similarly, the shadow side saturated density value calculation unit 63 has a ratio of the cumulative number of pixels from the shadow side of the density histogram for each color component obtained by the density histogram generation unit 61 to the total number of pixels smaller than the first predetermined value. The density value that becomes the set second predetermined value is calculated as the shadow side saturated density value. This second predetermined value: σ2 is also set for each color component (red component second predetermined value: σ2 # r, green component second predetermined value: σ2 # g, blue component second predetermined value: σ2 # b). Shadow side saturation density value: S is also calculated for each color component (red component shadow side saturation density value: Sr, green component shadow side saturation density value: Sg, blue component shadow side saturation density value: Sb). Any of the first predetermined values set for each color component is selected from 1 to 20%, and any of the second predetermined values set for each color component is selected from 0.01 to 1%. It is prescribed as follows.

  When the bridal correction is performed in consideration of the specified bridal gown area, the first predetermined value of each color component is related to the highlight area, that is, the expression of the white system, and thus is specified by the bridal gown area specifying unit 56. The first predetermined value may be set so that the bridal gown area has an appropriate white level. For this purpose, for example, a table may be created in which the first predetermined value is derived using the average density value of the bridal gown area for each color component and the density value of the appropriate color as parameters. On the other hand, when the bridal correction is performed without considering the specified bridal gown area, the first predetermined value is set regardless of the identified bridal gown area. Since the second predetermined value does not affect the white expression of the bridal gown in relation to the shadow region, that is, the black expression, a fixed low value can be set.

  The density correction curve generation unit 64 generates a density correction curve using the highlight side saturation density value as a highlight side saturation point and the shadow side saturation density value as a shadow side saturation point. Since the reproducibility of a pure white wedding dress is extremely important for a bridal photographed image, the shape of the density correction curve in the intermediate region between the highlight side saturation point and the shadow side saturation point may be a normal straight line. However, in the case where neutral color correction is also necessary, the statistical representative value of the input captured image as in the past, for example, the minimum value, maximum value, and average value obtained from the histogram of the entire captured image are used as parameters. It is also possible to generate a final density correction curve by adopting the density correction curve shape in the intermediate region obtained from the basic density correction curve generation unit 65 that determines the shape of the density correction curve.

  As shown in FIG. 6, the density correction module 80 stores the density correction curve set and adjusted by the bridal density correction curve generation module 60 or the normal density correction curve generation module 70 as a table accessible at high speed. A table 82 and a density correction execution unit 81 that executes density correction on a high-resolution captured image using the density correction curve table 82 are provided. Generally, the density correction execution unit 81 is realized by a program, and the density correction curve table 82 is realized by a data structure such as a lookup table. However, the present invention is limited to such a configuration form. is not.

A typical flow of density correction in the image processing unit 32 will be described with reference to the flowchart of FIG.
First, one of the input captured images is read as a density correction target captured image into the working area of the memory 30 (# 02), and face detection processing by the face detection module 40 is performed on the captured image (# 04). . If no face is detected from the photographed image (# 06 No branch), it is considered that the face is not a bridal photo with the bride as the subject, and normal normal density correction processing is performed (# 60). If a face is detected from the photographed image (# 06 Yes branch), then the bride photographing scene determination process is performed (# 08). In the bride photographing scene determination process, a determination condition in which the histogram characteristic calculated by the histogram characteristic calculating unit 51 and the image characteristic distribution of the divided section calculated by the divided image feature amount calculating unit 52 are defined as the bride photographing image characteristic, For example, if the histogram satisfies the condition that the highlight area and the shadow area are unevenly distributed, and the white color (the bride) and the black color (the groom) are unevenly distributed in the center section of the photographed image, the bride is It is determined that the bridal photograph as the subject is a photographed image including the bride in the photographed image, that is, a bride photographing scene (# 10). A technique for dividing the photographed image into a plurality of sections and evaluating the photographed scene from the image feature amounts of these sections is described in detail in Japanese Patent Laid-Open No. 2006-324987.

  If it is determined that this captured image is not a bride shooting scene (# 10 No branch), normal normal density correction processing is performed (# 60), but if this captured image is determined to be a bride shooting scene (# 10 Yes) (Branch), and a bridal gown determination process for checking whether the bridal gown exists in the photographed image is received (# 12). In this bridal gown determination process, the area having the color density pattern read from the bridal gown reference table 55 below the detected face area obtained from the face detection module 40 is also the same for the bridal gown read out from the bridal gown reference table 55. If it is similar to the shape, it is determined that a bridal gown is present in this photographed image. Further, when the bride's face image data is stored and set in advance, the presence / absence of the bridal gown can be determined by narrowing down the detected face area where the bride's face is detected using image processing technology such as pattern matching. it can.

  If it is determined that the bridal gown is not included in the photographed image (# 14 No branch), the photographed image is subjected to normal density correction processing (# 60). When it is determined that the bridal gown is included in the photographed image (# 14 Yes branch), the above-described bridal density correction processing (# 20) is performed on the photographed image. In the processing, when the information on the bridal gown area is used, the bridal gown area specifying unit 56 identifies the bridal gown area in advance, and the bridal gown area information is passed to the bridal density correction curve generation module (# 16).

Next, the flow of bridal density correction processing will be described using the flowchart of FIG.
First, a density histogram of a photographed image to be processed is generated for each color component (# 20). If the same histogram as that generated in the bride photographing scene determination process in step # 08 is used, it is used. Can be diverted. Subsequently, a first predetermined value σ1 determined and stored in advance by statistical investigation is set for each color component. At this time, when the bridal gown area information is used, the bridal gown area for each color component is set. The first predetermined value: σ1 may be corrected using the average density value and the density value of the appropriate color as parameters. (# 22). Subsequently, the number of pixels (frequency value) is cumulatively calculated from the density value on the highlight side of the density histogram (# 24). This cumulative calculation is performed until the ratio of the cumulative number of pixels to the total number of pixels: Z exceeds the first predetermined value of the corresponding color component (# 26). The density value when the ratio of the total number of pixels to the total pixel number Z exceeds the first predetermined value is stored as the highlight-side saturated density value of the corresponding color component (# 28). Next, the second predetermined value: [sigma] 2 is set for each color component so as to be a value considerably smaller than the first predetermined value: [sigma] 1 (1/10 to 1/1000 of the first predetermined value) (# 30). The number of pixels (frequency value) is cumulatively calculated from the density value on the shadow side of the density histogram (# 32). This cumulative calculation is performed until the ratio of the cumulative number of pixels to the total number of pixels: Z exceeds the second predetermined value of the corresponding color component (# 34). The density value when the ratio of the total number of pixels: Z to the second predetermined value exceeds the second predetermined value is stored as the shadow side saturated density value of the corresponding color component (# 36).

  The highlight-side saturation density value and the shadow-side saturation density value calculated in this way are set as the highlight-side saturation point and the shadow-side saturation point, respectively, and the intermediate region is a straight line (of course, the information from the basic density correction curve generation unit 65 is used. A density correction curve is generated (# 38). The generated density correction curve data is sent from the density correction curve generator 64 to the density correction curve table 52 of the density correction module 50, where it is tabulated so that it can be accessed at high speed (# 40). The density correction execution unit 82 performs density correction on the bridal photographed image developed in the memory 30 and using the density correction curve table 81 (# 42).

  When the specified captured image is subjected to density correction through either the bridal density correction process or the normal density correction process, the remaining uncorrected captured image is checked, and density correction is performed on all input bridal captured images. (# 80).

  In the density correction process shown in FIG. 7, a photographed image in which no face is detected or a photographed image that is not regarded as a bride photography scene based on the bride photography scene determination process is subjected to a normal density correction process. (# 10 No branch, # 60) Instead, as shown in the flowchart of FIG. 9, the bridal shooting scene determination is again performed on the shot image in which the face is not detected or the shot image not considered as the bride shooting scene. Processing may be performed to perform a recheck to receive either the bridal density correction process or the normal density correction process. In this bridal shooting scene determination process (# 18), image analysis such as histogram analysis may be adopted. However, if GPS position information is included in the attribute data of the shot image, the GPS position and bridal ceremony can be held. A bridal photographed image may be determined from a check with a different place. In any case, as a result of this bridal shooting scene determination process, if it is determined that this shot image is a bridal shooting scene (# 19 Yes branch), the bridal gown is more important than the shot image considered as a bride shooting scene. Since there is a high possibility that the subject is not a subject, when the bridal density correction process is performed in which the first predetermined value: σ1 is slightly adjusted to the highlight side and it is determined that the photographed image is not a bridal photographing scene (# 19 No branch), Naturally, normal density correction processing is performed.

  Although not shown in the flowcharts of FIGS. 7 and 9, when the photographed image is determined to be subjected to the bridal density correction process, the operator is informed of this by a buzzer or monitor screen display, and the bridal is displayed to the operator. You may employ | adopt the structure which notifies that the picked-up photograph was detected.

  In the description of the above-described embodiment, the bridal density correction is performed by the bridal density correction curve created based on the principle shown in FIG. When the area is specified, density correction may be performed to perform hue conversion so that the bridal gown area has a preset hue.

  Also, if face detection is successful, adjusting the shape of the bridal density correction curve according to the hue of the face area or adjusting the degree of hue conversion will produce a bridal photo with better overall color balance. Can be output.

  Further, the print station 1B exposes a photographed image to the photographic paper P by a print exposure unit 14 equipped with an exposure engine, and a so-called silver salt photographic printing system in which the photographic paper P after the exposure is subjected to a plurality of development processes. Of course, the print station 1B using the density correction curve determination technique according to the present invention is not limited to such a method. For example, an image is ejected onto a film or paper to eject an image. Various photographic printing methods such as an inkjet printing method for forming a film and a thermal transfer method using a thermal transfer sheet can be employed. Further, the bridal density correction curve generation technique according to the present invention may be mounted on a self-type photographic printing apparatus such as a kiosk terminal.

External view of photographic printing apparatus employing density correction curve generation technology according to the present invention Schematic diagram schematically showing the configuration of the print station of the photo printing device Functional block diagram explaining the functional elements built in the controller of the photo printing device Functional block diagram showing the bride scene detection module Schematic diagram showing the basic principle of the technique according to the present invention for generating a density correction curve Functional block diagram showing the functional configuration of the bridal density correction curve generation module Flow chart showing the overall flow of density correction processing Flow chart showing the flow of density correction processing of a bridal photographed image The flowchart which shows the flow of the whole density correction processing in another embodiment

Explanation of symbols

30: Memory 40: Face detection module 50: Bride scene detection module 51: Histogram characteristic calculation unit 52: Divided image feature amount calculation unit 53: Bride photographing scene determination unit 54: Bridal costume determination unit 55: Bride costume reference table 56: Bride Costume area specifying unit 60: bridal density correction curve generation module 70: normal density correction curve generation module 80: density correction module 81: density correction curve table 82: density correction execution unit

Claims (8)

  Of the input photographed images, those having the bride-photographed image characteristics are selected as bride-photographed images, and it is determined whether a bridal gown is present based on face detection information and color density information in the bride-photographed image, An image density correction method for performing density correction so that the bridal gown photographed image determined based on the determination result is an appropriate color.   Of the input photographed images, those having the bride-photographed image characteristics are selected as bride-photographed images, and it is determined whether a bridal gown is present based on face detection information and color density information in the bride-photographed image, An image density for specifying a bridal gown area to be corrected of the bridal gown photographed image determined as a bridal gown photographed image based on the determination result, and performing density correction so that the bridal gown in the bridal gown photographed image has an appropriate color Correction method.   The frequency characteristics of a shadow area and a highlight area in a histogram of a captured image and / or a pixel value distribution for each divided section of the captured image are used as the bride captured image characteristics. Image density correction method.   4. The face detection information used when determining whether or not a bridal gown exists from the photographed image of the bride is face detection information related to a face similar to a bride's face set in advance. The image density correction method according to any one of the above.   5. The image density correction method according to claim 1, wherein the bridal gown reference shape information is taken into account when determining whether or not a bridal gown is present in the bridal photographed image.   The image density correction method according to claim 1, wherein the appropriate color is adjusted according to a hue of a face region obtained based on the face detection information in the density correction.   A bride photographing scene determination unit that selects a photographed image of a bride as a photographed image of the bride among the input photographed images, and whether a bridal gown exists based on face detection information and color density information in the bride photographed image An image processing unit comprising: a bridal gown determination unit for determining whether or not; and a density correction module so that the bridal gown photographed image determined based on the determination result has a proper color.   A bride photographing scene determination unit that selects a photographed image of a bride as a photographed image of the bride among the input photographed images, and whether a bridal gown exists based on face detection information and color density information in the bride photographed image A bridal gown determination unit for determining whether or not, a bridal gown area specifying unit for specifying a bridal gown area to be corrected of the bridal shooting image determined based on the determination result, and the bridal gown shooting image. An image processing unit including a density correction module that allows the bride's costumes to have appropriate colors.

Images