JP2013201477A - Image processing apparatus, imaging apparatus, and image processing program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a technique capable of quickly and accurately obtaining an amount of adjustment of white balance processing on an image taken under plural light sources.SOLUTION: An image processor comprises: a scene recognition unit which calculates the feature amount of a target image being processed and recognizes the scene of the subject of the target image on the basis of the feature amount; a storage unit which previously stores weight information representing a likelihood that the subject is one of the plural scenes by each scene; a white balance calculation unit which proportionally divides a recognition result obtained by the scene recognition unit with the weight information to calculate and obtain a white balance adjustment value on the target image; and a white balance processing unit which on the basis of the white balance adjustment value, performs white balance processing on the target image.

Description

  The present invention relates to an image processing apparatus, an imaging apparatus, and an image processing program capable of acquiring a white balance adjustment value according to a scene.

  Conventionally, white balance processing for a captured image is performed based on, for example, the wavelength characteristics of one light source specified in advance by a user or specified by scene recognition for a subject. However, since the actual subject is illuminated with various light sources, the white balance processing is not performed correctly, and the subject of the captured image may differ from the actual subject.

  In order to solve this, for example, the captured image is divided into a plurality of blocks, and each light source illuminates the scene of each block based on the representative color obtained in each block and the typical color difference of each of the plurality of light sources. There is a technology that estimates the reliability and the contribution component for the entire image, and performs white balance processing by estimating the illumination color of the composite illumination that illuminates the scene of the entire image according to the reliability and the contribution component (for example, Patent Document 1).

Patent No. 4081213

  However, in the prior art such as Patent Document 1, it is necessary to obtain the reliability and contribution component for each light source in each block, and there is a problem that the processing amount increases as the image size increases and the processing takes time.

  In view of the problems of the prior art, it is an object of the present invention to provide a technique that can quickly and accurately acquire an adjustment amount of white balance processing for an image captured under a plurality of light sources.

  In order to solve the above-described problem, an aspect of an image processing apparatus illustrating the present invention provides a scene recognition unit that calculates a feature amount of a target image to be processed and recognizes a subject scene of the target image based on the feature amount. And a storage unit that preliminarily stores weight information indicating the probability of each scene in which the subject is one of a plurality of scenes, and a white balance adjustment for the target image by dividing the recognition result by the scene recognition unit into weight information. A white balance calculation unit that calculates and acquires a value; and a white balance processing unit that performs white balance processing of the target image based on the white balance adjustment value.

  Further, the weight information may be different depending on the luminance of the subject.

  In addition, the white balance calculation unit may detect white from the target image and apportion the recognition result by the scene recognition unit as weight information based on the number of pixels detected as white.

  Further, the scene recognition unit may recognize the scene of the subject based on the feature amount and the discrimination criterion obtained based on supervised learning.

  Further, the discrimination criterion may be obtained by applying a plurality of images captured by a light source other than the flash device and specifying a light source for illuminating the subject in advance to supervised learning.

  The scene may be a light source that illuminates the subject, and the scene recognition unit may identify the light source.

  One aspect of an imaging apparatus illustrating the present invention includes an imaging unit that captures an image of a subject and generates a target image, and the image processing apparatus of the present invention.

  In addition, a flash device for flashing a subject is provided, and when the imaging unit captures a target image using the flash device, a test image without flash is captured, and a white balance calculation unit is configured to calculate the test image and the target image. A change amount of the luminance distribution may be calculated, and the white balance adjustment value may be adjusted based on the change amount.

  One aspect of the image processing program illustrating the present invention is to calculate a feature amount of a target image to be processed, and to recognize a scene of a subject of the target image based on the feature amount, and a recognition result by the scene recognition procedure. , Based on the white balance adjustment value, the white balance calculation procedure for calculating and acquiring the white balance adjustment value for the target image by dividing with the weight information indicating the probability of each scene where the subject is one of a plurality of scenes The computer executes a white balance processing procedure for performing white balance processing on the target image.

  ADVANTAGE OF THE INVENTION According to this invention, the adjustment amount of a white balance process can be acquired rapidly and highly accurately with respect to the image imaged under the several light source.

1 is a block diagram showing a configuration of a digital camera according to an embodiment of the present invention. The figure which shows an example of the list of the weighting coefficient according to the luminosity of a photographic subject The figure which shows an example of the function of the degree which refers the scene recognition result according to the brightness | luminance of a to-be-photographed object The figure which shows an example of the function of the influence degree with respect to the white balance adjustment value by a flash device The flowchart of the processing operation of the digital camera which concerns on one Embodiment of this invention.

  FIG. 1 is a block diagram showing a configuration of a digital camera according to an embodiment of the present invention.

  The digital camera of this embodiment includes an imaging optical system 11, an imaging device 12, an AFE 13, a buffer memory 14, an image processing unit 15, a monitor 16, a flash emission control unit 17, a flash device 18, a CPU 19, an operation member 22, and a recording I / O. F23, a storage unit 25, and a bus. The buffer memory 14, the image processing unit 15, the monitor 16, the flash emission control unit 17, the CPU 19, the recording I / F 23, and the storage unit 25 are connected to each other via a bus so as to be able to transmit information. The operation member 22 is connected to the CPU 19.

  The imaging optical system 11 includes a plurality of lens groups including a zoom lens and a focusing lens. The lens position of the imaging optical system 11 is adjusted in the optical axis direction by a lens driving unit (not shown). For simplicity, the imaging optical system 11 is illustrated as a single lens in FIG.

  The image pickup device 12 is a device that picks up an image of a subject imaged by a light beam that has passed through the image pickup optical system 11. The output of the image sensor 12 is input to the AFE 13. Note that the image sensor 12 of the present embodiment may be a progressive scan type solid-state image sensor (CCD or the like) or an XY address type solid-state image sensor (CMOS or the like).

  A plurality of light receiving elements are arranged in a matrix on the light receiving surface of the image sensor 12. In each light receiving element of the image sensor 12, red (R), green (G), and blue (B) color filters are arranged according to a known Bayer array. Therefore, each light receiving element of the imaging element 12 outputs an image signal corresponding to each color by color separation in the color filter. Thereby, the image sensor 12 can acquire a color image.

  Here, in the shooting mode of the digital camera, the image pickup device 12 picks up a recorded image (main image) in response to a full pressing operation of a release button constituting the operation member 22. Further, the imaging element 12 in the shooting mode captures a composition confirmation image (through image) at a predetermined time interval even during shooting standby. The through image data is output from the image sensor 12 by thinning-out readout. The through image data is used for image display on the monitor 16 and various arithmetic processes by the CPU 19.

  The AFE 13 is an analog front end circuit that performs analog signal processing on the output of the image sensor 12. The AFE 13 performs correlated double sampling, image signal gain adjustment, and image signal A / D conversion. The output of the AFE 13 is temporarily recorded in the buffer memory 14. In the present embodiment, the imaging device 12 and the AFE 13 constitute an imaging unit. The buffer memory 14 can be used by appropriately selecting a general volatile semiconductor memory or the like, and temporarily stores image data in a pre-process or post-process of image processing by the image processing unit 15.

  The image processing unit 15 performs various types of image processing (for example, color interpolation processing, gradation conversion processing, contour enhancement processing, etc.) on the digital image signal for one frame stored in the buffer memory 14. In addition, the image processing unit 15 according to the present embodiment operates as a white balance processing unit that performs white balance processing on a digital image signal together with the above-described various image processing based on a white balance adjustment value acquired by a CPU 19 described later. .

  The monitor 16 is a display unit such as a liquid crystal monitor and displays various images in accordance with instructions from the CPU 19.

  The flash emission control unit 17 controls the flash operation of the flash device 18 according to the luminance of the subject measured by the CPU 19. When the flash device 18 is flashed to capture an image, the flash light emission control unit 17 causes the flash device 18 to perform pre-light emission for determining the amount of main light emission at the time of main image capturing. The CPU 19 calculates the light emission amount of the main light emission based on the luminance of the subject by the pre-light emission. At the time of actual imaging, the flash light emission control unit 17 causes the flash device 18 to perform main light emission based on the calculated light emission amount. As the light source of the flash device 18, a xenon lamp, a light emitting diode (LED), or the like can be appropriately selected and used.

  The CPU 19 is a processor that comprehensively controls each unit of the digital camera. The CPU 19 executes a control program and an image processing program to perform known automatic focus (AF) control by a phase difference detection method or a contrast detection method, automatic exposure (AE) calculation, and the like based on the data of the through image. Do. In addition, when the release button of the operation member 22 is pressed halfway by the user, the CPU 19 performs various arithmetic processes in the imaging mode, such as photometry of the luminance of the subject using a through image, and generates a file of the captured image. Perform processing and display processing on the setting screen. Furthermore, the CPU 19 of this embodiment operates as a scene recognition unit 20 and a white balance (WB) calculation unit 21 that acquire a white balance adjustment value used for white balance processing by the image processing unit 15 by executing an image processing program.

  The scene recognition unit 20 calculates feature quantities such as luminance and color difference of the main image (target image) captured by the image sensor 12, and uses the feature quantities including the exposure condition of the target image and a known scene recognition method. Based on this, the scene of the subject is recognized, and the light source that illuminates the subject is specified. Here, in this embodiment, for example, supervised learning such as pattern matching or neural network is captured using a light source other than the flash device, and the light source is pre-specified on a plurality of sample images of various subjects. It is assumed that a determination criterion for associating a feature amount such as an exposure condition, luminance, and color difference of each sample image with a wavelength characteristic of each light source is obtained in advance, and that the determination criterion is stored in the storage unit 25. . The scene recognition unit 20 identifies a light source that illuminates the subject of the target image based on the determination criteria stored in the storage unit 25.

  The WB calculation unit 21 calculates and acquires a white balance adjustment value for the target image based on the recognition result of the scene recognition unit 20 and the measured luminance of the subject. The operation of the WB calculation unit 21 of this embodiment will be described in detail later.

  The operation member 22 includes, for example, a release button, a command dial, a cross-shaped cursor key, a determination button, and the like. The operation member 22 receives various inputs of the digital camera from the user. For example, the operation member 22 is used for an imaging operation, a digital camera operation mode switching operation, an input operation on a setting screen, and the like.

  The recording I / F 23 is formed with a connector for connecting the storage medium 24. The recording I / F 23 writes / reads data to / from the storage medium 24 connected to the connector. The storage medium 24 is composed of a hard disk, a memory card incorporating a semiconductor memory, or the like. In FIG. 1, a memory card is shown as an example of the storage medium 24.

  The storage unit 25 stores main image data captured by a digital camera, a control program executed by the CPU 19, an image processing program, and the like. The storage unit 25 stores the list shown in FIG. 2 and the functions shown in FIG. 3 and FIG. 4 together with the data of the judgment criteria referred to by the scene recognition unit 20. Note that a nonvolatile semiconductor memory or the like can be used for the storage unit 25.

  Here, the list of FIG. 2 and the functions of FIGS. 3 and 4 stored in the storage unit 25 will be briefly described. FIG. 2 is a list of weighting factors indicating the probability of each light source that illuminates the subject according to the luminance of the subject. This list uses, for example, the plurality of sample images to determine the distribution of light sources that irradiate the subject of the sample image for each subject brightness (BV), and based on the distribution, the probability of each light source is a weighting factor. As a list. For example, when BV = 5, sample images taken under clear and cloudy weather are distributed at 40%, and sample images taken under illumination of fluorescent lamps and light bulbs are distributed at 20%. The value of the weighting factor of the light source is set. Then, as shown in the list of FIG. 2, as the BV value becomes smaller and the subject becomes darker, it becomes difficult to distinguish the light sources, and the value of the weight coefficient of each light source is set equal. On the other hand, as the BV value increases and the subject becomes brighter, the probability that the light source is clear or shaded is higher than that of the fluorescent light or electric light. The

  As shown in FIG. 2, the light source of this embodiment is a clear sky, a shade, a fluorescent lamp, or a light bulb. However, a light source such as a cloudy sky or a mercury lamp may be included. In this embodiment, the value of the weighting factor is set to 0 to 32. However, it is preferable that the weighting factor is appropriately set according to the required accuracy of determining the white balance correction value and the processing capability of the CPU 19.

  The function shown in FIG. 3 relates the degree of referring to the recognition result by the scene recognition unit 20 in accordance with the luminance of the subject when the WB calculation unit 21 acquires the white balance adjustment amount for the target image. That is, as described above, when the BV value is small and the subject is dark, it is difficult to specify the light source that illuminates the subject, and the reliability of the recognition result by the scene recognition unit 20 is low. Therefore, in the present embodiment, as shown in FIG. 3, when BV = 6 is smaller, the WB calculation unit 21 calculates and acquires the white balance adjustment value based only on the weight information shown in FIG. On the other hand, when the BV value is large and the subject is bright, it is easy to specify the light source that illuminates the subject. Therefore, the WB calculation unit 21 of the present embodiment attaches importance to the recognition result of the scene recognition unit 20 as the BV value increases to 6 or more, and proposes a weighting coefficient (weight information) corresponding to the BV value of FIG. And calculate and obtain the white balance adjustment value. In FIG. 3, the horizontal axis indicates the luminance (BV) of the subject, and the vertical axis indicates the reference degree α of the recognition result of the scene recognition unit 20.

  The function shown in FIG. 4 relates the degree of influence of the flash device 18 on the white balance adjustment value according to the contribution degree of the flash device 18 when the target image is captured using the flash device 18. is there. That is, when the target image is captured using the flash device 18, the WB calculation unit 21 of the present embodiment acquires a white balance adjustment value that takes into account the influence of the flash device 18 based on the function shown in FIG. . The degree of contribution of the flash device 18 is an index defined according to the amount of change in the luminance distribution in images captured before and after the flash device 18 emits light. In the present embodiment, a through image (test image) is used as an image before light emission of the flash device 18, and a target image is used as an image after light emission. In FIG. 4, the horizontal axis represents the contribution of the flash device 18, and the vertical axis represents the influence β of the flash device 18.

  Next, white balance processing by the digital camera of the present embodiment will be described with reference to the flowchart of FIG. However, it is assumed that the digital camera of this embodiment is set to the imaging mode in advance. In the following description, the digital camera according to the present embodiment will be described with respect to an operation in which a subject is captured in a single image to acquire a main image (target image) of a still image. Detailed description is omitted.

  The CPU 19 accepts a power-on instruction by turning on the power button on the operation member 22 by the user, and turns on the power of the digital camera. The CPU 19 reads and executes the control program and the image processing program from the storage unit 25, and initializes the digital camera. The CPU 19 causes the image sensor 12 to capture a through image at a frame rate such as 30 fps, and causes the monitor 16 to display the through image. The CPU 19 starts processing from step S101.

  Step S101: The CPU 19 determines whether or not a half press operation of the release button of the operation member 22 by the user has been accepted. When the CPU 19 receives a half-press operation from the user, the CPU 19 proceeds to step S102. On the other hand, if the CPU 19 does not accept the half-press operation (NO side), the CPU 19 waits until the half-press operation is accepted from the user.

  Step S102: When receiving a half-press operation from the user, the CPU 19 measures the luminance of the subject based on the through image captured by the image sensor 12. The CPU 19 calculates an exposure condition by performing an AE calculation based on the measured luminance. Note that if the CPU 19 determines to perform flash imaging by the flash device 18 as a result of subject photometry or calculation of exposure conditions, the through image used for photometry is temporarily recorded in the buffer memory 14.

  Step S103: The CPU 19 determines whether or not the user has fully pressed the release button of the operation member 22. When the CPU 19 accepts the full-press operation from the user, the CPU 19 causes the image sensor 12 to perform the actual imaging on the assumption that the subject imaging instruction has been accepted. The CPU 19 temporarily records the target image converted into a digital signal and output from the AFE 13 in the buffer memory 14, and proceeds to step S104 (YES side). On the other hand, when the CPU 19 does not accept the full-press operation (NO side), the CPU 19 stands by until an imaging instruction is received from the user.

  Step S104: The scene recognition unit 20 of the CPU 19 reads the target image from the buffer memory 14 and calculates feature quantities such as luminance and color difference of the target image. The scene recognition unit 20 recognizes the scene of the subject based on the feature amount including the exposure condition of the target image and the determination criterion of the storage unit 25, and identifies the light source that illuminates the subject.

  Step S105: The WB calculation unit 21 of the CPU 19 calculates and acquires a white balance adjustment value for the target image based on the recognition result of the scene recognition unit 20 and the measured luminance of the subject. First, the WB calculation unit 21 sets a weighting factor for the light source specified by the scene recognition unit 20 in order to apportion the recognition result of the scene recognition unit 20 with the weight information of the list shown in FIG. That is, for example, when the scene recognition unit 20 recognizes “sunny sky”, the WB calculation unit 21 assumes that the reliability of the recognition result is 100% (regardless of the BV value) (sunny sky, cloudy sky, fluorescent light, electric light) ) = (32, 0, 0, 0) and a weighting coefficient are set. Alternatively, when the scene recognizing unit 20 recognizes “fluorescent lamp”, the WB calculating unit 21 sets (0, 0, 32, 0) regardless of the BV value (clear sky, cloudy sky, fluorescent lamp, electric light). To do.

Next, the WB calculation unit 21 performs known white detection on the target image and extracts pixels having a white image value. For example, the WB calculation unit 21 specifies a light source corresponding to the color temperature of each pixel based on values such as G / R and G / B of each pixel detected as white. The WB calculation unit 21 counts the number of pixels in which white is detected for each light source. Here, n ₁ , n ₂ , n ₃ , and n ₄ are the numbers of pixels counted for each of clear sky, shade, fluorescent lamp, and electric lamp. In the following description, the light sources of clear sky, shade, fluorescent light, and electric light correspond to the symbols i = 1, 2, 3, 4, respectively.

The WB calculation unit 21 determines the reference degree α corresponding to the luminance of the subject measured in step S102 based on the function shown in FIG. 3, and the recognition result of the scene recognition unit 20 based on the following equation (1) The weight information of the list shown in FIG.
W _Gi = α × W _Si × n _i + (1−α) × W _i × n _i (1)
Here, W _Si , W _i , and W _Gi respectively indicate the weight coefficient of the recognition result of the scene recognition unit 20 for the light source i, the weight coefficient of the list of FIG. Indicates the coefficient. As an example, the recognition result of the scene recognition unit 20 is "sunny", when the photometry has been subject luminance _{BV = 9, (W S1,} W S2, W S3, W S4) = (32,0,0,0 ), (W ₁ , W ₂ , W ₃ , W ₄ ) = (32, 16, 0, 0), α = 0.25.

Finally, the WB calculation unit 21 calculates and obtains the white balance adjustment value ε for the target image, for example, by weighted averaging the wavelength characteristics of each light source based on the weighted weighting factor W _Gi of each light source i. To do.

  Step S106: The CPU 19 determines whether or not the flash device 18 flashes in the main imaging in step S103 based on the flag signal from the flash light emission control unit 17. The CPU 19 proceeds to step S107 (YES side) when the flag signal that the flash device 18 flashes in the main imaging is received from the flash light emission control unit 17, and proceeds to step S108 (NO) when the flag signal is not received. Side).

Step S107: The WB calculation unit 21 adjusts the white balance adjustment value ε acquired in step S105 according to the degree of contribution of the flash device 18 that has flashed light to the illumination of the subject. For this purpose, the WB calculation unit 21 calculates the luminance distribution of the subject before the flash device 18 emits light by applying the through image recorded in the buffer memory 14 in step S102 to the following equation (2).
Luminance distribution before light emission DBb = (average luminance in the central area of the through image) / (average luminance in the peripheral area of the through image) (2)
At the same time, the WB calculation unit 21 calculates the luminance distribution of the subject after light emission by applying the target image to the following equation (3).
Luminance distribution after light emission DBa = (average luminance in the central area of the target image) / (average luminance in the peripheral area of the target image) (3)
The WB calculating unit 21 calculates the amount of change in the luminance distribution of the subject due to the flash emission of the flash device 18 as a contribution based on the following equation (4).
Contribution = | luminance distribution DBa−luminance distribution DBb | (4)
The WB calculating unit 21 applies the contribution degree to the function shown in FIG. 4 and acquires the value of the influence degree β with respect to the white balance adjustment value ε by the flash device 18. The WB calculation unit 21 apportions the white balance adjustment value ε and the white balance adjustment value ε _{f of the} flash device 18 based on the following equation (5), and acquires the adjusted white balance adjustment value ε ′. It is assumed that the white balance adjustment value ε _f of the flash device 18 is stored in the storage unit 25 in advance.
ε ′ = (1−β) × ε + β × ε _f (5)
Step S108: The CPU 19 outputs the white balance adjustment value ε ′ acquired in step S105 or the white balance adjustment value ε ′ acquired in step S107 to the image processing unit 15. The image processing unit 15 performs white balance processing on the target image using various received image processing such as color interpolation processing, gradation conversion processing, and contour enhancement processing using the received white balance adjustment value.

  Step S109: The CPU 19 generates an image file of the target image subjected to image processing. The CPU 19 records the image file in the storage medium 24, the storage unit 25, and the like, and ends a series of processes.

  It should be noted that the CPU 19 according to the present embodiment, when generating an image file, uses the Exif (Exchangeable image file format for digital still) information such as the exposure condition and the determination result of the scene recognition unit 20 and the presence or absence of the flash operation of the flash device 18. Preferably, metadata compliant with the cameras) standard is added to the header area of the image file. Thus, the same white balance processing can be performed by causing the other computer to read the image file.

  As described above, in the present embodiment, the light source identified by scene recognition of the captured target image is divided into the weight information that is the probability of each light source in the photometric luminance, and is thus imaged under a plurality of light sources. The white balance adjustment amount can be acquired with high accuracy even for the target image that has been set.

  Further, by considering the influence of the flash device 18 individually, the white balance adjustment amount for the target image can be acquired with higher accuracy.

Further, by specifying the light source that irradiates the subject of the target image by scene recognition, the white balance adjustment amount for the target image captured under a plurality of light sources can be quickly acquired without causing an increase in the amount of calculation.
<< Additional items of embodiment >>
(1) In the above embodiment, an example in which the CPU 19 realizes each process of the scene recognition unit 20 and the WB calculation unit 21 by software has been described. However, each of these processes is realized by hardware using an ASIC. May be.

  (2) The image processing apparatus of the present invention is not limited to the example of the digital camera of the above embodiment. For example, the computer may function as the image processing apparatus of the present invention by causing the computer to read an image and causing the computer to execute an image processing program.

  (3) In the above embodiment, the CPU 19 measures the luminance of the subject from the through image of the image sensor 12. However, the present invention is not limited to this. For example, the digital camera measures the luminance of the subject together with the image sensor 12. A photometric sensor may be provided. In that case, in step S107, it is preferable to calculate the luminance distribution DBb before light emission calculated using Expression (2) from the image captured by the photometric sensor.

  (4) In the above embodiment, the functions shown in FIGS. 3 and 4 are used to obtain the reference degree α of the recognition result of the scene recognition unit 20 and the degree of influence β on the white balance adjustment value by the flash device 18. However, the present invention is not limited to this, and it is preferable to use a function corresponding to the subject of the target image, the required accuracy of white balance processing, and the like.

  (5) In the above embodiment, each light source in the list shown in FIG. 2 has a different weighting factor depending on the luminance, but the present invention is not limited to this. For example, each light source may have one weight coefficient regardless of luminance.

  From the above detailed description, features and advantages of the embodiments will become apparent. It is intended that the scope of the claims extend to the features and advantages of the embodiments as described above without departing from the spirit and scope of the right. Further, any person having ordinary knowledge in the technical field should be able to easily come up with any improvements and modifications, and there is no intention to limit the scope of the embodiments having the invention to those described above. It is also possible to use appropriate improvements and equivalents within the scope disclosed in.

DESCRIPTION OF SYMBOLS 11 ... Imaging optical system, 12 ... Imaging device, 13 ... AFE, 14 ... Buffer memory, 15 ... Image processing part, 16 ... Monitor, 17 ... Flash emission control part, 18 ... Flash apparatus, 19 ... CPU, 20 ... Scene recognition , 21 ... WB calculation unit, 22 ... operation member, 23 ... recording I / F, 24 ... storage medium, 25 ... storage unit

Claims (9)

A scene recognition unit that calculates a feature amount of a target image to be processed and recognizes a subject scene of the target image based on the feature amount;
A storage unit that preliminarily stores weight information indicating a probability for each scene in which the subject is one of a plurality of scenes;
A white balance calculation unit that divides the recognition result by the scene recognition unit into the weight information and calculates and acquires a white balance adjustment value for the target image;
A white balance processing unit that performs white balance processing of the target image based on the white balance adjustment value;
An image processing apparatus comprising: The image processing apparatus according to claim 1.
The image processing apparatus according to claim 1, wherein the weight information varies depending on luminance of the subject. The image processing apparatus according to claim 1 or 2,
The white balance calculation unit performs white detection on the target image, and based on the number of pixels detected as white, the recognition result by the scene recognition unit is divided into the weight information. apparatus. The image processing apparatus according to any one of claims 1 to 3,
The image processing apparatus, wherein the scene recognition unit recognizes the scene of the subject based on the feature amount and a discrimination criterion obtained based on supervised learning. The image processing apparatus according to claim 4.
The image processing is characterized in that the determination criterion is obtained by applying to the supervised learning a plurality of images that are captured by a light source other than a flash device and a light source that illuminates the subject is specified in advance. apparatus. The image processing apparatus according to any one of claims 1 to 5,
The scene is a light source that illuminates the subject,
The image processing apparatus, wherein the scene recognition unit identifies the light source. An imaging unit that images a subject and generates a target image;
An image processing apparatus according to any one of claims 1 to 6,
An imaging apparatus comprising: The imaging apparatus according to claim 7,
A flash device for flashing the subject;
The imaging unit, when capturing the target image using the flash device, captures a test image without flash,
The white balance calculation unit calculates an amount of change in luminance distribution between the test image and the target image, and adjusts the white balance adjustment value based on the amount of change. A scene recognition procedure for calculating a feature amount of a target image to be processed and recognizing a subject scene of the target image based on the feature amount;
A white balance obtained by calculating a white balance adjustment value for the target image by dividing the recognition result obtained by the scene recognition procedure into weight information indicating the probability of each scene where the subject is one of a plurality of scenes. Balance calculation procedure,
A white balance processing procedure for performing white balance processing of the target image based on the white balance adjustment value;
An image processing program for causing a computer to execute.

Images