JP2006262000A - Image processing apparatus and its processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing apparatus and its processing method by which a color-conditioned motion picture quality can be stably displayed. <P>SOLUTION: An imaging apparatus 10 implements the following operations of; performing control by a control signal 46 output from a CPU 16, inputting present frame image data 36 to a correction object detector 38 of a signal processor 14, retrieving past correction target information 48 contained in a first correction information output from a memory of the CPU 16, detecting a correction object in the correction object detector 38 based on the past correction target information 48, detecting a correction object to calculate image correction information of the present frame image data 36 for serving as a second correction information according to the result detected, supplying image correction information 50 calculated to a digital processor 40, and correcting the present frame image by using at least one of parameters of a gradation sequence and color correction of the image correction information 50 in the digital processor 40. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to an image processing apparatus and a processing method thereof. The image processing apparatus of the present invention relates to color conversion for a moving image taken by a digital camera or a video camera. The image processing method of the present invention also relates to a method applied to reproduction of captured moving images and recorded moving images.

  When an image recorded on a silver salt film such as a photographic negative film or a reversal film is color-printed, the captured image is subjected to color conversion so as to be similar to the scene at the time of shooting. However, when a digital image recorded by the image sensor is printed, a phenomenon may occur in which colors are biased by the light source and surrounding colors.

  The image processing apparatus of Patent Document 1 detects a face or the like from an image as a specific subject, and calculates a correction parameter so that the detected face color becomes a desired color. Using this calculated correction parameter, the entire digital image is corrected to improve the image quality.

In addition, the image processing apparatus disclosed in Patent Document 2 is arranged so that the face area has an appropriate color and density due to difficulty in improving the entire digital image, in other words, the face area color is not affected by the color density around the face area. Adjust the density.
JP 2000-182043 JP 2000-242775 A JP-A-5-307605

  By the way, these adjustments are improvements in image quality for still images. When this image quality improvement is applied to a moving image, there are times when a specific subject, that is, a face can be detected and when it cannot be detected. In a moving image, two color adjustments or color conversions are performed depending on whether or not this detection is possible.

  First, if the face can be detected, the color of the frame image is adjusted based on a desired color, for example, skin color, and second, the color based on another standard for the partial frame image when the face cannot be detected Adjust with. As a result, when a moving image color-adjusted by the two methods is displayed on the display device, the image quality varies in time series, causing abrupt fluctuations in color expression and making the moving image feel unstable.

  It can be considered that detection is possible only by a slight change in the position of the face and the size of the face taken in the continuous frame image, for example.

  SUMMARY OF THE INVENTION An object of the present invention is to provide an image processing apparatus and a processing method therefor that can solve such drawbacks of the prior art and can stably display the quality of a color-adjusted moving image.

  In order to solve the above-described problems, the present invention provides an image processing apparatus for performing image processing on a current frame image of a digital moving image, wherein the apparatus is a current frame image obtained from a past frame image corresponding to the current frame image. Correction information storage means for storing first correction information relating to the correction object, correction object detection means for detecting a correction object included in the current frame image and obtaining correction object information for the correction object, and correction object detection means Correction information calculating means for calculating second correction information for correcting the current frame image based on the correction object information and the first correction information in the obtained current frame image; and the current frame based on the second correction information. Image correction means for correcting the image.

  The image processing apparatus of the present invention inputs the current frame image to the correction target detection unit, reads the first correction information from the information storage unit, detects the correction target by the correction target detection unit, and detects the correction target according to the detection result. Information is obtained, second correction information in the current frame image is calculated based on the correction target information and the first correction information in the obtained current frame image, and the calculated second correction information is supplied to the image correction means. By correcting the current frame image using the second correction information by the image correction means, the image quality of the moving image can be displayed stably.

  In order to solve the above-described problems, the present invention provides an image processing method for performing image processing on a current frame image of a digital moving image, wherein the method is the current frame image obtained from a past frame image corresponding to the current frame image. A first step of reading the first correction information from the correction information storing means for storing the first correction information relating to the correction target of the frame image; and the correction target information for the correction target detected by detecting the correction target included in the current frame image A second step of calculating the second correction information for correcting the current frame image based on the correction target information and the first correction information in the current frame image, and a second correction information And a fourth step of correcting the current frame image based on the above.

  According to the image processing method of the present invention, the current frame image is input, the first correction information is read, the correction target included in the current frame image is detected, the correction target information is obtained according to this detection, and the correction target information and the first correction information are obtained. By calculating the second correction information for the current frame image from the correction information and correcting the current frame image with the second correction information, the image quality of the moving image is stabilized, and the first and second correction information are saved. This can greatly contribute to the correction of the next frame image.

  The present invention also provides the following digital camera. That is, this digital camera is obtained from an imaging means for imaging a scene to form a video signal of a digital moving image and a past frame image corresponding to the current frame image of the digital moving image upon receiving the video signal. Correction information storage means for storing first correction information relating to a correction target of the current frame image, correction target detection means for detecting a correction target included in the current frame image and obtaining correction target information for the correction target, correction Correction information calculating means for calculating second correction information for correcting the current frame image based on the correction target information and the first correction information in the current frame image obtained by the target detection means; And an image correcting means for correcting the current frame image based on the image processing unit, thereby performing image processing on the current frame image.

  Next, an embodiment of an image processing apparatus according to the present invention will be described in detail with reference to the accompanying drawings.

  In this embodiment, the image processing according to the present invention is applied to an imaging apparatus 10 that captures an object scene and forms a video signal of a digital moving image, for example, including a digital video camera or a digital electronic still camera having a moving image shooting function. The device is applied. The illustration and description of parts not directly related to the present invention are omitted. In the following description, the signal is indicated by the reference number of the connecting line in which it appears.

  As shown in FIG. 1, the imaging apparatus 10 of the present embodiment includes an image input unit 12, a signal processing unit 14, a central processing unit (CPU) 16, a memory 18, a driver 20, and a memory card interface (IF: InterFace) unit 22, display control unit 24, memory card 26, monitor 28, and operation unit 30.

  The image input unit 12 includes a lens stop 32 and an image sensor 34. The lens diaphragm 32 sets incident light from the object scene to a predetermined light amount. Incident light focused by the lens diaphragm 32 is transmitted to the image sensor 34 through a color filter (not shown), for example. The color filter of this embodiment is formed so that filter segments of the three primary colors RGB are arranged in a predetermined pattern. The filter segment has a one-to-one correspondence with a light receiving element described later. In the image input unit 12, the optical image is adjusted so as to be formed by the image sensor 34. The image sensor 34 two-dimensionally arranges light receiving elements for photoelectric conversion, and reads out signal charges accumulated in the light receiving elements over a set exposure time as pixels having color attributes. The image sensor 34 generates a two-dimensional image signal by this reading. The image input unit 12 removes signal noise (not shown) from the two-dimensional image signal, digitizes it, and generates image data. The image input unit 12 generates the image data 36 from the incident light in this way and outputs the generated image data 36 to the signal processing unit 14 and the CPU 16.

  The signal processing unit 14 has a function of appropriately performing image processing on the image data, and in particular, processing so that the color to be corrected is displayed as an image having a proper expression suitable for the scene. The signal processing unit 14 includes a correction target detection unit 38 and a digital processing unit 40. The correction target detection unit 38 has a function of detecting whether or not a correction target is included in the image data 36 and a function of calculating correction information. The correction target detection unit 38 includes a target detection function unit 42 and a correction information calculation function unit 44 as shown in FIG. 2 in accordance with these functions. The correction target detection unit 38 operates according to the control signal 46 from the CPU 16.

  In this embodiment, the case where the correction target is a face will be described as an example, but the correction target may be eyes and skin. The correction target may be one of a plurality of persons. The correction target information is information including, for example, the detection score indicating the probability of representing the brightness of the correction target, the region position and size of the correction target, or the facial appearance of the correction target, and the evaluation value of the facial appearance. The correction information is information including color correction parameters, gradation parameters, lens aperture (value), and shutter speed, for example. This information includes at least one of the above-described information.

  The object detection function unit 40 preferably includes components using the image processing method disclosed in Patent Document 3, for example. If this method is used, a predetermined subject candidate is extracted with high accuracy, so that the burden of detection in the next stage can be remarkably reduced, and the performance of the applied system can be improved.

The object detection function unit 42 includes the above-described components and inputs image data 36. The target detection function unit 42 reads the past correction target information 48 from the memory in the CPU 16. The object detection function unit 42 includes a detectable area calculation function unit 42a and a score calculation function unit 42b. The detectable area calculation function unit 42a calculates a correction target using the past correction target information 48 from the image represented by the input image data 34. Score calculation function unit 42b includes correction information Q _n provisional seeking detection score S _n in the current frame in accordance with the corrected calculated, a function of calculating the R _n.

  Here, information stored in the CPU 16, that is, past correction target information and past correction information are stored in a parallel data structure. Since the signal lines to be read are the same, the reference numerals are also described in the same way.

The correction information calculation function unit 44 has a function of calculating image correction information for the current frame image from the past correction information, detection score, and weighting coefficient. The past correction information 48 is read from the CPU 16, for example. The correction information calculation function unit 44 also operates according to the control signal 46 of the CPU 16. The correction information calculation function unit 44 uses either linear combination or non-linear combination algorithm for the image correction information P _n of the current frame image. Specific calculation will be further described later. The correction information calculation function unit 44 outputs the image correction information 50 for the calculated current frame image to the digital processing unit 40. The image correction information 50 includes at least one of a gradation correction parameter and a color correction parameter. Further, the lens aperture and shutter speed in the image correction information 50 are sent to the CPU 16 from the correction information calculation function unit 44 via the signal line 48.

  The correction information calculation function unit 44 may be calculated by the CPU 16 without being provided in the correction target detection unit 38. In this case, the detected detection score is supplied to the correction target detection unit 38 via the signal line 48 to the CPU 16. The CPU 16 sends the calculated result, that is, the correction information of the current frame image to the correction target detection unit 38 via the signal line 48.

  As shown in FIG. 3, the digital processing unit 40 includes an image correction function unit 52. The image correction function unit 52 is operated according to the control signal 46 and corrects the image data 36 based on the image correction information 50. The digital processing unit 40 performs color conversion so that the image data 36 is expressed in an optimum color. The digital processing unit 40 also performs, for example, synchronization. The signal processing unit 14 outputs image data 54 that is a processing result of the digital processing unit 40 to the memory 18.

  Returning to FIG. 1, the CPU 16 includes an arithmetic function unit, a ROM (Read Only Memory), and a RAM (Random Access Memory) (not shown). A procedure for operating the imaging apparatus 10, that is, a program is written in the ROM, and past correction information is stored in the RAM, for example. Image data 36 and operation instruction data 56 are input to the CPU 16. The CPU 16 outputs past correction information 48 and inputs correction information 48 for the current frame. The CPU 16 outputs control signals 46 and 58 based on the above-described input and program. The CPU 16 calculates how much the lens aperture and shutter speed should be based on the image data 36 and generates a control signal 58 that operates according to the calculated result. When the lens aperture and shutter speed in the correction information 48 of the current frame are supplied, the CPU 16 outputs a control signal 58 based on this correction information to the driver 20.

  The memory 18 has a buffer function for inputting and outputting the image data 54 in accordance with the control signal 46. The memory 18 is connected to the signal processing unit 14, the memory IF unit 22, and the display control unit 24. The memory 18 handles captured image data and image data reproduced from the memory card 26.

  The driver 20 generates drive signals 60 and 62 for the lens diaphragm 32 and the image sensor 34 in the image input unit 12. The driver 20 supplies a drive signal 60 to the mechanism of the lens diaphragm 32. In addition, the driver 20 outputs a drive signal 62 that defines the exposure time to the image sensor 34 and reads the signal charges accumulated with the exposure.

  The memory IF unit 22 is an interface for inputting and outputting image data 54 of the memory 18 and image data 64 of the memory card 26.

  The display control unit 24 converts the image data 54 into display data 66 for the monitor 28. The memory card 26 is a detachable semiconductor memory. The monitor 28 has a function of displaying the display data 66.

  The operation unit 30 includes a power-on switch, mode setting, a monitor display switch, a release shutter button, and the like. The operation unit 30 outputs instruction data 56 corresponding to the pressed state of the release shutter button. In the instruction data 56, the state of not pressing is S0, the half-pressed state is S1, and the fully-pressed state is S2. For example, a through image is displayed on the monitor 28 after the power is turned on. State S1 is a preliminary imaging state. The state S2 is the actual imaging timing in the still image mode, and the imaging start and end timing in the moving image mode. The operation unit 30 outputs the timing in these modes as instruction data 56 to the CPU 16.

  Next, an image processing method in the imaging apparatus 10 to which the image processing apparatus of the present invention is applied will be described. The present embodiment discloses moving image processing according to the presence or absence of a target included in an image. The moving image processing is performed on the image data 36 when the power-on button is turned on and the monitor display switch is turned on. By checking the state of the monitor display switch, useless power for monitor display can be suppressed. When the power is turned on and the monitor display switch is turned on, the imaging device 10 displays the moving image processed image on the monitor 28 even when the release shutter button is in the states S0 and S1. This state is a through image display. The imaging device 10 also performs moving image processing when displaying a through image in the still image mode or when displaying a captured image to be recorded on a monitor.

  The moving image processing in the moving image mode is performed regardless of the moving image monitor display. When the operation unit 30 sets the mode to moving image and the release shutter button in the operation unit 30 is in the state S2, that is, fully pressed, the imaging device 10 starts recording moving image shooting under the control of the CPU 16. When the release shutter button is fully pressed again, the imaging device 10 ends the recording of the moving image.

  The imaging device 10 performs image processing including moving image processing under the conditions in each mode described above. As an example, image processing when image data 36 from the image sensor 34 is displayed on the monitor 28 will be described. This image processing procedure is shown in FIG. A correction target is detected by the target detection function unit 40 of the correction target detection unit 38 with respect to the image data 36 supplied from the image sensor 34 (subroutine SUB1). The detection here refers to obtaining a correction target and obtaining a detection score in the target. The detection will be further described later.

  Next, the past correction information 48 is read from the CPU 16 (step S10). The past correction information 46 is information representing at least one of a color correction parameter and a gradation parameter in a correction target of at least about the past k = 10 samples from the current frame, and corresponds to parameters P, Q, and R described later. ing. The sampling interval is a time of a plurality of frames. This is because it takes a lot of time to detect the object. The CPU 16 outputs the read past correction information 48 to the correction information calculation function unit 42.

  The correction information calculation function unit 42 calculates the image correction information 50 in the current frame in accordance with the presence or absence of target detection (subroutine SUB2). The calculated image correction information 50 in the current frame is supplied to the digital processing unit 40 and the CPU 16, respectively.

  The digital processing unit 40 corrects the image data 36 according to the image correction information 50 (subroutine SUB3). After this correction, the imaging apparatus 10 stores the same information as the image correction information 50 for the current frame in the CPU 16 (step S12). After this saving, it is determined whether or not to end the moving image processing (step S14). The determination is made by the CPU 16. The CPU 16 makes a determination based on the value of the instruction data 56 supplied from the operation unit 30. Regardless of the mode of the operation unit 30, the moving image processing is continued when the power-on button is on and the monitor display button is on or when recording is in the moving image mode (NO: to subroutine SUB1) . In other cases (YES), in particular, when the instruction data 56 indicating that the power-on button is in the OFF state is supplied to the CPU 16, the imaging apparatus 10 proceeds to the end and stops its operation.

Next, the correction target detection process (subroutine SUB1) will be described. As shown in FIG. 5, the past correction target information 48 is read from the CPU 16 with respect to the frame image of the image data 36 (substep SS10). The past correction target information 48 is the position, size and detection score or evaluation value of the detection target. Again, the past correction target information 48 is the past correction information 48, that is, correction information Q calculated only at the correction target portion, which is at least one of color correction parameters, gradation parameters, lens aperture and shutter speed. Q _nk , Q _{n-k + 1} , ..., Q _n-1 , correction information calculated from the entire image R: R _nk , R _{n-k + 1} , ..., R _n-1 and image correction information Note that P: P _nk , P _{n−k + 1} ,..., P _n−1 is information recorded in parallel, for example, but the information content is different. Past corrected information 48, the probability or evaluation value of the face likeliness included in the frame image as a detection score S: is _{S nk, S nk + 1,} ···, even S _n-1.

  Here, if the detection score is zero, it indicates that no detection has been performed. Also, the number j of detection scores in the past correction target information 48 that are not zero is obtained.

  Next, it is compared whether or not the obtained number j is equal to or greater than the threshold number i of the number of detection scores that is not zero (sub step SS12). When the number j is equal to or greater than the threshold value i (YES), the process proceeds to region calculation processing (go to sub-step SS14). When the number j is smaller than the threshold value i (NO), the process proceeds to the detection process for the correction target (go to sub-step SS16). This comparison can avoid detection of the region, so that detection can be speeded up and power can be saved.

  In sub-step SS14, an area where the target can be detected is calculated based on the position and size included in the past correction target information. In the calculation of the detectable region, for example, the moving direction is obtained from the amount of movement of the target and the change in the position based on the difference in position and size. As a result, the object detection function unit 42 obtains the detectable area, position, and size of the correction object in the current frame. As a result, the presence or absence of a detectable region is known.

Next, a correction target in the current frame is detected (substep SS16). Obtaining a detection score S _n in the current frame by the detection. The detection method of the patent document 3 mentioned above is used for calculation of a detection score. In this detection routine, provisional correction information is further calculated according to the presence or absence of a detectable region. That is, if the detection score S _n obtained on the basis of this region, it is also calculated correction information Q _n provisional in the current frame. Further, provisional correction information R _n is calculated from the entire image for a detection score S _n = 0 where a possible area was not obtained.

In this way, the detection result in the current frame is stored (substep SS18). Here, the detection result is the presence / absence of the detection target, the detection position and the size obtained when the target is detected. Detection score S, lens aperture, shutter speed, provisional correction information Q _n , and R _n . Thereafter, the process proceeds to RETURN and ends the subroutine SUB1.

  Next, correction target calculation processing (subroutine SUB2) will be described. This calculation is performed by the correction information calculation function unit 44. First, in the calculation, it is determined whether or not there is a detection target as shown in FIG. 6 (substep SS20). When the correction target in the current frame is detected (YES), the process proceeds to calculation of correction information corresponding to the detection (subroutine SUB4). If the correction target in the current frame is not detected (NO), the process proceeds to calculation of correction information corresponding to non-detection (subroutine SUB5). The correction information includes image correction information. The image correction information includes at least one of a gradation correction parameter, a color correction parameter, a lens aperture, and a shutter speed.

  After one of these calculations, the process proceeds to return to end the subroutine SUB2.

Next, calculation of correction information corresponding to detection will be described with reference to FIG. The preferred correction information is calculated using only the correction target area detected in the current frame (sub-step SS40). And here the preferred correction information, a Q _n to calculate only the detection score S _n and corrected part in the current frame. The correction information Q _n is calculated only when the detection is successful.

  Next, past correction information is read (substep SS42). The past correction information to be read out is correction information Q, R, and P including the detection score S that is past correction target information.

Next, the combined image correction information P _n in the current frame is calculated (substep SS44). For the calculation, either linear combination or non-linear combination is used. In the calculation by linear combination, the correction information Q and P are multiplied by the weight coefficient W corresponding to the detection score by the image correction information _Pn . The linear combination is expressed by equation (1). Further, the weighting factor W is desirably a larger value as the detection score is larger, and is preferably larger as the detection score is closer to the current frame.

The image correction information P _n in the current frame is expressed by equation (1)
P _n = W _n (S _n ) Q _n + W _n-1 (S _n ) P _n-1 + ... + W _nk (S _n ) P _nk ... (1)
Is calculated by

Further, the image correction information P _n in the current frame is not limited to the equation (1), but may be performed by the equation (2) representing nonlinear coupling. Equation (2) is
P _n = f _D (S _n , Q _n , P _n-1 , P _n-2 , ..., P _nk ) (2)
It is represented by Here, f _D is a multi-dimensional non-linear function in a subject detection. After calculating in this way, the process proceeds to return and the subroutine SUB4 is terminated.

Next, calculation of correction information corresponding to non-detection will be described with reference to FIG. If the correction target is not detected in the current frame, balanced correction information is calculated from the entire image of the current frame (substep SS50). The well-balanced correction information here is the detection score S _n = 0 in the current frame and R _n calculated from the entire image.

  Next, past correction information is read (substep SS52). The past correction information read out here is also correction information Q, R, and P including the detection score S that is past correction target information.

Next, the combined image correction information P _n in the current frame is calculated (substep SS54). For the calculation, either linear combination or non-linear combination is used. In the calculation by linear combination, the image correction information _Pn is multiplied by the weighting factor W corresponding to the detection score and the fixed value C by the information P and R. The linear combination in this case is shown in Equation (3). Further, as described above, the weighting factor W is preferably set to a larger value as the detection score is larger, and is preferably set to a larger value as it is closer to the current frame in time series.

The image correction information P _n in the current frame is expressed by equation (3)
P _n = C (0) R _n + W _n-1 (S _n-1 ) P _n-1 + ... + W _nk (S _nk ) P _nk ... (3)
Is calculated by

Further, the image correction information P _n in the current frame in the non-detection is not limited to the equation (3), but may be performed by the equation (4) representing nonlinear coupling. Equation (4) is
P _n = f _ND (R _n , P _n-1 , P _n-2 , ..., P _nk ) (4)
It is represented by Here, f _ND is a multidimensional nonlinear function in non-detection of an object. After calculating in this way, the process proceeds to return and the subroutine SUB5 is terminated.

In this way, the correction information calculation function unit 44 calculates the image correction information P _n in the current frame corresponding to detection or non-detection of the correction target. The calculated image correction information P _n is supplied to the CPU 16 and the digital processing unit 40. The CPU 16 also calculates the lens aperture and shutter speed for the next frame based on the calculated image correction information _Pn . The CPU 16 may generate the control signal 58 with the lens aperture and shutter speed calculated here.

Next, using the calculated image correction information P _n (50), the image data 36 in the current frame is corrected (subroutine SUB3). In this correction, as shown in FIG. 9, the current frame image of the image data 36 is color-corrected with image correction information P _n (50) (substep SS30).

  Since correction information is calculated so that correction for the image data 36 sampled in time series is performed by an algorithm according to the presence or absence of a correction target, and tone correction and color correction are performed with this correction information, time series An appropriate moving image can be obtained without abrupt fluctuation of the moving image depending on whether or not each of the frame images has a correction target. The present invention can be applied not only to moving images but also to actual imaging of still images.

  Needless to say, the above-described moving image processing procedures and component operations can be realized by software using a program.

1 is a block diagram illustrating a schematic configuration of an imaging apparatus in an embodiment to which an image processing apparatus of the present invention is applied. It is a block diagram which shows the schematic structure in the correction target detection part of FIG. It is a block diagram which shows the schematic structure in the digital processing part of FIG. 2 is a flowchart for explaining an operation in the image processing apparatus of FIG. 1. 6 is a flowchart in the object detection routine of FIG. 5 is a flowchart for calculating correction information in a current frame image based on the correction target in FIG. 4. It is a flowchart explaining the calculation procedure of the correction information based on the detection of the correction object of FIG. It is a flowchart explaining the calculation procedure of the correction information based on the non-detection of the correction target of FIG. 5 is a flowchart in the color correction routine of FIG.

Explanation of symbols

10 Imaging device
12 Image input section
14 Signal processor
16 CPU
18 memory
20 drivers
22 Memory IF section
24 Display controller
26 Memory card
28 Monitor
30 Operation unit

Claims (24)

In an image processing apparatus that performs image processing on a current frame image of a digital moving image, the apparatus includes:
Correction information storage means for storing first correction information relating to a correction target of the current frame image obtained from a past frame image corresponding to the current frame image;
Correction target detection means for detecting the correction target included in the current frame image and obtaining correction target information for the correction target;
Correction information calculating means for calculating second correction information for correcting the current frame image based on the correction target information and the first correction information in the current frame image obtained by the correction target detecting means;
And an image correction unit for correcting the current frame image based on second correction information. The apparatus according to claim 1, wherein the correction information calculation unit is configured to calculate temporary correction information for the current frame based on the correction target information in the current frame image;
An image processing apparatus comprising: second correction information calculating means for calculating second correction information for correcting the current frame image based on the first correction information and the provisional correction information. The apparatus according to claim 1, wherein the correction target detection unit reads out correction target information in the past frame image for the first correction information as past information;
Based on the past information, an area calculating means for calculating a detectable area of the correction target in the current frame image;
An image processing apparatus comprising: an information calculation unit that detects a correction target of the current frame image from the detectable region and calculates correction target information for the correction target.   The image processing apparatus according to claim 1, wherein the correction target includes at least one of a face, an eye, and a skin color region.   5. The apparatus according to claim 1, wherein the correction target information is at least one of information related to a face, eyes, and a skin color region, and an evaluation value that indicates a probability of detection of the correction target. An image processing apparatus comprising:   6. The apparatus according to claim 1, wherein the correction information calculation unit is configured to determine the temporary information for the current frame image according to a result of whether or not the correction target is detected in the current frame image. An image processing apparatus that calculates correction information.   7. The apparatus according to claim 1, wherein the first and second image correction information includes at least one of a gradation correction parameter, a color correction parameter, a lens diaphragm, and a shutter speed. An image processing apparatus.   8. The apparatus according to claim 1, wherein the apparatus realizes the functions of the correction target detection unit, the correction information calculation unit, and the image correction unit by a program. apparatus. In an image processing method for performing image processing on a current frame image of a digital moving image, the method includes:
A first step of reading first correction information from correction information storage means for storing first correction information relating to a correction target of the current frame image obtained from a past frame image corresponding to the current frame image;
A second step of detecting the correction target included in the current frame image and obtaining correction target information for the correction target;
A third step of calculating second correction information for correcting the current frame image based on the correction target information and the first correction information in the current frame image;
And a fourth step of correcting the current frame image based on second correction information. The method according to claim 9, wherein the second step is a fifth step of reading out correction target information in the past frame image for the first correction information as past information;
A sixth step of calculating a detectable region of the correction target in the current frame image based on the past information;
A seventh step of detecting a correction target of the current frame image from the detectable region and calculating correction target information for the correction target. The method according to claim 9 or 10, wherein the third step includes the step of changing the current frame based on the correction target information in the current frame image when the processing result in the second step indicates the presence of the correction target. An eighth step of calculating provisional correction information for the image;
And a ninth step of calculating second correction information for correcting the current frame image based on the first correction information and the provisional correction information. The method according to claim 9 or 10, wherein the third step is a provisional process for the current frame image based on the entire current frame image when the processing result in the second step does not indicate the presence of the correction target. A tenth step of calculating correction information of
And an eleventh step of calculating second correction information for correcting the current frame image based on the first correction information and the provisional correction information.   13. The image processing method according to claim 9, wherein the correction target includes at least one of a face, an eye, and a skin color region as the correction target.   The method according to any one of claims 9 to 13, wherein the method includes, as the correction target information, information relating to a face, eyes, and a skin color area, and an evaluation value indicating a probability in detecting the correction target. An image processing method comprising at least one of the following. 15. The method according to claim 14, wherein in the third step, the correction target information is an evaluation value indicating a probability in the detection of the correction target,
When the processing result in the second step indicates the presence of the correction target in the current frame image, the provisional correction information in this case and the multiplication result of the weighting factor based on the evaluation value, the first correction information, An image processing method characterized by linearly combining a multiplication result of weighting factors based on the evaluation value and calculating second correction information based on the linear combination. 15. The method according to claim 14, wherein in the third step, the correction target information is an evaluation value indicating a probability in the detection of the correction target,
If the processing result in the second step does not indicate the presence of the correction target in the current frame image, the provisional correction information and the multiplication result of the fixed weighting factor in that case, the first correction information, and the evaluation An image processing method characterized by linearly combining a multiplication result of weighting factors based on values and calculating second correction information based on the linear combination. 15. The method according to claim 14, wherein in the third step, the correction target information is an evaluation value indicating the probability of detection of the correction target,
When the processing result in the second step indicates the presence of the correction target in the current frame image, the provisional correction information, the evaluation value for the current frame image, and the first correction information in that case are respectively used as dependent variables. An image processing method characterized in that the second correction information is calculated by a multidimensional nonlinear function. 15. The method according to claim 14, wherein in the third step, the correction target information is an evaluation value indicating a probability in the detection of the correction target,
When the processing result in the second step does not indicate the presence of the correction target in the current frame image, a multidimensional nonlinear function using each of the temporary correction information and the first correction information in that case as dependent variables An image processing method comprising calculating the current image correction information.   The method according to any one of claims 9 to 18, wherein the first and second correction information includes at least one of a gradation correction parameter, a color correction parameter lens aperture, and a shutter speed. A featured image processing method.   The method according to any one of claims 11 to 18, wherein the method selects and selects one of the correction targets detected in a second step as the temporary correction information. An image processing method comprising calculating the provisional correction information for a correction target.   The method according to any one of claims 11 to 18, wherein the method selects two or more of the correction targets detected in a second step as the temporary correction information, and selects the selected correction target. An image processing method characterized in that the provisional correction information is calculated for the corrected object.   The method according to claim 20 or 21, wherein the method selects the correction target based on at least one of brightness, size, position, and evaluation value of the correction target. Processing method.   23. The image processing method according to claim 9, wherein the method is also used in actual imaging of a still image. Imaging means for imaging a scene to form a video signal of a digital moving image;
Correction information storage means for receiving the video signal and storing first correction information relating to a correction target of the current frame image obtained from a past frame image corresponding to the current frame image of the digital moving image;
Correction target detection means for detecting the correction target included in the current frame image and obtaining correction target information for the correction target;
Correction information calculating means for calculating second correction information for correcting the current frame image based on the correction target information and the first correction information in the current frame image obtained by the correction target detecting means;
An image correction means for correcting the current frame image based on second correction information, thereby performing image processing on the current frame image.

Images