JP2006018739A - Object detection means, device, and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To solve the problem that the color balance of a photographed image collapses according to photographic conditions or the status of a scene when photographing. <P>SOLUTION: This object detecting means is provided with an analyzing means for analyzing an input image, a setting means for setting a correction parameter according to the analysis result, an image correcting means for correcting the input image by using the set parameter and a detection means for detecting a preset object by using the corrected image. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a means, an apparatus and a method for detecting an object.

In the object detection process for detecting a human face or the like in a photographic image, the process is conventionally performed on the assumption that the image is captured in an appropriate state. Then, detection processing is performed with reference to luminance information, color information, edge information, and the like therein.
Japanese Patent Laid-Open No. 10-232930

  However, the color balance of the captured image may be disturbed depending on the shooting conditions and the state of the scene at the time of shooting, such as a photo taken under an indoor fluorescent light or a photo taken outdoors in the shade. However, since the photo includes dark noise in the image, there is a problem that the detection rate is remarkably lowered depending on the state of the image.

  In the present invention, in order to solve the problem, an analysis unit that analyzes an input image, a setting unit that sets a correction parameter according to the analysis result, and the input using the set correction parameter It is an object of the present invention to provide an object detection means comprising image correction means for correcting an image and detection means for detecting a preset object using the corrected image.

  The preset object is a human face.

  According to the present invention, for an image in which the color balance of the photographed image is lost depending on the photographing condition and the state of the scene at the time of photographing, the image is corrected to an appropriate color balance, and detection is performed after reducing dark noise in the image. Therefore, it is possible to always detect a human face with a good detection rate.

  FIG. 1 shows a configuration of a main part of an image processing apparatus according to an embodiment of the present application.

  An image processing apparatus shown in FIG. 1 includes an image input unit (2), an image output unit (3), a histogram creation unit (4), a histogram holding unit (5), a lookup table creation unit (6), and a lookup table. From the holding unit (7), the image correction unit (8), the image buffer (9), the image buffer for copying (10), the filter processing unit (11), the face detection processing unit (12), and the photographing information holding unit (13) Composed.

  The image input unit (2) reads data from the input image (1) and writes it into the image buffer (9).

  The image output unit (3) writes the data stored in the image buffer (9) to the output image (14).

  The histogram creation unit (4) creates a histogram based on the image data stored in the image buffer (9), and stores the result in the histogram holding unit (5).

  The histogram holding unit (5) holds a histogram of the entire image data and a histogram of the mask area.

  The lookup table creation unit (6) calculates parameters necessary for correction based on the histogram stored in the histogram holding unit (5), and creates a lookup table. Further, it is combined with the look-up table for correction of gradation characteristics and color reproducibility held in advance in the look-up table holding unit (7), and the combined look-up table is stored in the look-up table holding unit (7).

  The look-up table holding unit (7) holds the look-up table created by the look-up table creating unit (6) and the gradation characteristic and color reproducibility correction look-up table.

  The image correction unit (8) corrects the image stored in the image buffer (9) based on the lookup table stored in the lookup table holding unit (7), and the result is copied to the image buffer for copying ( 10).

  The image buffer (9) and the copy image buffer (10) hold image data.

  The filter processing unit (11) performs enlargement / reduction processing and edge enhancement of the highlight portion.

  The face detection processing unit (12) detects whether a human face exists based on the image data stored in the image buffer (9). If it exists, mask data is created.

  The shooting information holding unit (13) holds the camera orientation, flash use information, focus distance, focus position (two-dimensional), and the like at the time of shooting.

  The image processing apparatus can be realized by supplying a program for realizing processing or control of each unit in FIG. In that case, the CPU of the personal computer performs each process or control based on the contents of the program.

  Hereinafter, embodiments of the invention will be described in detail using specific examples.

<Processing flow of this embodiment>
FIG. 2 shows the flow of processing in this embodiment.

  In step S1, the image input unit (2) reads the shooting information and image data of the input image (1) and develops them in the image buffer (9). Details of the operation of the image input unit (2) are shown in FIG.

  In step S2, the image is corrected in the histogram creation unit (4), the lookup table creation unit (6), and the image correction unit (8). Details of the operation are shown in FIG.

  In step S3, the face detection processing unit (12) performs face detection processing. Details of the operation of the face detection processing section (12) are shown in FIG.

  In step S4, it is determined whether a face exists as a result of the face detection process. If at least one face exists, the process proceeds to step S5. If no face exists, the process proceeds to step S6.

  In step S5, the histogram creation unit (4), the lookup table creation unit (6), and the image correction unit (8) correct the image using the mask data. Details of the operation are shown in FIG. 14 and will be described later.

  In step S6, the filter processing unit (11) performs filter processing. Details of the operation of the filter processing unit are shown in FIG.

  In step S7, the image output unit (3) outputs the image data and shooting information stored in the image buffer (9) to the output image (14).

<Shooting Information and Image Data Stored in Input Image (1)>
FIG. 20 shows shooting information and image data stored in the input image (1). The shooting information is stored in the shooting information holding unit (13), and the image data is stored in the image buffer (9). In addition to this, the image buffer (9) stores mask data.

  The shooting information includes the orientation of the camera at the time of shooting (FIG. 23 (a) is the normal position, and there are four types (a) to (d)), flash use information, focus distance, focus position, input device ID, and the like. Stored.

<Image expansion method>
Details of the operation in the image input unit (2) are shown in FIG. This shows step S1 in FIG. 2 in detail.

  In step S11, shooting information is input from the input image (1) and stored in the shooting information holding unit (13).

  In step S12, the camera orientation at the time of shooting stored in the shooting information holding unit (13) is determined, and the image data obtained from the input image (1) is developed in the image buffer (9) according to the determination result.

  When the camera is in the position shown in FIG. In FIG. 23B, the image is rotated 90 ° to the left, 90 ° to the right in FIG. 23C, and 180 ° in FIG. 23D. Expand so that the vertical direction coincides with the time of. By developing in this way, at the time of the subsequent face detection process, it can be processed that the person is always facing the normal position regardless of the orientation of the camera at the time of shooting.

  At the time of development, processing for reducing dark noise peculiar to a digital camera is also performed. As this process, for example, a process of smoothing a part where the luminance is slightly varied in a low luminance part can be considered.

<Image correction processing>
FIG. 4 shows the flow of image correction processing.

  In step S21, the histogram creation unit (4) creates a histogram based on the image data stored in the image buffer (9), and stores the result in the histogram holding unit (5). Details of the operation of the histogram creation unit (4) will be described later with reference to FIG.

  In step S22, the lookup table creation unit (6) creates a lookup table. Details of the operation of the lookup table creation unit (6) will be described later with reference to FIG.

  In step S23, the image correction unit (8) corrects the image. Details of the operation of the image correction unit (8) will be described later with reference to FIG.

<Operation in Histogram Creation Unit (4)>
The operation in the histogram creation unit (4) is shown in FIG. This shows step S21 in FIG. 4 in detail.

  In step S31, image data is extracted from the image buffer (9) one pixel at a time. The image data stores the luminance (R, C, B) of each RGB color.

In step S32, the luminance L of the pixel is obtained from the RGB values of the image data according to the following formula.
L = (3 * R + 6 * G + 1 * B) / 10
In step S33, the histogram stored in the histogram holding unit is updated. The histogram holding unit holds the histogram HistL of the calculated luminance L and HistR, HistG, and HistB that store the accumulated luminance value of each color of RGB for each luminance L of the pixel. The initial state is all zero. The histogram is updated according to the following formula.
HistR [L] = HistR [L] + R
HistG [L] = HistG [L] + G
HistB [L] = HistB [L] + B
HistL [L] = HistL [L] +1
In step S34, it is checked whether all pixels have been completed. If finished, exit. If not, the process proceeds to step S31.

  An example of the created histogram HistL is shown in FIG.

<Method for creating lookup table for tone reproduction and color reproducibility correction>
The look-up table for gradation reproduction and color reproducibility correction refers to the preferred color reproduction, and gradation reproduction and color of printed matter when an image input from an input device such as a digital camera is output by an output device such as a printer. It is a look-up table for correcting reproduction to an image that approaches gradation reproduction and color reproduction realized by an output target such as a silver salt positive film. A method of creating a lookup table for tone reproduction and color reproducibility correction will be described in detail with reference to FIG.

  In step S121, the output target profile shown in FIG. 9 (the relationship between the patch number and the XYZ colorimetric values) and the output device profile shown in FIG. 11 (the relationship between the XYZ colorimetric values and RGB luminance) are synthesized. The synthesizing method searches the output device profile for a patch having the same Y value as the Y value of the XYZ value (there is a value for each patch) of the output target profile. Then, the RGB values corresponding to the XYZ values of the patches found from the output device profile and the patch numbers of the original output target profile are taken and combined. The synthesized table has a relationship between the patch number and the luminance (RGB value) of the output device.

  In step S122, the table synthesized in step S121 and the input device profile shown in FIG. 10 are synthesized. Since the input device profile has a relationship between the patch number and the luminance (RGB value) of the input device, the composition method is the luminance of the input device (RGB value) and the luminance of the output device (RGB value) corresponding to the same patch number. Will collect the relationship. The created table is stored in the look-up table holding unit (7) and used when correcting an image. An actual use example will be described later.

<Example of lookup table for correcting gradation characteristics and color reproducibility>
The look-up table shown in FIG. 22 is a look-up table for correcting gradation characteristics and color reproducibility, LUTRDD, LUTGDD, and LUTBDD created by the above method. It should be noted that conversion data is created by interpolating with straight lines between points where no data is stored in the table. At both end points, interpolation is performed with straight lines (255, 255) and (0, 0), respectively. In this embodiment, linear interpolation is used as the interpolation processing. However, for example, non-linear interpolation processing using a spline curve or a Bezier curve may be used.

<Operation in Lookup Table Creation Unit (6)>
The operation in the lookup table creation unit (6) is shown in FIG. This shows step S22 of FIG. 4 in detail.

  In step S41, the maximum luminance of the image is obtained from the histogram HistL stored in the histogram holding unit (5). In the histogram shown in FIG.

  In step S42, a predetermined amount (in FIG. 6, (predetermined amount) = 10) is subtracted from the maximum luminance obtained in S41, and the luminance when the maximum luminance becomes larger (in FIG. LH ′, that is, in this example, 255, 245, 235,..., Are lowered and compared with the maximum luminance each time (in FIG. 6, LH ′ = 245). A region including a percentage of pixels (1% of the total number of pixels in FIG. 6) is obtained. The minimum luminance of the area is a highlight point (LH, LH = 234 in FIG. 6). Then, RGB average luminance (RH, GH, BH) in the region (region where the luminance is not less than LH and not more than LH ′) is calculated according to the following formula.

ステップＳ４３において、ヒストグラム保持部（５）に格納されているヒストグラムＨｉｓｔＬから、画像の最小輝度を求める。図４に示したヒストグラムにおいては、最小輝度は６になる。

In step S43, the minimum luminance of the image is obtained from the histogram HistL stored in the histogram holding unit (5). In the histogram shown in FIG.

  In step S44, the minimum luminance obtained in S43 and a predetermined amount from 0 ((predetermined amount) = 10 in FIG. 6) are added, and the luminance when the minimum luminance becomes smaller (in FIG. LS ′, LS ′ = 10 in FIG. 6 is obtained, and a region including a predetermined ratio of pixels (1% of the total number of pixels in FIG. 6) is obtained. The maximum luminance of the area is a shadow point (LS, LS = 22 in FIG. 6). Then, the average luminance (RS, GS, BS) of RGB in the region (region where the luminance is LS ′ or more and LS or less) is calculated according to the following formula.

ステップＳ４５において、求めたＲＨ、ＧＨ、ＢＨ、ＲＳ、ＧＳ、ＢＳからＲＧＢそれぞれのルックアップテーブルＬＵＴＲ、ＬＵＴＧ、ＬＵＴＢを作成する。作成したルックアップテーブルの例を図８（ａ）に示し、詳細を後述する。

In step S45, RGB lookup tables LUTR, LUTG, and LUTB are created from the obtained RH, GH, BH, RS, GS, and BS. An example of the created lookup table is shown in FIG. 8A and will be described in detail later.

  In step S46, the average luminance of the image is obtained from the histogram HistL stored in the histogram holding unit (5). Then, an exposure correction lookup table LUTL is created based on the obtained average luminance value. An example of the created lookup table is shown in FIG. 8B and will be described in detail later.

In step S47, the gradation characteristics and color reproducibility stored in advance in the lookup tables LUTR, LUTG, and LUTB created in step S45, the lookup table LUTL created in step S46, and the lookup table holding unit (7). The correction look-up tables LUTRDD, LUTGDD, and LUTBDD are synthesized according to the following formula. The synthesized lookup tables LUTR2, LUTG2, and LUTB2 are stored in the lookup table holding unit (7).
LUTR2 [i] = LUTRDD [LUTL [LUTR [i]]]
LUTG2 [i] = LUTGDD [LUTL [LUTG [i]]]
LUTB2 [i] = LUTBDD [LUTL [LUTB [i]]]
In the above formula, i is a value not less than 0 and not more than the maximum luminance.

<Example of created lookup table>
An example of the created lookup tables LUTR, LUTG, LUTB, and LUTL is shown in FIG. Look-up tables LUTR, LUTG, and LUTB shown in FIG. 8A are for correcting contrast and color cast. Here, the highlight gamma is set in the order of G, B, and R. In this way, by strengthening B and G with respect to R, it is possible to correct, for example, the fogging of a cyan fogging image. At the same time, the contrast can be corrected.

  On the other hand, the look-up table LUTL shown in FIG. 8B is a look-up table for exposure correction. Depending on the average brightness of the image, there is a possibility of underexposure when the average brightness is low. Create a table, and when the average brightness is low, there is a possibility of overexposure, so create a lookup table that darkens the whole.

<Operation in Image Correction Unit (8)>
FIG. 12 shows the operation of the image correction unit (8). This shows details of step S23 in FIG. 4 and step S73 in FIG.

  In step S51, one pixel of image data stored in the image buffer (9) is extracted. The image data stores the luminance (R, G, B) of each color of RGB.

In step S52, the image data extracted from the image buffer (9) is corrected based on the lookup tables LUTR2, LUTG2, and LUTB2 synthesized in step S46 or step S98. The result is stored in the copy image buffer (10).
R = LUTR2 [R]
G = LUTG2 [G]
B = LUTB2 [B]
In step S53, it is checked whether or not all pixels have been completed. End if all pixels are finished. If not completed, the process proceeds to step S51.

  In the present embodiment, a lookup table is prepared for each component in order to increase the speed of the lookup table creation process and the image correction process.

<Operation of Face Detection Processing Unit (12)>
13 and 24 show the operation of the face detection processing unit (12). This shows step S3 in FIG. 2 in detail.

  In step S61, an area having a low luminance is marked locally based on the image data stored in the copy image buffer (10) (that is, the image data corrected in step S2).

  In step S62, two regions marked in step S61 are grouped, and it is determined whether the two groups are eyes from the uniformity of the size of the regions, the difference in luminance, the angle with respect to the horizontal, etc. (FIG. 24 (a)).

  In step S63, for the group determined to be an eye, a rectangular area centering on the two is set. Then, the edge near the boundary of the region and the color information in the region are determined. The edge determines whether there is a face outline. Further, the color information determines whether or not the average value of the RGB values in the rectangular area is within a preset skin color area. Then, the face area is determined. The face area is given by the rectangular area (FIGS. 24B and 24C).

  In step S64, mask data for masking the determined face area is created, and the created mask data is stored in the image buffer (9).

<Image correction processing using mask data>
FIG. 14 shows the flow of image correction processing when mask data in this embodiment is used.

  In step S71, the histogram creation unit (4) creates a histogram based on the image data stored in the image buffer (9) (image data before being corrected in step S2), and the result is stored in the histogram holding unit. Store in (5). Details of the operation of the histogram creation unit (4) will be described later with reference to FIG.

  In step S72, the lookup table creation unit (6) creates a lookup table. Details of the operation of the lookup table creation unit (6) will be described later with reference to FIG.

  In step S73, the image correction unit (8) corrects the image. The details of the operation of the image correction unit (8) are the same as those shown in FIG.

<Histogram creation process using mask data>
The operation in the histogram creation unit (4) is shown in FIG. This shows step S71 of FIG. 14 in detail.

  In step S81, image data and mask data are extracted from the image buffer (9) one pixel at a time. The image data stores the luminance (R, G, B) of each RGB color, and the mask data stores the presence or absence of a mask.

In step S82, the luminance L of the pixel is obtained from the RGB values of the image data according to the following formula.
L = (3 * R + 6 * G + 1 * B) / 10
In step S33, the histogram stored in the histogram holding unit is updated. The histogram holding unit stores the histogram HistL of the calculated luminance L and the accumulated luminance value of each color of RGB for each luminance L of the pixel. HistR, HistG, HistB, and a histogram HistLMsk of the luminance L of the area masked by the mask data are held. The initial state is all zero. The histogram is updated according to the following formula.
HistR [L] = HistR [L] + R
HistG [L] = HistG [L] + G
HistB [L] = HistB [L] + B
HistL [L] = HistL [L] +1
In step S84, it is checked whether or not the pixel is masked from the mask data. If masked, the process proceeds to step S85, and if not masked, the process proceeds to step S86.

In step S85, the histogram of the mask area stored in the histogram holding unit is updated.
HistLMsk [L] = HistLMsk [L] +1
In step S86, it is checked whether or not all pixels have been completed. If finished, exit. If not, the process proceeds to step S81.

  An example of the created histogram HistL is the same as that shown in FIG.

<Operation in lookup table creation unit (6) when mask data is used>
FIG. 16 shows the operation of the lookup table creation unit (6) when mask data is used. This shows step S72 in FIG. 14 in detail.

  In step S91, the maximum luminance of the image is obtained from the histogram HistL stored in the histogram holding unit (5). In the histogram shown in FIG.

  In step S92, the maximum brightness obtained in S91 and a predetermined amount (in FIG. 6, (predetermined amount) = 10) is subtracted from 255, and the brightness when the maximum brightness is larger (in FIG. LH ′, that is, in this example, 255, 245, 235,..., Are lowered and compared with the maximum luminance each time (in FIG. 6, LH ′ = 245). A region including a percentage of pixels (1% of the total number of pixels in FIG. 6) is obtained. The minimum luminance of the area is a highlight point (LH, LH = 234 in FIG. 6).

  In step S93, the minimum luminance of the image is obtained from the histogram HistL stored in the histogram holding unit (5). In the histogram shown in FIG.

  In step S94, the minimum luminance obtained in S93 is added to a predetermined amount from 0 ((predetermined amount) = 10 in FIG. 6), and the luminance (when the minimum luminance becomes smaller) ( LS ′, LS ′ = 10 in FIG. 6 is obtained, and a region including a predetermined ratio of pixels (1% of the total number of pixels in FIG. 6) is obtained. The maximum luminance of the area is a shadow point (LS, LS = 22 in FIG. 6).

In step S95, the average values (RG, GG, BG) of the respective RGB colors of the pixels satisfying the following expression are obtained for the pixels having the luminance in the range from LS to LH.
Max (R, G, B) −Min (R, G, B) = <Thr
However, Max (R, G, B) is the maximum value among RGB values, Min (R, G, B) is the minimum value among RGB values, and Thr is a certain threshold value.

  In step S96, lookup tables LUTR, LUTG, and LUTB are created from the obtained LS, LH, RG, GG, and BB. An example of the created lookup table is shown in FIG. 17A and will be described in detail later. Then, the result is stored in the lookup table holding unit (7). Also, a lookup table LUTTmp for HistL correction is created based on LH and LS.

  In step S97, HistLMsk is corrected using LUTTmp. This is used later when calculating the exposure correction amount.

  In step S98, the average luminance of the mask area is obtained from the corrected HistLMsk, and an exposure correction lookup table LUTLMsk is created according to a predetermined method. Then, the result is stored in the lookup table holding unit (7). An example of LUTLMsk is shown in FIG. 10B and will be described in detail later.

In step S99, the lookup tables LUTRDD, LUTG, LUTB, and LUTLMsk created in step S96, and the gradation characteristics and color reproducibility correction lookup table LUTRDD stored in advance in the lookup table holding unit (7), LUTGDD and LUTBDD are synthesized according to the following formula. The synthesized lookup tables LUTR2, LUTG2, and LUTB2 are stored in the lookup table holding unit (7).
LUTR2 [i] = LUTRDD [LUTLMsk [LUTR [i]]]
LUTG2 [i] = LUTGDD [LUTLMsk [LUTG [i]]]
LUTB2 [i] = LUTBDD [LUTLMsk [LUTB [i]]]
In the above formula, i is a value not less than 0 and not more than the maximum luminance.

<Example of created lookup table>
An example of the created lookup table is shown in FIG. The lookup tables LUTR, LUTG, and LUTB shown in FIG. 17A are for correcting contrast and color cast. Here, halftone gamma is set in the order of G, B, and R. In this way, by strengthening G and B with respect to R, it is possible to correct the color cast of an image that is, for example, cyan. At the same time, the contrast can be corrected.

  On the other hand, the lookup table shown in FIG. 17B is a lookup table LUTLMsk for optimally correcting the exposure of the mask area. Here, for example, a person area may be set as a mask area, and LUTLMsk may be set so that the exposure of the person can be optimally corrected.

<Operation of Filter Processing Unit (11)>
FIG. 18 shows a flow of the operation of the filter processing unit (11). This shows step S6 in FIG. 2 in detail.

  In step S101, image enlargement / reduction processing is performed. For processing, a method such as bicubic or bilinear is used.

  In step S102, edge enhancement processing is performed. Details of the edge enhancement processing are shown in FIG. 19 and will be described later.

<Edge enhancement processing>
FIG. 19 shows an operation flow of edge enhancement processing. This shows step S102 in FIG. 18 in detail.

  First, in step S111, RGB luminance data for one pixel is read from the copy image buffer (10).

  In step S112, the read RGB luminance data and the threshold are compared for each of RGB. As a result, if the RGB luminance data are all smaller than the threshold value or are pixels at the end of the image, the process proceeds to step S114. In other cases, the process proceeds to step S113.

  In step S113, the color component data having a value equal to or greater than the threshold value is subjected to filter processing using a 3 × 3 edge enhancement filter. Note that filter processing is not performed on color component data that is equal to or less than the threshold value.

  In step S114, the processed RGB luminance data is written into the image buffer (9). When the edge enhancement processing is not performed, the RGB luminance data read in step S111 is written as it is in the image buffer (9).

  In step S115, it is determined whether all pixels have been completed. If all the pixels have not been completed, the process proceeds to step S111.

  Note that the threshold value used in step S112 may be the same threshold value for each color component, or may be set independently for each color component.

(Second Embodiment)
In the embodiment of the present invention, when creating the lookup table, the corrected highlight point and shadow point (HP and SP in FIG. 8A) always use certain constants. Using the information stored in the holding unit (13), the corrected highlight point and shadow point can be made variable. The embodiment will be described in detail below.

  FIG. 25 shows the configuration of the main part of the image correction apparatus according to the second embodiment of the invention.

  The image processing apparatus shown in FIG. 25 includes an image input unit (2), an image output unit (3), a histogram creation unit (4), a histogram holding unit (5), a lookup table creation unit (6), and a lookup table. Holding unit (7), image correcting unit (8), image buffer (9), image buffer for copying (10), filter processing unit (11), face detection processing unit (12), photographing information holding unit (13), It is comprised from a correction information holding part (15).

  Among them, the operations other than the look-up table creation unit (6) and the correction information holding unit (15) perform the same operations as those described in the above embodiment, and thus the description thereof is omitted.

<Lookup Table Creation Method in Second Embodiment>
FIG. 26 shows information held in the correction information holding unit (15).

  In FIG. 26, the presence / absence of flash use and the range of the focus distance are combined, and the corrected highlight point and shadow point are held according to each combination. For example, when using a flash and the focus distance is short, there is a strong possibility that the highlight is flying, so lower the HP value so that the contrast is not emphasized too much, or the focus distance is not used when the flash is not used. Since the possibility of a landscape photograph is strong when it is long, it is possible to enhance the contrast by lowering the SP value and increasing the HP value. The HP and SP are used by being read from the correction information holding unit (15) when the lookup table creation unit (6) creates the lookup table (steps S45 and S96).

(Other embodiments)
<Luminance concept>
In the above embodiment, the luminance data is implemented with digital values from 0 to 255. However, other than this, for example, the maximum value of the data is not limited to 255, and not only the luminance but also halftone dot density, etc. Needless to say, it is good.

<Calculation method of luminance value>
In step S32 and step S82, the luminance value is calculated by weighted averaging with a weight of R: G: B = 3: 6: 1, but needless to say, the calculation is performed with other weights, for example. Alternatively, the average value of the maximum value and the minimum value of RGB may be obtained.

<How to show face detection results>
In the above-described embodiment, the face detection result is indicated by a rectangle. However, it is needless to say that the face detection result is not limited to a rectangle.

<Camera orientation during shooting and image data stored in image buffer>
In FIG. 23, four cases are described with respect to the direction of the camera at the time of shooting and the image data stored in the image buffer. However, it is needless to say that the four types are not necessarily limited to, for example, tilted by 45 °. When the image is taken, the image may be stored in the image buffer by being inclined 45 ° in the reverse direction.

<How to use shooting information>
In the above embodiment, only the position of the camera is used as the shooting information. However, as shown in FIG. 20, the output target profile shown in FIG. 9 is switched using the camera ID. Good. In addition, by using the in-focus position, the exposure correction lookup table LUTL may be created around the in-focus position.

<About information held in the correction information holding unit (15)>
In the second embodiment, the correction information holding unit (15) holds the corrected highlight point (HP) and shadow point (SP), but it goes without saying that it is not limited to HP and SP. Yes. For example, the exposure correction method is also maintained, and when the flash is not used and the focus distance is long, there is a strong possibility of underexposure. Therefore, the exposure correction may be set brighter than usual.

<Others>
The present invention may be applied to a system composed of a plurality of devices (for example, a host computer, an interface device, a reader, a printer, etc.) or an apparatus composed of a single device (for example, a copying machine, a facsimile machine).

  In addition, program code of software for realizing the functions of the embodiment is supplied to an apparatus or a computer in the system connected to the various devices so as to operate the various devices so as to realize the functions of the above-described embodiments. However, the present invention also includes those implemented by operating the various devices according to a program stored in a computer (CPU or MPU) of the system or apparatus.

  In this case, the program code of the software itself realizes the functions of the above-described embodiments, and the program code itself and means for supplying the program code to the computer, for example, a memory storing the program code The medium constitutes the present invention.

  As a storage medium for storing the program code, for example, a floppy (registered trademark) disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a magnetic tape, a nonvolatile memory card, a ROM, or the like can be used.

  Further, by executing the program code supplied by the computer, not only the functions of the above-described embodiments are realized, but also the OS (operating system) in which the program code is running on the computer, or other application software, etc. It goes without saying that the program code is also included in the embodiment of the present invention even when the functions of the above-described embodiment are realized in cooperation with the embodiment.

  Further, after the supplied program code is stored in the memory provided in the function expansion board of the computer or the function expansion unit connected to the computer, the CPU provided in the function expansion board or function storage unit based on the instruction of the program code Needless to say, the present invention includes a case where the functions of the above-described embodiment are realized by performing part or all of the actual processing.

The figure which shows the structure of the image processing apparatus of this-application embodiment. The figure which shows the flow of a process in this embodiment. The figure which shows the operation | movement in an image input part (2). The figure which shows the flow of image correction. The figure which shows the operation | movement in a histogram preparation part (4). The figure which shows the example of a histogram. The figure which shows operation | movement in a look-up table preparation part (6). The figure which shows the example of the lookup table currently hold | maintained by the lookup table holding | maintenance part (7). Example output target profile. An example of an input device profile. An example of an output device profile. The figure which shows operation | movement in an image correction part (8). The figure which shows operation | movement in a face detection process part (12). The figure which shows the flow of an image correction process in the case of using mask data. The figure which shows the operation | movement in a histogram process part (4) in the case of using mask data. The figure which shows operation | movement in the lookup table preparation part (6) in the case of using mask data. The figure which shows the example of the lookup table hold | maintained by the lookup table holding | maintenance part (7) in the case of using mask data. The figure which shows the operation | movement in a filter process part (11). The figure which shows the flow of an edge emphasis process. The figure which shows the example of the imaging | photography information and image data which are stored in the image buffer (9). The figure which shows the synthetic | combination method of the look-up table for gradation reproduction and color reproducibility correction | amendment. The figure which shows the example of the look-up table for the synthesized gradation reproduction and color reproducibility correction | amendment. The figure which shows the relationship between the direction at the time of imaging | photography of a digital camera, and the expansion | deployment method to an image buffer (9). The figure which shows the flow of face detection. The figure which shows the structure of the image processing apparatus of 2nd Embodiment of this application. The figure which shows the information hold | maintained in a correction information holding part (15).

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Input image 2 Image input part 3 Image output part 4 Histogram creation part 5 Histogram holding part 6 Lookup table creation part 7 Lookup table holding part 8 Image correction part 9 Image buffer 10 Copy image buffer 11 Filter processing part 12 Face detection Processing unit 13 Shooting information holding unit 14 Output image 15 Correction information holding unit

Claims (8)

  Analysis means for analyzing an input image, setting means for setting a correction parameter according to the analysis result, image correction means for correcting the input image using the set correction parameter, and the corrected image Object detection means comprising: detection means for detecting an object set in advance using the object detection means. The analysis means creates a histogram of the input image,
The setting means calculates the highlight, shadow, exposure state of the image from the created histogram, and creates a correction lookup table based on the calculation result and the detection result,
The object detection unit according to claim 1, wherein the image correction unit corrects the input image using the created correction lookup table.   2. The image correction according to claim 1, wherein the detection unit detects a position and a size of the preset target, and creates a mask image that masks the detected preset target portion. 3. means.   The image correction means according to claim 3, wherein the mask has a rectangular shape.   The image correction means according to claim 4, wherein the preset object is a human face.   The image correction means according to claim 5, wherein the mask has a shape of a human face.   An analysis unit that analyzes an input image, a setting unit that sets a correction parameter according to the analysis result, an image correction unit that corrects the input image using the set correction parameter, and the corrected image An object detection apparatus comprising: a detection unit that detects an object set in advance by using the detection unit.   An analysis method for analyzing an input image, a setting method for setting a correction parameter according to the analysis result, an image correction method for correcting the input image using the set correction parameter, and the corrected image And a detection method for detecting an object set in advance using the target detection method.

Images