JP2005244536A - Image composition for generating composite image by overlapping image - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce wrong feeling in a composite image in which images are overlapped. <P>SOLUTION: Dot-matrix-like pixel image data for composing a frame image are read, a see-through region in the frame image is determined, and the image region of an original image to be composed corresponding to the perspective region is determined as a region to be seen through. The image data of the original image to be composed included in the region to be seen through determined in such manner are sampled, the image data are processed according to the result (histogram), and the contrast of the image is for example adjusted. The new image of the target to be composed is generated, based on the processed image data, and the new image is composed while overlapping below the frame image. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a technique for generating a composite image by superimposing images.

  In recent years, with the spread of digital still cameras and the like, images can be easily obtained as digital data. Because of this situation, a technique for adjusting the color shift and color balance of an image when displaying or printing the image has been proposed (for example, Patent Document 1).

JP-A-10-271524

  In addition, a technique for superimposing a photographed image on a so-called frame has been proposed (for example, Patent Document 2).

JP 2003-233366 A

  When shooting an image, there is little consideration for subsequent synthesis with a frame. Therefore, when a photographed image is combined with a frame after adjusting its color shift and color balance, the composite image may be uncomfortable. For example, although adjustments such as color balance are performed on the entire image, only a partial region of the adjusted image can be visually recognized in the composite image with the frame. For this reason, when the balance adjustment performed with great effort is limited to the visual recognition area of the image by the frame, the balance may be unbalanced.

  The present invention has been made to solve the above-described problems, and an object thereof is to reduce a sense of incongruity in a composite image in which images are superimposed.

  In order to solve at least a part of the problem, in the image generation device of the present invention, both the first image to be superimposed and the second image having the fluoroscopic region of the first image are included in the second image. Prior to the synthesis that overlaps one image, the first image located below is processed. In this processing, after determining the corresponding region of the first image so as to correspond to the fluoroscopic region of the second image with a predetermined relationship, the first image data constituting the first image is converted into the first image data in the determined corresponding region. Adjust based on the properties of one image. For this reason, the image adjustment for the first image that is visible only in the perspective area of the second image is based on the nature of the image in the first image in the visible area, that is, in the corresponding area. Even if the visual recognition area where the image is visually recognized is limited by the second image, the uncomfortable feeling of the first image after the image adjustment can be reduced.

  In such an image generation apparatus of the present invention, the image adjustment of the first image is executed by correcting the first image data constituting the first image, and at the time of correction of the first image data during such image adjustment. Is the corresponding area of the first image. It is preferable to adjust the first image by correcting the first image data based on the statistical characteristics obtained from the first image data in the sampling region.

  In this way, since the property of the image (first image) in the corresponding area can be grasped more accurately by the first image data, the image adjusted first when the visual recognition area of the first image is limited by the second image. The uncomfortable feeling of the image can be reduced more reliably.

  Further, the image data correction of the first image based on the image data situation of the sampling area is performed on the image data including at least the first image data of the fluoroscopic area among the first image data constituting the first image. It is preferable to do so. In this case, image data correction may be performed on all the first image data constituting the first image, and among the first image data constituting the first image, the first image data of the fluoroscopic region (that is, the first image data) Image data correction performed on sampling area data) may be performed. With the latter correction method, the processing load can be reduced because there is little data to be corrected.

  The image generation apparatus of the present invention described above further includes a storage unit that stores a plurality of images with different perspective areas, and an image selection unit that selects one of the stored images as a second image. You can also In this way, even when the second image having a different perspective area is combined with the first image, the uncomfortable feeling can be reduced and the application range can be expanded. Moreover, it is more preferable if the storage unit can input an image from the outside.

  In addition, the second image data constituting the second image can include information for determining a corresponding area for adjusting the first image, and the second image data includes the first image adjustment in the first image adjustment. It is also possible to include correction participation data relating to the contents of correction performed on the image data, and to correct the first image data by reflecting this correction participation data to adjust the first image. In this way, the image adjustment becomes simple.

  Further, with respect to a second image area including at least a fluoroscopic area surrounding image portion around the fluoroscopic area of the second image, the second image area is set as the second image sampling area, and the second image area includes the second image sampling area. The statistical characteristics of the image obtained from the image data can be obtained by the second image sampling unit, and the obtained statistical characteristics can be reflected in the correction of the first image data at night by the adjustment unit. By so doing, the nature of the image of the second image region including at least the fluoroscopic region peripheral image portion in the second image is reflected in the first image data adjustment, so that the uncomfortable feeling of the first image can be further alleviated.

  In addition, in order to solve at least a part of the above problems, the present invention may be configured as follows. That is, at the time of image composition in which a plurality of images are superimposed, the first image is adjusted, and the processed first image is superimposed on the second image to be superimposed, and both images are synthesized. At this time, the second image includes an overlapping region defined so that the image portion of the first image can be seen by overlapping the image portion of the partial region of the first image on the second image, A process for cutting out the image portion of the first image corresponding to the overlapping area and an adjustment process for the first image based on the property of the first image in the area to be cut out are executed, and the image adjustment and cut-out have been performed. The image portion of the partial area of the first image is superimposed on the overlapping area of the second image and synthesized.

  That is, in the present invention described above, the second image having the fluoroscopic region of the first image is superimposed on the first image, but the present invention is the opposite idea, and the image of the first image A first image that has been subjected to the same image adjustment as described above is applied to a second image having a superimposed region that is defined so that the image portion of the first image can be seen by overlapping the portion. Cut and stack. Even in this case, the same effect as the previous invention can be obtained.

  The above-described image composition of the present invention can be realized in various forms. For example, in addition to the image composition method, a computer program for causing a computer to realize the functions of the image composition apparatus and method, and the program are recorded. It can be realized in the form of a recording medium or the like.

Hereinafter, embodiments of the present invention will be described in the following order based on examples.
A. First embodiment:
A1. Configuration of image processing device:
A2. Image processing:
B. Image processing of the second embodiment:
C. Image processing of the third embodiment:
D. Variation:

A. First embodiment:
A1. Configuration of image processing device:
FIG. 1 is an explanatory diagram showing an image processing system 100 as a first embodiment of the present invention. As shown in the figure, this image processing system 100 includes an image database 20 that supplies image data such as moving images and still images, and a personal computer as an image processing apparatus that executes image processing on images input from the image database 20. 30, a user interface 40 for a user to instruct execution of image processing, a color printer 50 for outputting an image subjected to image processing, and the like.

  The image database 20 includes devices that handle images such as a digital video camera 21, a digital still camera 22, a DVD 23, a hard disk 24, and a mobile phone 25, and supplies image data to the personal computer 30. Note that the image data held in the image database 20 of the first embodiment is still image data acquired by the digital still camera 22. The personal computer 30 is configured to output a composite image, which will be described later, to the digital still camera 22 or the mobile phone 25 of the image database 20.

  The personal computer 30 includes a CPU 31 for executing image processing, a ROM 32, a RAM 33, a hard disk 34 for installing image processing software, an image database 20, an I / F circuit for exchanging with external devices such as a user interface 40 and a color printer 50. 35 etc. The image processing of the software installed in the hard disk 34 is a process of generating a single high-resolution still image by combining a plurality of input still images, specifically two still images. The personal computer 30 installed with this software has the functions of “image processing unit”, “compositing unit”, “region determining unit”, and “adjusting unit” as an image processing apparatus. The flow of this image processing will be described in detail later.

  The user interface 40 includes a keyboard 41 and a mouse 42 for a user to perform an image processing execution operation, a display 43 that displays an image before image processing and a still image after composition processing, and the like.

A2. Image processing:
FIG. 2 is a flowchart of image processing according to the first embodiment for generating a single still image by combining the first image and the second image to be superimposed, and FIG. 3 illustrates the contents of the image processing according to the first embodiment. It is explanatory drawing for doing. The first image is an image that is superimposed so as to be below the second image, and is referred to as a composition target original image in the following description. The second image is an image having a fluoroscopic region that allows a partial region of the first image (composition target original image) to be seen through, and is referred to as a frame image. In the image processing system 100 having the above-described hardware configuration, when the user operates the keyboard 41 or the mouse 42, the image processing installed in the personal computer 30 is started.

  When image processing starts, the personal computer 30 displays each frame image stored as a frame image group in the image database 20 on the display 43 (step S200). For this display, for example, a list of thumbnails may be displayed in the display area on the right half of the display 43, or each image may be displayed and switched one by one. Since the user specifies a desired frame image by operating the mouse or the like, the personal computer 30 determines whether or not this frame image is selected (step S210) and waits.

  The personal computer 30 reads data (image data) of pixels (for example, dot matrix-like pixels) constituting the selected frame image (step S220). This image data is composed of RGB color data and alpha channel data that determines whether image transmission is possible. The personal computer 30 determines an area (perspective area in the frame image) that is supposed to be transmitted by the alpha channel data by reading the image data of the frame image, and has a predetermined relationship with the image of the original image to be synthesized corresponding thereto. The region is determined as a see-through region (step S230).

  Referring to FIG. 3, the personal computer 30 reads the image data of each pixel of the end point coordinates (Xe, Ye) from the start point coordinates (Xs, Ys) of the matrix coordinates, and coordinates (Xa1, Yb1), (Xa1). , Ybi), (Xai, Yb1), (Xai, Ybi), the perspective area of the rectangular frame image is set as the perspective area of the composition target original image. This is because the image data of this rectangular area in the frame image has alpha channel data as if all image transmission is performed. In determining the fluoroscopic region, in addition to the above-described rectangular vertex coordinates, information for determining the position, for example, the relationship between the start point coordinates (Xs, Ys) and the rectangular corner coordinates (Xa1, Yb1) It is determined. In this case, if the fluoroscopic region of the frame image is determined to be rectangular, the fluoroscopic region of the composition target original image can be determined by the diagonal coordinates (Xa1, Yb1) and (Xai, Ybi). If the perspective area of the frame image is an irregular shape, the perspective area of the irregularly shaped frame image and the perspective area of the composition target original image can be determined by using the coordinates on the outline. The fluoroscopic region and the fluoroscopic region in these cases have a relationship that their shapes are the same.

  The fluoroscopic area and the fluoroscopic area corresponding to the fluoroscopic area are different if the frame images are different. Therefore, the fluoroscopic area is determined in step S230 for each frame image used for image synthesis. The fluoroscopic area once determined is stored in the hard disk 34 in association with the identification data (for example, frame name data) of the frame image, and from the next time the identification data of the frame image is determined, It can also be configured to determine the see-through region of the read compositing target original image.

  The personal computer 30 displays each composition target original image stored as a composition target original image group in the image database 20 on the display 43 in the same manner as the frame image display / selection standby (step S240). The image selection waits (step S250). In this case, the composition target original image may be displayed prior to the main image composition process, and the main image composition process may be executed upon selection of the composition target original image by the user. With this configuration, steps S240 and 250 are omitted.

  The personal computer 30 reads data (image data) of pixels (for example, dot matrix-like pixels) constituting the selected composition target original image (step S260). Then, among the read image data, the image data included in the fluoroscopic region determined in step S230 is sampled, and data processing based on the statistical characteristics of the image obtained as a result of the sampling is performed (step S270). . This sampling data processing will be described with reference to FIG. 3. The perspective coordinates determined in step S230 are reflected in the read composition source original image, and specifically, the vertex coordinates that define the rectangular-shaped transparent region described above. And a region to be synthesized are allocated to the original image to be synthesized based on the information for determining the position, the image data (RGB color data) of the pixels included in the region are sampled for each pixel, and the histogram of the luminance distribution as a statistical characteristic thereof Get. For example, this histogram (median m0) is subjected to data processing so as to approach the normal distribution histogram (median m), and the brightness is adjusted. In addition, it is also possible to perform contrast adjustment by performing data processing that widens the maximum and minimum values of the histogram to the left and right. Such data processing is referred to as correction α for convenience.

  The luminance distribution obtained in step S270 is for the image data of the pixels included in the see-through region, and is different from that obtained for the entire image of the composition target original image as the sampling region. Explaining in the figure, since the background image is seen on the right side or above in the figure as shown in FIG. 3 in the composition target original image, if the whole image of the composition target original image is sampled, the image of the pixels of such background image The data also rides on the histogram and affects the luminance distribution. However, since the sampling in step S270 is limited to the image data of the pixels included in the fluoroscopic region, the image data of the pixels of the background image is not on the histogram and is not reflected in the luminance distribution. In other words, the result of the data processing in step S270 (the adjustment result of brightness, contrast, etc.) is as if the transparent area is an entire image. Therefore, when the transparent area is viewed, The image is balanced. On the other hand, since the data processing result of the entire composition target original image including the background image reflects the image data of the pixels of the background image, assuming that the image of the processing result is cut out and viewed in the fluoroscopic region. The cut image has a balance affected by the background image, and the balance tends to be lost.

  Subsequent to step S270, the personal computer 30 generates a new image to be synthesized based on the image data that has been processed (step S280). This new image generation may be performed as follows. As shown in FIG. 3, correction α by data processing is performed on the fluoroscopic region window range Wa in the composition target original image, and a new image including the image of the fluoroscopic region window range Wa that has been corrected α is generated. be able to. In addition, it is also possible to generate a new image including an image obtained by performing correction α on the extended window range Wb extended from the see-through region window range Wa. Alternatively, the correction α can be performed on the entire area of the composition target original image to generate a new image. These new image generation methods may be determined by a mouse operation of the user interface 40 or the like.

  The personal computer 30 combines the new image to be combined generated in this way and the frame image selected in step S210 so that the frame images are on top (step S290), and ends this routine. The obtained composite image is stored in the hard disk 24 or the like, and is output as a print image to the color printer 50 or as a wallpaper image of the display 43 or the mobile phone 25 based on an instruction from the user interface 40.

  In this embodiment in which this series of image composition processing is performed, when superimposing the frame image and the image to be synthesized, the image data of the composition target original image included in the fluoroscopic region visually recognized by the frame image is sampled. Then, data processing (correction α) reflecting the statistical characteristics is performed, and a new image to be synthesized is generated (steps S270 to 280). Therefore, the image that is visually recognized by being combined with the frame image is subjected to correction α (image adjustment) based on the properties of the image of the visible region to be visually recognized, so that the balance of the image is not disturbed and the user feels uncomfortable. Can be reduced. Further, since the image adjustment to be visually recognized from the see-through region is based on the data processing of the image data of the original image, the uncomfortable feeling can be more reliably reduced.

  In the present embodiment, the first method for performing data processing (correction α) based on the sampling result of the composition target original image of the fluoroscopic region for the fluoroscopic region window range Wa in the composition target original image, and the fluoroscopic region window One of the second method performed for the extended window range Wb extended from the range Wa and the third method performed for the entire area of the composition target original image can be taken. Therefore, variations in image generation through the correction α can be increased. In particular, the new image generation by the first method is preferable because the data processing of the image data, that is, the number of data to be subjected to the correction α can be reduced, which can reduce the calculation load. In addition, in the new image generation by the second method, the calculation load can be reduced as compared with the third method, and the image area of the new image after generation can be made wider than the perspective area of the frame image, so that the perspective image as the sampling area can be obtained. The calculation at the time of area determination can be made rough, and the calculation load can be reduced accordingly.

  Furthermore, in the present embodiment, a plurality of frame images that can be frame images are stored in the image database 20, and the see-through area of the composition target original image corresponding to the see-through area of the frame image selected from the frame images is determined. (Steps S200 to 230). Therefore, even when frame images having different fluoroscopic regions are synthesized with the same composition target original image, the uncomfortable feeling can be reduced and the application range can be expanded. In addition, since various frame images are stored in the DVD 23 and the hard disk 24 of the image database 20 and used in the image composition, various composite images obtained by compositing various frame images and the original image to be combined are used. An image can be generated. It should be noted that various frame images can be stored in the hard disk 34 and used.

B. Image processing of the second embodiment:
The hardware configuration of the second embodiment is not different from that of the first embodiment described above, and the contents of the image processing are partially different. FIG. 4 is a flowchart of image processing according to the second embodiment for generating a single still image by synthesizing the first image and the second image to be superimposed, and FIG. 5 illustrates the contents of the image processing according to the second embodiment. It is explanatory drawing for doing.

  As shown in FIG. 4, even in the second embodiment, the personal computer 30 first displays each frame image stored as a frame image group in the image database 20 on the display 43 (step S200). It waits for the user to select a frame image (step S210). When the frame image is selected, the personal computer 30 reads the image data included in the selected frame image and analyzes the image data (step S225).

  In this embodiment, the image data structure of the frame image has a feature. As shown in FIG. 5, the image data includes meter data including data related to the image processing of the composition target original image, RGB color data, and the like. Frame image data. The metadata includes the frame name data of the frame image, as well as the perspective area data (start point coordinates (Xs, Ys) described above and the coordinates of the perspective area (Xa1) necessary for determining the perspective area of the composition target original image. , Yb1), (Xai, Ybi)) and correction data related to the image processing of the composition target original image.

  In this embodiment, since it is assumed that the frame image has a rectangular perspective area, it is sufficient to have the rectangular start point coordinates (Xa1, Yb1) and end point coordinates (Xai, Ybi) as the perspective area data. The data structure becomes simple. In this case, if the fluoroscopic region has an irregular shape, coordinate data along the contour of the irregular shape is accumulated in the metadata. The correction data includes data for further correcting the image processing (the correction α described above) of the composition target original image based on the color tone, contrast, and the like of the frame image, and a correction execution area that is an image processing target of the composition target original image. (Described above, data indicating the see-through region window range Wa, the extended window range Wb, and the entire image range Wall). For example, when the frame image has a bright color tone, the color tone is corrected to be brighter when the color tone is corrected by the correction α for the image data in the fluoroscopic region of the composition target original image described above. Data (for example, correction coefficient etc.) instructing (correction α → correction α0) is stored as correction data. As the correction execution area data to be subjected to image processing, any one of the already described fluoroscopic area window range Wa, the extended window range Wb, and the entire image range Wall is stored.

  When the personal computer 30 obtains data relating to the fluoroscopic region of the composition target original image and data correction as described above by reading and analyzing such various data, it is stored in the image database 20 as a composition target original image group. Each of the compositing target original images is displayed (step S240), and the user waits for selection of the compositing target original image (step S250). Thereafter, the image data of the selected composition target original image is read (step S260), the image data included in the fluoroscopic region analyzed and determined in step S235 is sampled, and the data processing is performed (step S275). For this sampling, a histogram of luminance distribution is obtained as in step S270. However, in the data processing using this histogram, the correction α is set to the correction α0 reflected by the correction data analyzed in step S235. As a result, the image data correction based on the sampling of the image data included in the fluoroscopic region of the composition target original image and the image data correction based on the frame image surrounding the fluoroscopic region of the composition target original image are reflected in the new image. Will be.

  Subsequent to step S275, the processes of steps S280 and S290 described above are executed, and the new image to be combined and the frame image generated through the correction α0 are combined so that the frame image is on top. This routine is terminated. In this case, in the new image generation in step S280, a new image is generated by executing correction α0 for the area determined by the correction execution area data acquired in step S225.

  Even in the second embodiment in which this series of image composition processing is performed, the same effects as in the first embodiment can be obtained. In particular, in the second embodiment, the image data correction based on the sampling of the image data included in the fluoroscopic area of the composition target original image is added with the image data correction based on the frame image surrounding the fluoroscopic area of the composition target original image. Therefore, the sense of incongruity can be reduced as a whole image after synthesis. Further, since the data of the frame image includes the data about the image processing of the composition target original image, it is simple.

C. Image processing of the third embodiment:
The third embodiment is characterized in that image composition is performed such that a partial image of the composition target original image is superimposed on the image area of the frame image, and the hardware configuration is the same as that of the first embodiment described above. FIG. 6 is a flowchart of image processing according to the third embodiment for generating a single still image by synthesizing the first image and the second image to be superimposed, and FIG. 7 illustrates the contents of the image processing according to the third embodiment. It is explanatory drawing for doing.

  As shown in FIG. 6, even in the third embodiment, the personal computer 30 first displays each frame image stored as a frame image group in the image database 20 on the display 43 (step S200). It waits for the user to select a frame image (step S210). When the frame image is selected, the personal computer 30 executes reading of image data included in the selected frame image (step S220).

  In this embodiment, the image data structure of the frame image has a feature. As shown in FIG. 7, the frame image overlaps the visual recognition area image SG that is visually recognized even after the synthesis and a partial image of the synthesis target original image. And a viewable region image HG that is no longer visible. Although the frame image has RGB color data for each of the viewing area image SG and the viewing area image HG, for the viewing area image HG, the synthesis target element is obtained by overlapping the partial image portion of the composition target original image. It includes data (superimposition data) indicating that it is a superposition region that allows this partial image portion of the image to be viewed.

  The personal computer 30 determines the overlapping area of the composition target original image based on the image data of the read frame image (step S300). This region determination method is not different from the method described in the first embodiment. As shown in FIG. 7, the region-of-view image HG grasped by the frame image superimposition data is superimposed on the composition source image. It corresponds to the region HG0. At this time, as described in the first embodiment, data such as the start point coordinates of the image area is used.

  When the overlapping area HG0 of the composition target original image is determined in this way, the personal computer 30 displays each composition target original image stored as a composition target original image group in the image database 20 (step S240), and the composition target by the user. The original image selection is waited (step S250). Thereafter, the image data of the selected composition target original image is read (step S260), and among the read image data, an image obtained with the image data included in the overlapping area HG0 of the composition target original image is read from the composition target original image. Cut out (step S310).

  The personal computer 30 samples this cut-out image (cut-out image), that is, image data included in the overlapping area HG0 of the composition target original image, and performs data processing thereof (step S320). For this sampling, as in the previous step S270, a luminance distribution histogram is obtained. For example, brightness adjustment is performed so that the histogram histogram (median m0) approaches the normal distribution histogram (median m), or the histogram maximum Contrast adjustment data processing (correction β) is performed to widen the value and minimum value to the left and right.

  Following step S320, the personal computer 30 generates a new cut-out image to be synthesized based on the image data that has been processed (step S330). This new image generation may be performed as follows. As shown in FIG. 7, correction β by data processing is performed on the image data of the clipped image in step S <b> 310 to generate a corrected new image with a clipped shape.

  The personal computer 30 adds the cut-out new image to be synthesized and the frame image selected in step S210 to the synthesis target cut-out new image in the region of the viewable region image HG in the frame image. In such a manner, they are combined (step S340), and this routine is terminated.

  Even in the third embodiment in which this series of image composition processing is performed, the same effects as in the first embodiment can be obtained.

D. Variation:
Although some examples have been described above, the present invention is not limited to these examples and embodiments, and can be implemented in various modes without departing from the scope of the invention. For example, the following modifications are possible.

  FIG. 8A is an explanatory diagram for explaining a modified example of the method for determining the perspective area of the composition target original image (step S230) at the time of image composition in which the frame image is superimposed on the composition target original image, and FIG. FIG. 8C is an explanatory diagram for explaining another modified example of the fluoroscopic region determining method (step S230). These modified examples take a method of determining the fluoroscopic region of the composition target original image as a rectangular shape when the fluoroscopic region of the frame image has an irregular shape.

  FIG. 8A shows that when the frame image has an irregularly shaped perspective region TR shown in the figure, the rectangular-shaped perspective region HTR in the composition target original image has a maximum rectangular shape that fits in the irregularly shaped perspective region TR. An example is shown. Further, in this modification, even when the frame image has a rectangular perspective region as shown in FIG. 1, the rectangular perspective region HTR in the composition target original image is defined as the rectangular perspective region of the frame image. It also means that they do not necessarily need to be matched. In addition, if the rectangular shape is a rectangle, the rectangle may have a shape in which the four sides are in contact with the innermost side of the irregularly shaped perspective region TR.

  In this way, the sampling area (that is, the fluoroscopic area HTR) at the time of image data processing in step S270 can be narrowed, so that the calculation load can be reduced. In this case, when data processing (correction α) is performed to generate a new image, the area to be corrected may be the extended window range Wb wider than the fluoroscopic area TR or the entire area of the composition target original image. . Note that the extended window range Wb can also be the see-through region HTR. In this case, the rectangular-shaped see-through region HTR in the composition target original image is set to the minimum size of the rectangular shape surrounding the deform-shaped see-through region TR. This is an example.

  FIG. 8B shows an example in which, when the frame image has the irregularly shaped perspective region TR shown in the figure, the rectangular perspective region HTR in the composition target original image interferes with the outline of the irregularly shaped perspective region TR. Show. In other words, in FIG. 8A, the minimum rectangular-shaped perspective region HTR of the composition target original image indicated by a solid line, and the intermediate rectangular-shaped perspective region HTR between the maximum rectangular-shaped perspective region HTR indicated by a two-dot chain line, It is also an example.

  8C shows a rectangular-shaped perspective region in the composition target original image when the frame image has an irregularly shaped perspective region TR illustrated in FIG. 8A. The minimum rectangular shape perspective region of the composition target original image shown by a solid line in FIG. It has HTR1 and a transparent region HTR2 having an intermediate rectangular shape shown by a solid line in FIG. 8B, and on that basis, determination of the transparent region in step S230 and weighting of data processing (correction α) in step S270 are performed. An example is given. That is, in FIG. 8C, the image data included in the transparent region HTR1 having the smallest rectangular shape is set as the transparent region in step S230, and the image data included in the transparent region (the transparent region HTR1 having the smallest rectangular shape) is Sampling is performed in step S270. Then, the correction α for the image data included in the fluoroscopic region HTR1 having the minimum rectangular shape is subjected to data processing by the method of the correction α described above with the weighting set to 1. However, the correction α for image data included in the intermediate rectangular shape of the transparent region HTR2 and included in the region away from the minimum rectangular shape of the transparent region HTR1 is weighted, for example, 0.7. Data processing. This has the following advantages.

  When data correction is performed on the compositing target original image based on the sampling result in step S270, an image included in the region without interference with the fluoroscopic region of the frame image (an image corresponding to the fluoroscopic region HTR1 having the smallest rectangular shape) is weighted. Is an image that has been subjected to the correction α data processing with 1 and the outer image (the image between the intermediate rectangular-shaped transparent region HTR2 and the minimum rectangular-shaped transparent region HTR1) has different weights. It is also possible to match the color tone / contrast of the image to the transition of the image in the image corresponding to the transparent region HTR1 having the smallest rectangular shape and the frame image. Therefore, it is possible to further reduce the uncomfortable feeling in the composite image.

  In addition, when determining the perspective area of the composition target original image, the start point coordinates (Xs, Ys) of the matrix coordinates of the perspective area in the frame image and the perspective area of the composition target original image are the same. You can also In general, an image incorporated in a see-through area of a frame image and seen through is a human face. Therefore, if the composition target original image includes a person image, the processing in step S230 includes processing for determining a fluoroscopic region in the frame image, and specifying a human face description portion included in the composition target original image, It is also possible to divide the face drawing portion into adaptive processing that is determined so as to correspond to the fluoroscopic region in the frame image. This is preferable because the face of the person in the composition target original image can be synthesized near the center of the perspective area of the frame image. It should be noted that the identification of the face depiction portion of the person included in the composition target original image is executed through analysis such as a state in which matrix-like pixel data is distributed vertically and horizontally as skin color data.

  The following modifications are also possible. In the second embodiment, the image data correction (correction α) based on the sampling of the image data included in the fluoroscopic region of the composition target original image is corrected to the image data correction based on the frame image surrounding the fluoroscopic region of the composition target original image (correction α). In consideration of the correction α0), the correction α0 is included in the frame image data. However, such correction α0 can be modified so as to be calculated from the frame image itself.

  In other words, the image characteristic of the image obtained from the image data included in the region (for example, luminance distribution) is set as the sampling region, the image region of the frame image including at least the fluoroscopic region of the frame image, for example, the image region around the fluoroscopic region. ) Then, this statistical characteristic is used in place of the correction α0 in the second embodiment to perform image data correction of the composition target original image. Even in this case, the same effect as the second embodiment can be obtained. In addition to the luminance distribution, the composition target original image may be subjected to complementary color correction (data correction) according to the color tone of the image around the fluoroscopic region in the frame image obtained as a statistical characteristic.

It is explanatory drawing which shows the image processing system 100 as 1st Example of this invention. It is a flowchart of the image processing of 1st Example which synthesize | combines the 1st image used as a superimposition object, and a 2nd image, and produces | generates one still image. It is explanatory drawing for demonstrating the content of the image processing of 1st Example. It is a flowchart of the image processing of 2nd Example which synthesize | combines the 1st image used as a superimposition object, and a 2nd image, and produces | generates one still image. It is explanatory drawing for demonstrating the content of the image processing of 2nd Example. It is a flowchart of the image processing of 3rd Example which synthesize | combines the 1st image used as a superimposition object, and a 2nd image, and produces | generates one still image. It is explanatory drawing for demonstrating the content of the image processing of 3rd Example. It is explanatory drawing explaining the 1st modification of the see-through | perspective area | region determination method of the synthetic | combination object original image at the time of the image composition which overlaps a frame image on a compositing object original image. It is explanatory drawing explaining the other modification of the to-be-transparent area | region determination method in the case of the superimposition synthesis | combination of an image. It is explanatory drawing explaining another modification of the to-be-viewed area | region determination method in the case of the superimposition synthesis | combination of an image.

Explanation of symbols

20 ... Image database 21 ... Digital video camera 22 ... Digital still camera 23 ... DVD
24 ... Hard disk 25 ... Mobile phone 30 ... Personal computer 34 ... Hard disk 35 ... I / F circuit 40 ... User interface 41 ... Keyboard 42 ... Mouse 43 ... Display 50 ... Color printer 100 ... Image processing system SG ... Viewing area image HG ... Viewing area image TR ... Perspective area HTR ... Perspective area

Claims (13)

An image generation device that generates a composite image by superimposing images,
An image processing unit that processes the first image located below when the first image and the second image to be superimposed overlap;
A synthesis unit that synthesizes the second image and the first image after adjustment so that the second image having a fluoroscopic region for seeing through the first image overlaps the first image;
The image processing unit
An area determining unit that determines a predetermined corresponding area on the first image so as to correspond to the fluoroscopic area of the second image in a predetermined relationship;
An image generation apparatus comprising: an adjustment unit that adjusts first image data constituting the first image based on the property of the first image in the determined corresponding region. The image generation apparatus according to claim 1,
The adjustment unit is
When adjusting the first image by performing correction on the first image data, the corresponding area is set as a sampling area, and the image data is based on statistical characteristics of an image obtained from the first image data in the sampling area. An image generation apparatus that adjusts the first image by correction. The image generation apparatus according to claim 2,
The adjustment unit is
An image generation apparatus that performs image data correction based on an image data state of the sampling area for image data including at least the first image data of the corresponding area among the first image data constituting the first image. The image generation device according to claim 3,
The adjustment unit is
Of the first image data constituting the first image, the first correction of the corresponding region is performed in the first correction performed on the first image data constituting the first image. An image generation apparatus that is selectively executed from the second correction performed on image data. The image generation apparatus according to any one of claims 2 to 4,
Furthermore,
A storage unit for storing a plurality of images having different perspective areas;
An image generation apparatus comprising: an image selection unit that selects one of the plurality of stored images as the second image. The image generation apparatus according to claim 5, wherein
The image generation apparatus, wherein the storage unit is capable of inputting an image having a different perspective area from the outside. The image generation apparatus according to claim 2, wherein:
The second image data constituting the second image includes information for determining the corresponding area. The image generation apparatus according to any one of claims 2 to 7,
The second image data includes correction-related data related to the content of correction performed on the first image data during the first image adjustment,
The image generation apparatus that adjusts the first image by correcting the first image data by reflecting the correction participation data. The image generation apparatus according to any one of claims 2 to 7,
Furthermore,
With respect to a second image region including at least a fluoroscopic region surrounding image portion of the second image around the fluoroscopic region of the second image, the second image sampling region is defined as the second image sampling region. A second image sampling unit for obtaining a statistical characteristic of an image obtained from the second image data of the second image included in
The image generation device adjusts the first image by reflecting the statistical characteristic obtained by the second image sampling unit in the correction of the first image data. An image generation device that generates a composite image by superimposing a plurality of images,
An adjustment unit for adjusting the first image during image synthesis;
A synthesis unit that synthesizes the first image after processing and the second image to be superimposed;
The second image is
An overlapping region defined so that the image portion of the first image can be seen by overlapping an image portion of the partial region of the first image on the second image;
The adjustment unit is
Performing a process of cutting out the image portion of the first image corresponding to the overlapping area, and an adjustment process of the first image based on the property of the first image in the area to be cut out;
The synthesis unit is
An image generation apparatus that combines the image portion of the partial area of the cut out first image with the overlapping area of the second image. An image generation method for generating a composite image by superimposing a plurality of images,
(A) With respect to the first image positioned below when the first image and the second image to be overlapped overlap each other, the second image has a predetermined relationship with the fluoroscopic region that allows the first image to be seen through Determining a predetermined corresponding area on the first image to correspond with
(B) Among the first image data constituting the first image, the corresponding area determined in the step (a) is set as a sampling area, and the statistical value of the image obtained from the first image data in the sampling area is determined. Adjusting the first image by correcting image data based on characteristics;
(C) combining the second image that overlaps the first image and the first image that has been adjusted in the step (b) so that the second image is on the first image. An image generation method comprising: A computer program for causing a computer to execute image composition processing for superimposing a plurality of images,
(A) With respect to the first image positioned below when the first image and the second image to be overlapped overlap each other, the second image has a predetermined relationship with the fluoroscopic region that allows the first image to be seen through A function of determining a predetermined corresponding area on the first image so as to correspond with
(B) Out of the first image data constituting the first image, the corresponding area determined by the function (a) is set as a sampling area, and the statistical value of the image obtained from the first image data in the sampling area A function of adjusting the first image by correcting image data based on characteristics;
(C) A function of synthesizing the second image overlapping the first image and the first image adjusted by the function (b) so that the second image is on the first image. When,
A computer program for causing the computer to realize the above.   A computer-readable recording medium on which the computer program according to claim 12 is recorded.

Images