JP2007323583A - Image converter and image conversion program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image converter with a high degree of freedom for performing different image conversion in respective partial areas of an original image. <P>SOLUTION: This image converter is provided with a splitting grid generation means generating a first splitting grid for splitting the original image into a plurality of partial areas and a second splitting grid for defining a converted image as one or more split areas, a selection means selecting one or more partial areas for performing image conversion among a plurality of partial areas split by the first splitting grid, an area correspondence decision means associating each of the split areas with the selected partial area, a pixel value extraction means calculating positional coordinates of the partial areas corresponding to respective pixels constituting the split area by using the corresponding partial area as to each of the split areas and extracting pixel values of pixels in the obtained positional coordinates, and a means giving the pixel values extracted respectively by the pixel value extraction means as pixel values of the respective pixels constituting the split area as to each of the split areas and generating a converted image formed of these split areas. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an image conversion apparatus that extracts a predetermined area included in an original image and creates a converted image, and in particular, an image conversion apparatus that performs different image conversion for each of a plurality of partial areas of an original image. And an image conversion program.

  As this type of image conversion apparatus, for example, there is a general-purpose image processing tool equipped with various image processing functions, such as Photoshop (registered trademark) manufactured by Adobe. This Photoshop (registered trademark) is provided with a technique for creating a converted image by various geometric transformations for a single range designated by an original image.

  In Patent Document 1, a three-dimensional solid model is defined by connecting a triangular or quadrangular small figure in a three-dimensional grid, and the three-dimensional grid pattern is developed on the original two-dimensional image data. A technique for mapping pixels of an original image to a three-dimensional grid while associating with the two-dimensional grid pattern thus disclosed is disclosed.

On the other hand, Patent Document 2 discloses a method of performing geometric correction on an image having optical distortion caused by a lens when a planar object is photographed by a camera. When projecting to the screen, when determining the correction amount of the source image, the grid pattern is projected and the amount of deviation from the orthogonality is measured, so that the source image is deformed in advance, Each of the technologies capable of maintaining verticality is disclosed.
JP-A 64-29976 Patent No. 3429280 JP 2006-33357 A

  In the existing image processing tool described above, it is possible to apply the same conversion process to a plurality of areas. However, when performing different conversion processes for each area, the conversion is individually repeated for each area. It was necessary to define the contents of processing. For this reason, there is a problem that processing is complicated when different conversions are performed in a plurality of areas, and further, image connectivity between the areas cannot be maintained. The conversion function is also limited to affine transformation that can be defined by linear algebra.

  Therefore, the present invention has been made in view of the above problems, and provides a highly flexible image conversion apparatus and image conversion program capable of performing different image conversion for each of a plurality of partial regions of an original image. Purpose.

  In order to solve the above-described problem, the invention according to claim 1 is an image conversion apparatus that creates a converted image by extracting a predetermined region included in an original image, and divides the original image into a plurality of partial regions. Generating a first divided grid and generating a second divided grid for defining the converted image as at least one or more divided regions; and a plurality of the plurality of divisions divided by the first divided grid Of the partial areas, selection means for selecting at least one partial area for image conversion, area correspondence determining means for associating each of the divided areas with the selected partial area, and each of the divided areas , Using the partial area corresponding to the divided area, the position coordinates of the partial area corresponding to each pixel constituting the divided area are calculated and included in the partial area. A pixel value extracting means for extracting a pixel value of a pixel located at a calculated position coordinate, and the pixel value extracting means for each divided area as a pixel value of each pixel constituting the divided area And a converted image creating means for creating the converted image composed of these divided regions, to which the pixel values extracted respectively in the above are applied.

  According to this, since the image conversion is performed for each area using the first divided grid of the original image and the second divided grid of the converted image, different image conversion is performed for each of the plurality of partial areas of the original image. Can be applied quickly.

  In order to solve the above-mentioned problem, the invention according to claim 2 is the image conversion apparatus according to claim 1, wherein the divided grid generation means uses the plurality of grid lines to form the first divided grid and the second divided grid. A grid is generated, and within the range where the grid lines do not intersect, the positions of the intersections of the grid lines and the grid lines indicating the outlines of the divided grids and the contact points of the grid lines are changed. An image conversion apparatus having a correction unit that corrects the shape of a divided area.

  According to this, since the first divided grid and the second divided grid having various shapes can be generated, the degree of freedom of image conversion is improved, and various and non-linear image conversion processing can be performed with a simple operation. A possible image conversion device can be provided.

  In order to solve the above problem, the invention according to claim 3 is the image conversion apparatus according to claim 1 or 2, wherein the number of partial areas selected by the selection unit and the divided areas of the converted image are determined. This is an image conversion device characterized in that the numbers are equal.

  According to this, it is possible to quickly perform different image conversions on the selected partial areas among the plurality of partial areas of the original image.

  In order to solve the above problem, the invention according to claim 4 is the image conversion device according to any one of claims 1 to 3, wherein the partial area of the original image is divided by the first divided grid. Is equal to the number of divided areas of the converted image defined by the second divided grid, and all the partial areas are selected by the selection means.

  According to this, since the image conversion is performed for each area using the first divided grid of the original image and the second divided grid of the converted image, different image conversion is performed for each of the plurality of partial areas of the original image. Can be applied quickly.

  In order to solve the above-mentioned problem, the invention according to claim 5 is the image conversion apparatus according to any one of claims 1 to 4, wherein the original image is a moving image composed of a plurality of image frames. If there is a representative image determination unit for determining a representative image frame from the plurality of image frames, the division grid generation unit is configured to divide the determined representative image frame into a plurality of partial regions. The divided grid is generated, the second divided grid for defining the converted image frame corresponding to the representative image frame as at least one or more divided areas is generated, and the representative grid generated by the divided grid generating means Storage means for storing the first divided grid and the second divided grid related to the image frame, and the representative image frame and the other image frames When generating the corresponding converted image frame, the converted image generating means uses the first divided grid and the second divided grid stored in the storage means to generate the representative image frame and the other image frames. A corresponding converted image frame is created, and a converted moving image is created based on the created converted image frame.

  According to this, since the divided grid used in the image conversion of the representative image frame can be used as it is for the image conversion of the other image frames, the original image is for a moving image composed of a plurality of image frames. Even for an image, it is possible to quickly perform image processing on each image frame and create a converted moving image based on each created converted image frame.

  In order to solve the above-mentioned problem, the invention according to claim 6 is the image conversion apparatus according to claim 5, wherein the representative deciding means is the first of the plurality of image frames constituting the moving image. An image conversion apparatus that determines an image frame as the representative image frame.

  According to this, since the division grid of the first image frame can be used as it is for image conversion of other image frames, image processing can be performed more quickly on each image frame.

  In order to solve the above problem, according to a seventh aspect of the present invention, in the image conversion device according to the sixth aspect, the divided grid generation means is the last of the plurality of image frames constituting the moving image. Generating the first divided grid for dividing the image frame into a plurality of partial areas, and defining the converted image frame corresponding to the last image frame as at least one of the divided areas A grid is generated, and the storage unit stores a first divided grid and a second divided grid relating to the last image frame generated by the divided grid generation unit, and the converted image generation unit stores the representative image. When creating a converted image frame corresponding to an image frame other than the frame and the last image frame, the storage means stores the A first divided grid according to the front image frame, a linear interpolation grid obtained by linear interpolation of the first divided grid according to the last image frame, a second divided grid according to the representative image frame stored in the storage means, Using the linear interpolation grid obtained by linearly interpolating the second divided grid for the last image frame, each of the converted image frames corresponding to the other image frames is created, and based on each of the created converted image frames Thus, an image conversion apparatus is characterized in that a converted moving image is created.

  According to this, a converted image frame corresponding to an intermediate image frame that is another image frame is created using a divided grid of the first and last image frames and a linear interpolation grid obtained by linearly interpolating these divided grids, Since the converted moving image is created based on each created converted image frame, it is possible to quickly and easily create a converted moving image in which the image is gradually deformed with the converted image time.

  In order to solve the above problem, an invention according to claim 8 is an image conversion program that causes a computer to function as the image conversion apparatus according to any one of claims 1 to 7.

  According to the present invention, it is possible to quickly perform different image conversion for each of a plurality of partial regions of an original image.

  DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described with reference to the accompanying drawings.

[1. Configuration and function of image conversion apparatus]
FIG. 1 is a block diagram showing a configuration of an image conversion apparatus according to the present invention. The image conversion apparatus can use a general-purpose computer.

  As shown in the figure, the image conversion apparatus 1 stores a CPU (Central Processing Unit) as a computer having a calculation function, a working RAM (Random Access Memory), various data and programs (including the image conversion program of the present invention). And a storage unit for storing various image data including a conversion source image (an example of an original image of the present invention) for performing image conversion, and the like. 12. An input unit 13 through which an operator gives various input instructions when generating a divided grid used for image conversion processing, the generated divided grid, a conversion source image, and a conversion destination image obtained as a result of image conversion (the present invention) The display unit 14 displays an example of the converted image), and the control unit 11, the storage unit 12, the input unit 13, and the display unit 14 are connected via a bus 15. They are connected to each other.

  The storage unit 12 functions as a storage unit of the present invention, and a conversion source image storage unit 121 for storing / storing the conversion source image data, and an image conversion process described later to perform image conversion on the conversion source image. The divided grid storage unit 122 for storing / storing the divided grid generated by the above and the conversion destination image storage unit 123 for storing / storing the conversion destination image data obtained as a result of the image conversion.

  And the control part 11 produces | generates the control information for controlling each structural member, outputs the said control information to a corresponding structural member via the bus | bath 15, and carries out control over the operation | movement of each said structural member, By executing an image conversion program, which will be described later, stored in a ROM or the like, the divided grid generation means, selection means, area correspondence determination means, pixel value extraction means, and conversion image creation of the present invention in cooperation with other components Means and correction means.

[2. Split grid]
First, a divided grid (an example of a first divided grid of the present invention) of a conversion source image that is an object of image conversion will be described with reference to FIG.

  FIG. 2 is an explanatory diagram of the division grid of the conversion source image.

  As shown in the figure, the division grid of the conversion source image has the same size as the conversion source image, and includes a plurality of grid lines for dividing the conversion source image into a plurality of partial areas. Depending on the number and position of the grid lines, various sizes, shapes, and the like of the partial areas into which the conversion source image is divided can be designated.

  The image conversion process of the present invention extracts pixel values of pixels in the partial area of the conversion source image, and uses the pixel values as divided areas of the conversion destination image corresponding to each partial area (an example of the second divided grid of the present invention). ), It is possible to perform image conversion such as enlargement / reduction / cut-out for each partial area.

  FIG. 3 is an explanatory diagram of the division grid of the conversion source image and the division grid of the conversion destination image.

  The division grid of the conversion destination image is for defining the conversion destination image as a plurality of divided areas, and is composed of a plurality of grid lines, like the division grid of the conversion source image.

  Each divided area is associated with a partial area of the conversion source image, and the pixel value of each pixel constituting the divided area includes any pixel constituting the partial area corresponding to the divided area. The same pixel value is given. The calculation of the position coordinates of “any pixel constituting the partial area” will be described in detail in an image conversion process described later.

  As described above, since image processing for the conversion source image is performed in association with the divided areas having independent sizes and shapes for each partial area using the divided grid, different image conversions ( Enlargement / reduction, etc.) can be performed in a single process.

[2-1. Generate split grid]
Next, generation of a divided grid will be described with reference to an explanatory diagram of divided grid generation in FIG.

  First, the outline of the divided grid is defined with the same size as that of the conversion source image, and a plurality of grid lines (including grid lines serving as outlines) are generated in the horizontal direction and the vertical direction of the conversion source image (FIG. 4). (A)). If there is a portion (margin portion) that is not subject to image conversion around the conversion source image, the portion excluding that portion is set as the conversion source image. In addition, the grid line is generated by displaying the conversion source image on the display unit 14, and operating the pointer on the conversion source image displayed on the display unit 14 by operating the mouse or the like of the input unit 13 by the operator. It is preferable that the grid line be created while checking the grid line.

  Then, the intersection of each grid line (indicated by Δ in FIG. 4B) is moved (corrected) by operating (clicking and dragging) with the mouse of the input unit 13 or the like (FIG. 4). 4 (C)), a division grid for dividing the conversion source image into a plurality of partial regions Ro is generated (FIG. 4D). In addition to the intersection of each grid line, the position of the contact point (excluding the vertex of the outline) between the grid line indicated by ◯ in FIG. 4B and the grid line applied to the outline can be corrected. Each partial area Ro shares a vertex (intersection of the grid lines) and a side with four adjacent partial areas Ro.

  Note that the correction of the intersection and the contact is performed within a range where the grid lines do not intersect. For example, in the example shown in FIG. 5A, the correction of the intersection is performed within a range where the grid lines do not intersect, so that the correction is correctly performed (suitable). However, in the example shown in FIG. 5B, since the intersection is corrected beyond the grid line, the correction is inappropriate (unsuitable). When moving the intersection with the mouse of the input unit 13 or the like, a drag restriction is provided so that each intersection cannot be moved beyond the grid line, or a warning is displayed when the grid line is crossed during the drag (for example, crossed Preferably, the grid lines are displayed in a different color from the other grid lines), and the operator is assisted to correct the intersections and contacts correctly.

  The divided grid of the conversion destination image also defines a divided area with each grid line by the same method as described above. In addition, the divided grid of the conversion destination image can also correct the intersections and contact points of the respective grid lines in the same manner as described above to define divided regions having various nonlinear shapes, but the divided grid of the conversion destination image has a square size. Are all defined by a 90-degree square, the calculation calculation of pixel position coordinates, which will be described in detail later, becomes easier.

  Then, the partial area Ro desired to be converted is selected from the partial areas Ro of the conversion source image, and the selected partial area Ro is associated with the divided area of the conversion destination image.

  For example, FIG. 6 shows an example in which all the partial areas Ro are selected and each is associated with the divided area Rr of the conversion destination image. The selection of the partial area Ro and the association between the selected partial area Ro and the divided area Rr are also performed by the operator selecting and linking them by operating the input unit 13. Also, as shown in the figure, when all the partial areas Ro are selected and associated with any of the divided areas Rr of the conversion destination image, for example, numbers (1) are sequentially assigned from the upper left partial area Ro. ), (2),..., (12), and the divided region Rr is similarly assigned numbers (1), (2),. You may comprise so that the partial area | region Ro and the division | segmentation area | region Rr which have the same number may be matched autonomously.

[2-2. Examples of various division grids and image conversion]
As another example, as shown in FIG. 7, only the four-corner partial areas Ro are selected by selecting the four-corner partial areas Ro from among the partial areas Ro of the conversion-source image and associating them with the divided areas Rr of the conversion-destination image. The image can be expanded and extracted.

  Furthermore, as shown in FIG. 8, among the partial areas Ro of the conversion source image, the partial areas Ro ((i, j) where the position coordinates (i, j) of the partial areas in the conversion source image are both even. = (2, 2), (2, 4), (4, 2), (4, 4)), and by associating each with the divided area Rr of the conversion destination image, Image conversion in which only specific coordinate positions are extracted can be performed.

  Further, as shown in FIG. 9, the conversion destination image is defined as one divided region Rr (1), one partial region Ro (5) is selected from the partial regions Ro of the conversion source image, and the partial region is selected. If Ro is associated with the divided region Rr (1), only the partial region Ro (5) of the conversion source image can be cut out and extracted (trimmed).

  As described with reference to FIGS. 6 to 9, the number of selected partial regions Ro and the number of divided regions Rr need to be the same.

  As described above, since the image processing for the conversion source image is performed by associating with the divided areas having independent sizes and shapes for each partial area, different image conversion processes are performed for each partial area at a time. (Enlargement / reduction, etc.) can be applied.

  Furthermore, a specific example is given and demonstrated.

  The example shown in FIG. 10 generates a divided grid that becomes a larger partial area Ro toward the lower side for the conversion source image, and a rectangle having the same length in the vertical direction for the conversion destination image. It is an example of image conversion when a divided grid that can define the divided region Rr by shape is generated. By using each of such divided grids, when the conversion source image is a photograph with a person as a subject as indicated by a dotted line in the figure, image conversion can be performed by enlarging the background of the subject (zoom out). it can.

  FIG. 11 is an example of image conversion when the conversion source image is a map. FIG. 11A is a conversion source image, FIGS. 11B and 11C are diagrams showing the conversion source image divided grid and the conversion source image superimposed, and FIG. (C) shows the state after correction of the intersection and tangent of the grid lines (division grid completed state). FIG. 11D is a diagram in which the divided grid defining the conversion destination image and the conversion destination image obtained as a result of the image conversion are superimposed.

  The partial areas Ro (1) to (9) of the conversion source image are all selected as the objects of image conversion, and each partial area Ro is divided into the divided areas Rr (1) to (9) according to the division numbers (1) to (9). 9).

  Then, as shown in FIGS. 11C and 11D, only the partial area Ro (5) of the part to be enlarged in the conversion source image is the rectangular shape of the corresponding divided area Rr (5) of the conversion destination image. It is configured to be smaller than the size of.

  When image conversion is performed in such a state, the divided region Rr (5) is displayed as an image obtained by enlarging the partial region Ro (5). Further, each partial area Ro shares the vertex (intersection of the grid lines) and the side with the four adjacent partial areas Ro, so that it corresponds without missing one pixel constituting the conversion source image. In addition to being able to extract (that is, to perform image conversion) as a pixel value of any pixel in the divided region, it further corresponds to each of the four partial regions Ro adjacent to each other and each partial region Ro sharing a vertex and side. Similarly, each of the attached divided regions Rr also shares vertices and sides with four adjacent divided regions Rr. Therefore, it is possible to perform different image processing (enlargement / reduction) for each partial region Ro while maintaining the connection of images between the partial regions. Therefore, as shown in FIGS. 11C and 11D, the conversion source image is divided into a plurality of partial regions Ro by the division grid, but since the vertex and the side are shared, the route on the map, the road, It is possible to perform enlargement / reduction image processing while maintaining continuity without dividing the coastline, character information, and the like at the boundary of each partial region Ro.

[3. Image conversion processing]
Next, an example of a specific method of image conversion processing will be described with reference to the drawings.

  FIG. 12 is a flowchart showing image conversion processing. In this process, when the operator operates the input unit 13 to instruct the start of the process, the following process based on the control of the control unit 11 is started.

  First, the control unit 11 acquires (captures) conversion source image data to be subjected to image conversion from the conversion source image storage unit 121 of the storage unit 12 (step S1). In order to perform image conversion of the acquired conversion source image data (hereinafter referred to as “conversion source image”), the control unit 11 functions as a division grid generation unit and a correction unit, and converts the division grid and conversion destination of the conversion source image. An image division grid is generated (step S <b> 2), and the generated division grid is stored in the division grid storage unit 122.

  Then, image conversion processing is executed using the divided grid stored in the divided grid storage unit 122 (step S3).

  In “execution of image conversion process” in step S3, first, the control unit 11 functions as a selection unit, and among the partial areas Ro divided by the division grid of the conversion source image, the partial area Ro for which image conversion is desired. (N) is selected (step S11). Next, the control unit 11 functions as an area correspondence determining unit, and associates the selected partial area Ro (n) with the divided area Rr (n) defined by the divided grid of the conversion destination image (step S12). .

  Then, a variable “n” indicating the number of sets of the partial region Ro (n) and the divided region Rr (n) is initialized as 1 (step S13), and the pixel P (X, Y) constituting the divided region Rr (n) is initialized. ) Is initialized as (x1, y1) (step S14). In the following process, pixel values are assigned in order from the pixel P located at the position coordinate (x1, y1) of the divided region Rr (n).

  First, in the partial area Ro (n) corresponding to the divided area Rr (n), position coordinates (Xo, Yo) corresponding to the pixels P constituting the divided area Rr (n) are calculated (step S15).

  Here, the calculation method of the position coordinates (Xo, Yo) will be described in detail with reference to FIG.

  13A and 13B are explanatory diagrams of the divided region Rr and the partial region Ro corresponding to the divided region Rr.

  The coordinates (xo1, yo1), (xo2, yo2), (xo3, yo3), (xo4, yo4) of the partial area Ro, the coordinates (x1, y1), (x2, y2) of the four vertices of the divided area Rr ), (X3, y3), and (x4, y4) are known values already known when the divided grid is generated.

Therefore, by obtaining the unknown coordinate values xo12, yo12, xo43, and yo43, the position coordinates (Xo, Yo) of the partial region Ro corresponding to the pixel P (X, Y) in the divided region Rr (n) are expressed as follows: It can be obtained by a calculation formula.
<Calculation of point (xo12, yo12)>
xo12 = (xo2-xo1). (X-x1) / (x2-x1) + xo1
yo12 = (yo2-yo1). (X-x1) / (x2-x1) + yo1
<Calculation of point (xo43, yo43)>
xo43 = (xo3-xo4). (X-x1) / (x2-x1) + xo4
yo43 = (yo3-yo4). (X-x1) / (x2-x1) + yo4
∴
<Calculation of position coordinates (Xo, Yo) corresponding to point P>
Xo = (xo43−xo12) · (Y−y1) / (y2−y1) + xo12
Yo = (yo43−yo12) · (Y−y1) / (y2−y1) + yo12

  The control unit 11 functions as a pixel value extracting unit, extracts the pixel value of the pixel Po in the partial region Ro located at the position coordinate (Xo, Yo) corresponding to the point P (step S16), A pixel value of the extracted pixel Po is assigned (step S17).

  Then, it is determined whether or not pixel values have been assigned to all the pixels in the divided region Rr (n) (step S18), and pixels that have not yet been given a pixel value in the divided region Rr (n). If there is any (step S18: No), the pixel P (X, Y) is updated (step S19), and the steps until the pixel value assignment for all the pixels in the divided region Rr (n) is completed. The processes from S15 to S18 are repeated.

  On the other hand, when pixel value assignment has been completed for all the pixels in the divided region Rr (n) (step S18: Yes), pixel value assignment processing is performed for all the divided regions Rr in the conversion destination image. It is determined whether or not the process has been completed (step S20). When there is a divided region Rr that has not been processed (step S20: No), 1 is added to the variable “n” (step S21), and the conversion destination Steps S14 to S20 are repeated until the processing is completed for all the divided regions Rr of the image.

  When the pixel value assignment processing has been completed for all the divided regions Rr of the conversion destination image (step S20: Yes), the control unit 11 functions as converted image creation means, and gives a pixel value to each pixel. A conversion destination image is generated based on each divided area Rr, the conversion destination image is stored in the conversion destination image storage unit 123 as conversion destination image data, and the process ends (step S22).

  As described above, according to the present embodiment, the division grid for dividing the conversion source image into a plurality of partial regions Ro and the division grid for defining the conversion destination image as at least one or more division regions Rr. Of the plurality of partial regions Ro, the partial region Ro selected as the partial region Ro to be subjected to image conversion is associated with the divided region Rr, and each divided region Rr is divided into the corresponding partial regions Ro. Since the position coordinate (Xo, Yo) corresponding to the pixel P constituting the region Rr is calculated and the pixel value of the pixel Po located at the position coordinate is assigned to the pixel P, each partial region Ro is assigned. On the other hand, different image conversions (enlargement / reduction, etc.) can be performed quickly and reliably in a single process.

  In addition, since the divided grid is generated with grid lines and the intersections of the grid lines can be corrected, the partial areas Ro and the divided areas Rr of various sizes and shapes can be determined, and image conversion processing with a high degree of freedom is possible. Can be realized.

  In the image conversion process described above, the step of selecting the partial area Ro (n) for which image conversion is desired is performed in “execution of the image conversion process” among the partial areas Ro divided by the division grid of the conversion source image. (Step S11) equalizes the number of divided regions Rr defined by the divided grid of the conversion destination image and the number of partial regions Ro (n) selected as the partial region Ro (n) for which image conversion is desired. Therefore, when the divided grid of the conversion source image is generated in the divided grid generation step in step S2, the partial area Ro (n) for which image conversion is desired is selected, and the selected partial area Ro is selected. This information may also be stored in the divided grid storage unit 122 together with the divided grid.

[4. Modified example]
4-1. Modification 1
The calculation of the position coordinates (Xo, Yo) in step S15 of the image conversion process described above has been described for the case where the divided region Rr is a quadrangle with all four corners being 90 degrees. The calculation of the position coordinates (Xo, Yo) when Ro is a quadrangle with all four corners being 90 degrees will be described.

  14A and 14B are explanatory diagrams of the divided region Rr and the partial region Ro corresponding to the divided region Rr.

  The coordinates (xo1, yo1), (xo2, yo2), (xo3, yo3), (xo4, yo4) of the partial area Ro, the coordinates (x1, y1), (x2, y2) of the four vertices of the divided area Rr ), (X3, y3), (x4, y4) are known values due to the creation of the divided grid.

Therefore, by obtaining the unknown coordinate values x12, y12, x43, y43, the position coordinates (Xo, Yo) of the partial region Ro corresponding to the pixel P (X, Y) in the divided region Rr (n) are It can be obtained by a calculation formula.

<Calculation of point (x12, y12)>
x12 = (x2-x1). (Xo-xo1) / (xo2-xo1) + x1 (Formula 1)
y12 = (y2-y1). (Xo-xo1) / (xo2-xo1) + y1 (Formula 2)

<Calculation of point (x43, y43)>
x43 = (x3-x4). (Xo-xo1) / (xo2-xo1) + x4 (Formula 3)
y43 = (y3-y4). (Xo-xo1) / (xo2-xo1) + y4 (Formula 4)

<Coordinate value of pixel P (X, Y)>
X = (x43-x12). (Yo-yo1) / (yo2-yo1) + x12 (Formula 5)
Y = (y43-y12). (Yo-yo1) / (yo2-yo1) + y12 (Formula 6)
∴
Yo = (X-x12) (yo2-yo1) / (x43-x12) + yo1 (Formula 7)
Yo = (Y−y12) (yo2−yo1) / (y43−y12) + yo1 (Formula 8)
∴
(Y-y12) (x43-x12) = (X-x12) (y43-y12) (Formula 9)

Then, (Equation 1) to (Equation 4) are substituted into (Equation 9),

{Y- (y2-y1). (Xo-xo1) / (xo2-xo1) + y1} {(x3-x4). (Xo-xo1) / (xo2-xo1) + x4- (x2-x1). (Xo -Xo1) / (xo2-xo1) -x1}
= {X- (x2-x1). (Xo-xo1) / (xo2-xo1) + x1}
{(Y3-y4). (Xo-xo1) / (xo2-xo1) + y4- (y2-y1). (Xo-xo1) / (xo2-xo1) -y1}

{(Y + y1) (xo2-xo1)-(y2-y1). (Xo-xo1)} {(x3-x4). (Xo-xo1)-(x2-x1). (Xo-xo1) + (x4- x1) (xo2-xo1)}
= {(X + x1) (xo2-xo1)-(x2-x1). (Xo-xo1)}
{(Y3-y4). (Xo-xo1)-(y2-y1). (Xo-xo1) + (y4-y1) (xo2-xo1)}

{(Y + y1) (xo2-xo1)-(y2-y1). (Xo-xo1)} {(x3-x4-x2 + x1). (Xo-xo1) + (x4-x1) (xo2-xo1)}
= {(X + x1) (xo2-xo1)-(x2-x1). (Xo-xo1)}
{(Y3-y4-y2 + y1). (Xo-xo1) + (y4-y1) (xo2-xo1)}

here,

x = Xo-xo1
C1 = (Y + y1) (xo2-xo1)
C2 = (y2-y1)
C3 = (x3-x4-x2 + x1)
C4 = (x4-x1) (xo2-xo1)
C5 = (X + x1) (xo2-xo1)
C6 = (x2-x1)
C7 = (y3-y4-y2 + y1)
C8 = (y4-y1) (xo2-xo1)

After all,

(C1-C2x) (C3x + C4) = (C5-C6x) (C7x + C8)
(C6 · C7−C2 · C3) x ² + (C1 · C3 · C2 · C4−C5 · C7 + C6 · C8) x + C1 · C4−C5 · C8 = 0

Furthermore,

K1 = C6 ・ C7−C2 ・ C3
K2 = C1, C3-C2, C4-C5, C7 + C6, C8
K3 = C1, C4-C5, C8

Then, a quadratic equation for the unknown x is obtained.

Therefore, the solution Xo is

Xo = {− K2 ± (K22−4K1 · K3)} / (2K1) + xo1 (Formula 10)

And one of the two solutions included in the region is determined as Xo.

  Then, based on the obtained Xo, the values of x12, y12, x43, and y43 are determined using (Expression 1) to (Expression 4), and Yo is determined using (Expression 7) or (Expression 8). decide.

4-2. Modification 2
Although the case where either one of the partial region Ro or the divided region Rr is a quadrangle whose squares are all 90 degrees has been described, as a second modification, as shown in FIGS. 15A and 15B, The calculation of the position coordinates (Xo, Yo) in the case where both the partial region Ro and the divided region Rr are not quadrangles whose four corners are all 90 degrees will be described.

  In the case of the present embodiment, the position coordinates are calculated by interposing a quadrilateral having 90 degrees of all four corners as the mediation region Rm as shown in FIG. 16 (A) or (B). FIG. 16A is an explanatory diagram for replacing (mediating) the divided region Rr (FIG. 15B) with the mediation region Rm, and FIG. 16B illustrates the partial region Ro (FIG. 15A) as the mediation region. It is explanatory drawing replaced (mediated) by Rm.

  When the divided region Rr is replaced with the mediation region Rm (FIG. 16A), Po, Xo, Yo, xo1, yo1, xo2, and yo2 in FIG. 14 of Modification 1 are respectively changed to Pm, Xm, and Ym. , Xm1, ym1, xm2, ym2 and (Xm, Ym) is determined using (Expression 1) to (Expression 10).

  On the other hand, when replacing the partial region Ro with the mediation region Rm (FIG. 16B), P, X, Y, x1, y1, x2, and y2 in FIG. 13 are respectively replaced with Pm, Xm, Ym, and xm1. , Ym1, xm2, ym2 and based on the value of (Xm, Ym) determined in FIG. 16A, <calculation of point (xo12, yo12)> in step S15> <point (xo43, yo43) )> <Xo, Yo) is determined by using the following formulas: <Calculation of position coordinates (Xo, Yo) corresponding to point P>.

4-3. Modification 3
In the above-described embodiment, grid lines are generated in the horizontal direction and the vertical direction of the conversion source image (or the conversion destination image), and the division grid is generated so that the partial area Ro and the division area Rr have a rectangular shape. As shown in FIG. 17, a division grid may be used in which the partial region Ro and the division region Rr are triangular.

  In this case, each vertex of the partial region Ro (or divided region Rr) is shared with six adjacent partial regions Ro (or divided regions Rr), and each side of the partial region Ro (or divided region Rr) is adjacent. It is shared with three partial areas Ro (or divided areas Rr).

  FIG. 18 is an explanatory diagram of a right triangle triangular divided region Rr and a triangular (not right triangle) partial region Ro corresponding to the divided region Rr.

  The coordinates (xo1, yo1), (xo2, yo2), and (xo3, yo3) of the three vertices of the partial region Ro are set, and the coordinates (x1, y1), (x2, y1), (x1, y1), and (x1, y1) of the four vertices of the divided region Rr are set. y2) is a known value due to the creation of a divided grid.

Therefore, by obtaining the unknown coordinate values xo12, yo12, xo13, and yo13, the position coordinates (Xo, Yo) of the partial region Ro corresponding to the pixel P (X, Y) in the divided region Rr (n) are obtained as follows. It can be obtained by a calculation formula.

<Calculation of point (xo12, yo12)>
xo12 = (xo2-xo1). (X-x1) / (x2-x1) + xo1
yo12 = (yo2-yo1). (X-x1) / (x2-x1) + yo1

<Calculation of point (xo13, yo13)>
xo13 = (xo3-xo1). (Y-y1) / (y2-y1) + xo1
yo13 = (yo3-yo1). (Y-y1) / (x2-x1) + yo1

<Calculation of position coordinates (Xo, Yo) corresponding to point P>
Xo = xo13 + xo12-xo1
Yo = yo13 + yo12−yo1

  FIG. 19 is an explanatory diagram of a triangular (not a right triangle) divided region Rr and a right triangular partial region Ro corresponding to the divided region Rr.

  The coordinates (xo1, yo1), (xo2, yo1), (xo1, yo2) of the three vertices of the partial area Ro are set, and the coordinates (x1, y1), (x2, y2), (x3, x4) of the four vertices of the divided area Rr are set. y3) is a known value due to the creation of a divided grid.

Therefore, by obtaining the unknown coordinate values xo12, yo12, xo13, and yo13, the position coordinates (Xo, Yo) of the partial region Ro corresponding to the pixel P (X, Y) in the divided region Rr (n) are obtained as follows. It can be obtained by a calculation formula.

<Calculation of point (x12, y12)>
x12 = (x2-x1). (Xo-xo1) / (xo2-xo1) + x1 (Formula 11)
y12 = (y2-y1). (Xo-xo1) / (xo2-xo1) + y1 (Formula 12)

<Calculation of point (x13, y13)>
x13 = (x3-x1). (Yo-yo1) / (yo2-yo1) + x1 (Formula 13)
y13 = (y3-y1). (Yo-yo1) / (yo2-yo1) + y1 (Formula 14)

<Coordinate value of pixel P (X, Y)>
X = x13 + x12-x1 (Formula 15)
Y = y13 + y12−y1 (Formula 16)
When (Expression 11) to (Expression 14) are substituted into (Expression 15) and (Expression 16),
X = (Yo-yo1) (x3-x1) / (yo2-yo1) + x1
+ (Xo-xo1) (x2-x1) / (xo2-xo1) (Formula 17)
Y = (Yo-yo1) (y3-y1) / (yo2-yo1) + y1
+ (Xo-xo1) (y2-y1) / (xo2-xo1) (Formula 18)

A binary simultaneous equation for the unknowns Xo and Yo is obtained.

here,

x = Xo-xo1
y = Yo-yo1
C1 = (X-x1)
C2 = (x3-x1) / (yo2-yo1)
C3 = (x2-x1) / (xo2-xo1)
C4 = Y-y1
C5 = (y3-y1) / (yo2-yo1)
C6 = (y2-y1) / (xo2-xo1)

After all,

C1 = C2y + C3x
C4 = C5y + C6x

The solution Xo, Yo is Xo = (C1 · C5−C2 · C4) / (C3 · C5−C2 · C6) + xo1 (Equation 19)
Yo = {C1-C3 (C1, C5-C2, C4) / (C3, C5-C2, C6)} / C2 + yo1 (Formula 20)
As obtained.

  Next, as shown in FIG. 20, calculation of position coordinates (Xo, Yo) when both the partial region Ro and the divided region Rr are not right triangles will be described.

  In the case of the present embodiment, the position coordinates are calculated by interposing a right triangle intermediary region Rm as shown in FIG. 21 (A) or (B). FIG. 21A is an explanatory diagram for replacing (mediating) the divided region Rr (FIG. 20B) with the mediation region Rm, and FIG. 21B illustrates the partial region Ro (FIG. 20A) as the mediation region. It is explanatory drawing replaced (mediated) by Rm.

  When replacing the divided region Rr with the mediation region Rm (FIG. 21A), Po, Xo, Yo, xo1, yo1, xo2, and yo2 in FIG. 19 are respectively replaced with Pm, Xm, Ym, xm1, and ym1. , Xm2, ym2, and (Xm, Ym) is determined using (Expression 11) to (Expression 20).

  Further, when replacing the partial region Ro with the mediation region Rm (FIG. 21B), P, X, Y, x1, y1, x2, and y2 in FIG. 13 are respectively replaced with Pm, Xm, Ym, and xm1. , Ym1, xm2, ym2 and based on the value of (Xm, Ym) determined in FIG. 21A, <calculation of point (xo12, yo12)> in step S15> <point (xo43, yo43) )> <Xo, Yo) is determined by using the following formulas: <Calculation of position coordinates (Xo, Yo) corresponding to point P>. At this time, xm1, ym1, xm2, and ym2 are the same as those in FIG.

4-4. Modification 4
A case where the conversion source image is a moving image composed of a plurality of image frames will be described.

  FIG. 22 is an explanatory diagram of a moving image (conversion source image) composed of a plurality of image frames.

  First, the control unit 11 functions as a representative determining unit, determines (selects) one representative image frame from a plurality of image frames, and generates a divided grid of the representative image frame. The representative image frame is preferably the first image frame f1 as shown in FIG.

  Then, a converted image frame corresponding to the representative image frame is created. The creation method is the same as in the above embodiment.

  At this time, the division grid of the representative image frame and the division grid of the converted image frame corresponding to the representative image frame are stored (saved) in the division grid storage unit 122 of the storage unit 12 (storage unit).

  When image conversion is performed on the other image frames f2, f3, and f4, the divided grid of the representative image frame stored in the divided grid storage unit 122 and the converted image corresponding to the representative image frame. Using the frame division grid, converted image frames corresponding to the other image frames f2, f3, and f4 are created.

  Therefore, since the division grid used for image conversion of the representative image frame can be used as it is for image conversion of other image frames, the conversion source image is a moving image image composed of a plurality of image frames. Even if it exists, it can image-process rapidly with respect to each image frame, and can produce | generate the image for conversion moving images based on each created conversion image frame.

  As another modified example, as shown in FIG. 23, divided grids of the first image frame f1 and the last image frame f4 are generated from among a plurality of image frames, and intermediate image frames are applied by applying these divided grids. It can also be configured to generate f2 and f3 split grids.

  FIG. 24 is an explanatory diagram for generating a divided grid of an intermediate image frame.

  This figure is an example when the deformation of the divided grid increases with time. The divided grids of the intermediate image frames f2 and f3 are generated by linear interpolation of the divided grid of the first image frame f1 and the divided grid of the last image frame f4.

More specifically, the vertex coordinates of the divided grid of the first image frame f1 of the n image frames are (x1, y1), and the vertex coordinates of the divided grid of the last image frame fn are (xn, yn). The vertex coordinates of the divided grid of the intermediate image frame fm (2 ≦ m <n) are
[X1 + (xn-x1) (m-1) / (n-1)
, Y1 + (yn−y1) (m−1) / (n−1)]
It becomes.

  In this way, the converted image frame corresponding to the intermediate image frame is created using the divided grids of the first and last image frames and the linear interpolation grid obtained by linearly interpolating these divided grids. Then, a converted moving image is created based on each created converted image frame.

  Therefore, it is possible to create a converted moving image image in which the image is gradually deformed with the converted image time.

  As described above, the present invention can perform different image processing (for example, image extraction, trimming, zoom out, zoom up, enlargement, reduction, etc.) for each partial area of the conversion source image. Graphic processing for still image production such as printed materials, Web images, presentation images (shooting / scanning image correction, layout correction, special effects, etc.), image processing such as angle correction, special effects, etc. for CG / video production When applied to, particularly remarkable effects can be obtained.

It is a block diagram which shows the structure of the image converter which concerns on this invention. It is explanatory drawing of the division | segmentation grid of a conversion source image. It is explanatory drawing of the division | segmentation grid of a conversion source image, and the division | segmentation grid of a conversion destination image. It is explanatory drawing of division | segmentation grid production | generation. It is explanatory drawing of a correction process. FIG. 6 is a diagram for explaining association between a partial region Ro and a divided region Rr. It is an example of another division grid and image conversion. It is an example of another division grid and image conversion. It is an example of another division grid and image conversion. It is a specific example of image conversion. It is a specific example of image conversion. It is a flowchart which shows an image conversion process. It is explanatory drawing of division area Rr and the partial area Ro corresponding to the said division area Rr. It is explanatory drawing of the division area Rr concerning the modification 1, and the partial area Ro corresponding to the said division area Rr. It is explanatory drawing of the division area Rr concerning the modification 2, and the partial area Ro corresponding to the said division area Rr. It is explanatory drawing of position coordinate (Xo, Yo) calculation via the mediation area | region Rm concerning the modification 2. It is explanatory drawing of the division | segmentation grid of the conversion source image concerning the modification 3, and the division | segmentation grid of a conversion destination image. It is explanatory drawing of the division area Rr concerning the modification 3, and the partial area Ro corresponding to the said division area Rr. It is explanatory drawing of the division area Rr concerning the modification 3, and the partial area Ro corresponding to the said division area Rr. It is explanatory drawing of the division area Rr concerning the modification 3, and the partial area Ro corresponding to the said division area Rr. It is explanatory drawing of position coordinate (Xo, Yo) calculation via the mediation area | region Rm concerning the modification 3. It is explanatory drawing of the image for moving images (conversion original image) comprised from several image frames. It is explanatory drawing of the division | segmentation grid of the first image frame f1 and the last image frame f4. It is explanatory drawing of the division | segmentation grid generation of an intermediate | middle image frame.

Explanation of symbols

1 Image converter
11 Control Unit 12 Storage Unit 13 Input Unit 14 Display Unit 15 Bus Ro Partial Area Rr Divided Area

Claims (8)

In an image conversion apparatus that creates a converted image by extracting a predetermined region included in an original image,
A divided grid generating means for generating a first divided grid for dividing the original image into a plurality of partial areas, and generating a second divided grid for defining the converted image as at least one divided area;
Selecting means for selecting at least one or more partial areas for image conversion from among the plurality of partial areas divided by the first divided grid;
For each of the divided areas, area correspondence determining means for associating with the selected partial area;
For each of the divided areas, the position coordinates of the partial areas corresponding to the pixels constituting the divided area are calculated using the partial areas corresponding to the divided areas, and pixels included in the partial area are Pixel value extraction means for extracting the pixel value of the pixel located at the calculated position coordinate;
For each of the divided areas, the pixel value extracted by the pixel value extracting unit is assigned as the pixel value of each pixel constituting the divided area, and the converted image is created by creating the converted image including the divided areas. Means,
An image conversion apparatus comprising: The image conversion apparatus according to claim 1,
The divided grid generation means generates the first divided grid and the second divided grid using a plurality of grid lines, and sets the intersection of the grid lines and the outline of the divided grid within a range where the grid lines do not intersect. An image conversion apparatus comprising: a correction unit that corrects the shape of the partial area and the divided area by changing a position of a contact point between the grid line to be displayed and each grid line. The image conversion apparatus according to claim 1 or 2,
An image conversion apparatus characterized in that the number of partial areas selected by the selection means is equal to the number of divided areas of the converted image. In the image conversion device according to any one of claims 1 to 3,
The number of partial areas of the original image divided by the first divided grid is equal to the number of divided areas of the converted image defined by the second divided grid, and all the parts are selected by the selection unit. An image conversion apparatus characterized in that a region is selected. In the image conversion device according to any one of claims 1 to 4,
When the original image is a moving image composed of a plurality of image frames, the image processing apparatus includes a representative determining unit that determines a representative image frame from the plurality of image frames.
The divided grid generation means generates the first divided grid for dividing the determined representative image frame into a plurality of partial regions, and at least one or more of the divided image frames corresponding to the representative image frame Generating the second divided grid for defining as a region;
Storage means for storing a first divided grid and a second divided grid according to the representative image frame generated by the divided grid generation means;
When creating the converted image frame corresponding to the representative image frame and the other image frames, the converted image creating means uses the first divided grid and the second divided grid stored in the storage means, An image conversion apparatus that generates converted image frames corresponding to the representative image frame and the other image frames, and generates a converted moving image based on the generated converted image frames. The image conversion apparatus according to claim 5, wherein
The image conversion apparatus characterized in that the representative determining means determines the first image frame as the representative image frame among the plurality of image frames constituting the moving image. The image conversion apparatus according to claim 6, wherein
The divided grid generation means generates the first divided grid for dividing the last image frame into a plurality of partial areas among the plurality of image frames constituting the moving image, and the last image frame Generating the second divided grid for defining the converted image frame corresponding to as at least one or more of the divided regions;
The storage means stores a first divided grid and a second divided grid relating to the last image frame generated by the divided grid generating means,
When the converted image creating means creates a converted image frame corresponding to an image frame other than the representative image frame and the last image frame, the converted image creating means applies a first to the representative image frame stored in the storage means. A divided grid, a linear interpolation grid obtained by linearly interpolating the first divided grid for the last image frame, a second divided grid for the representative image frame stored in the storage unit, and a first for the last image frame. A conversion image frame corresponding to the other image frame is created using a linear interpolation grid obtained by linearly interpolating a two-part grid, and a converted moving image is generated based on each of the created conversion image frames. An image conversion apparatus characterized by being created.   An image conversion program for causing a computer to function as the image conversion apparatus according to any one of claims 1 to 7.

Images