JP2976512B2 - 2D image transformation device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial applications]

The present invention relates to an image transformation device that changes the shape of a two-dimensional bitmap image.

[Prior art]

Various types of deformation functions for two-dimensional bitmap images, such as enlargement / reduction and rotation, are considered depending on the application and purpose, and many methods have been proposed for realizing those functions. One of the functions is a function to parse a two-dimensional plane image. This is necessary when the image is deformed so as to have a shape when viewed from various viewpoints in the three-dimensional space of the two-dimensional image, or when the two-dimensional image is deformed so as to be pasted and synthesized with a three-dimensional object. Secondary projection transformation is known as a general method for realizing this function, and its software has been created in the image processing subroutine package "SPIDER" developed by the Electronic Technology Research Institute of the National Institute of Advanced Industrial Science and Technology. ing. Although the secondary projective transformation does not use three-dimensional information in the calculation formula, it is calculated based on the arrangement of images in a three-dimensional space, so that an accurate perspective can be expressed. However, in addition to this accurate calculation method, there has been a demand that the parsing should be artificially performed by calculation using only two-dimensional information. In addition, in order to make various types of deformable shapes, not only give the perspective a real thing,
It was considered that a function that allows the user to set the condition of the perspective was desirable.

[Problems to be solved by the invention]

The present invention provides various deformations of a two-dimensional image such as a deformation that parses a two-dimensional bitmap image and a deformation that attaches a two-dimensional image to a three-dimensional image such as a sphere or a cylinder in a two-dimensional plane. It is an object of the present invention to provide an image deforming apparatus capable of adjusting the degree of a three-dimensional effect on an image having the three-dimensional shape.

[Means for Solving the Problems]

The two-dimensional image deforming apparatus according to the present invention includes a pixel density changing unit (first unit) that changes a pixel density of an original image with a predetermined function.
FIG. 1301), an image shape changing means for changing the shape of the image processed by the pixel density changing means (FIG. 1302), and the predetermined image density adjusting means for adjusting the rate of image density change by the pixel density changing means. Input means (FIG. 1) for inputting an indication point for determining a coefficient of the function.

[Action]

In the two-dimensional image deforming apparatus of the present invention, the pixel density changing means performs a process for giving a sense of depth to the image by changing the pixel density of the original image with a predetermined function. The image shape changing means performs a process of giving a sense of distance to the entire shape of the image. By combining the processing of the density and the external shape, it is possible to transform the image into a three-dimensional image. Point information indicating the degree of the three-dimensional effect is input by the input unit, and the coefficient of the function of the pixel density changing unit is adjusted based on the input information. By adjusting the designated point, the stereoscopic effect can be adjusted.

【Example】

 First Embodiment In this embodiment, a horizontal original image is
And the pixel density in both the vertical and vertical directions
The target shape is changed continuously, and the external shape is further deformed.
Non-rectangular shape using pseudo-affine transformation
The shape of the image is transformed by converting
Related to a device for causing the same. FIG. 2 is an example of an image before and after the deformation processing, and FIG.
(A) is an original image which is a rectangle before deformation, and FIG.
This is a final deformed image with a source. Fig. 2
Where 21, 22, 23, and 24 are the vertices P at the four corners of the original image₁, P_Two, P_Three, P_FourIn
25,26,27,28 are the vertices P of the original image₁, P_Two, P_Three, P_FourCorresponding to
Vertex Q at the four corners of the transformed image₁, Q_Two, Q_Three, Q_ThreeIt is. Second
A method for performing the modification shown in the figure will be described. First, a two-dimensional original image is created without using three-dimensional information.
The principle for three-dimensional deformation will be briefly described. Second
Compare the deformed image of FIG. 2B with the original image of FIG. 2A.
It can be seen that the outer shape of the entire image has changed. Also
Although not shown in FIG. 2, the pixel density in the image also changes.
You. That is, 25 points Q in the deformed image of FIG.₁
Is the point closest to the viewpoint in 3D space, 27 points Q_ThreeConversely
The farthest point can be considered,
If the distances are different, the pixel density differs between them.
So we decided to use these two properties to transform
I do. Fig. 3 shows an original rectangle that takes advantage of these two properties.
Shows how the image is transformed from perspective to perspective
This is an example. FIG. 3 (a) is an original image which is a rectangle before deformation.
It is an image and is the same as FIG. 2 (a). Fig. 3
(B) shows that the pixel density in both the horizontal and vertical directions is continuously changed.
To give the difference in pixel density inside the image.
Things. In other words, it gives a sense of depth. Fig. 3
(C) shows the entire outer shape of the image subjected to the pixel density change.
FIG. 2 (b) is a final deformed image. Next, the operation content for performing image deformation will be described.
You. FIG. 4 shows an image given by the operator before and after the deformation process of FIG.
It is the one that added the indicated point. Original image of Fig. 4 (a)
The selection of the two vertices located on the diagonal of the rectangular area, ie, the point
P₁And P_ThreeOr point P_TwoAnd P_FourOf FIG. 4 (b)
In the image, 4 vertices P of the original image₁, P_Two, P_Three, P_Four4 vertices Q corresponding to
₁, Q_Two, Q_Three, Q_FourAnd two points Q_Five, Q₆And instruct and give. Point Q_Five
Is two points P of the original image₁And P_TwoCorresponding to a point in the middle of
Then, the side Q of the deformed image₁Q_FourSide Q_TwoQ_ThreePixels in the direction toward the side
This specifies the degree of change in density. Also point Q₆Is the original picture
Two points P of the image₁And P_FourCorresponding to the point at the middle position of
Image side Q₁Q_TwoSide Q_FourQ_ThreePixel density changes in the direction
It specifies the condition. 2 points Q_Five, Q₆Is a perspective tool
In other words, to adjust the sense of depth,
By arranging this point appropriately, the pixel distribution inside the image
Can be changed to some extent, diversifying the deformed shape
be able to. Now we will explain how to calculate the coordinates to get the deformed image
I do. First, a process of changing the pixel density is performed. Figure 5 is a drawing
It is an image before and after the process of changing the density. Fig. 5
(A) is an image before performing this processing, that is, FIG.
(A) is the original image, and (b) in FIG.
It is an image after having made it. In Fig. 5, 51,52,53,54 are original images
Vertex P at the four corners of₁, P_Two, P_Three, P_FourAnd 55,56,57,58 are original images
Each vertex P of₁, P_Two, P_Three, P_FourAfter the pixel density change processing corresponding to
Four vertices R of the image₁, R_Two, R_Three, R_FourIt is. In this process the original image
Pixel density in the horizontal and vertical directions
You. Each coordinate calculation formula for performing the conversion process is
Equations (1) and (2) are shown. Horizontal direction x '= Ax^Two+ Bx + C (1) Vertical direction y '= Dy^Two+ Ey + F (2) where x and y are horizontal and vertical coordinates before the pixel density change processing.
X 'and y' are coordinates after processing. Where equation (1)
The coefficients A, B, and C in equation (1) are input coordinates x of 0,
_1xP_6x|, | P_1xP_4xWhen the value of |, the value of the output coordinate variable x 'is
0, | Q₁Q₆|, | Q₁Q_FourIt is a value that satisfies | P₁Is the point P
₁Represents the horizontal coordinate value of. | P_1xP_6x| Is side P₁P₆Length of
It is. | Q₁Q₆| Is side Q₁Q₆Is the length of Also, in equation (2)
Coefficients D, E, and F are input coordinate coefficients y in equation (2) are 0, | P_1yP
_5y|, | P_1yP_2yWhen the value of |, the value of the output coordinate variable y 'is 0, | Q
₁Q_Five|, | Q₁Q_TwoIt is a value that satisfies | Before calculating equations (1) and (2), lower left of the original image
Corner vertex P₁Original image so that is the origin position on the xy plane coordinates
It is necessary to keep the image in place. After performing the coordinate calculation to change the pixel density,
Calculate coordinates to change. Figure 6 shows the whole image
It is an image before and after the process of changing the outer shape into a non-rectangular shape. No.
FIG. 6A shows an image before deformation in this processing.
FIG. 6 (b) is the same as the image shown in FIG.
The images with the perspective obtained in FIG.
63 and 64 are the vertices R at the four corners before this deformation₁, R_Two, R_Three, R_FourAnd
65,66,67,67 is each vertex R₁, R_Two, R_Three, R_FourVertices of the four corners corresponding to
Q₁, Q_Two, Q_Three, Q_FourIt is. In this conversion, the pixels inside the image
Use a pseudo-affine transformation located at the shape location. here
The calculation formulas for the conversion process in Eq. (3) and (4) are respectively
Show.  In the formula, x 'and y' are the horizontal and vertical coordinates before the shape change processing.
X and Y are coordinates after processing. x 'and y' are (1)
The output coordinates of equations (2) and (3) are used as inputs. Outline change processing
Is finished, the vertex Q of the deformed image₁Is the coordinate value on the xy plane coordinates
Q₁The deformed image is arranged so as to come to the position. As described above, the process of changing the pixel density and the overall
Deformation is performed by combining with further processing. No.
According to FIG. 3, the image with the density changed by the former process is
And then tell the user to make a shape change to that image.
, But without actually generating an intermediate image, one pixel
, The coordinates of equations (1) to (4) are continuously calculated. In this embodiment, which performs image deformation based on the principle described above,
FIG. 8 shows an example of the hardware configuration. Smell
81 is a computer, 82 is a central processing unit, 83 is image transformation
Processing unit, 84 is image memory, 85 is input device, 86 is display
Ray, 87 is a scanner, 88 is a printer, 89 is a bus.
The input device 85 is for giving commands and the like to the computer.
You. The display 86 outputs and displays an image.
The scanner 87 is for inputting an image, and the printer 88 is for inputting an image.
Output. The bus 89 stores image data and control information.
Play the role of delivery. The image memory 84 stores image data
The central processing unit 82 is an image processing device.
It controls the entire device and performs general calculations. The image deformation processing unit 83 including the configuration characteristic of the present invention
It has the configuration shown in FIG. That is, the image transformation processing unit 83
Are the coordinates of the representative points of the image before and after
Initial deformation value input section for inputting information on the size and shape of the switch
11 and the procedure for selecting a small plane patch from the input initial value
Control and selection of integer coordinate points in the deformed facet patch
The trace control unit 12 that controls the
Deformation point that calculates the coordinates after transformation for the coordinates of the vertices
The target calculation unit 13 and the selected facet patch and deformed facet patch
Inverse mapping to find the inverse mapping function of the transformation function from the vertex coordinates
The image function calculator 14 calculates the integer coordinate points in the deformed facet patch.
Integer coordinate detection unit 15 to detect and detection using inverse mapping function
To calculate the coordinates on the original image from the coordinates of the integer coordinate points
The mapping coordinate calculation unit 16 performs interpolation calculation from the calculated coordinate values.
Interpolation processing unit 17 that calculates new pixel values on the image after the transformation
And pixel input / output to control writing and reading of pixel values
It comprises a force control unit 18. 19 is the image data before and after deformation
This is an image memory for storing data. In the above configuration, coordinate calculation from reading pixel values
About the entire processing flow from writing to writing
explain. FIG. 7 shows the flow. (701) The initial value required for deformation
Enter period information. Information is representative of the image before and after deformation
There are coordinates of the designated point. Positioned on the diagonal of the rectangular area in the original image
A set of two points P₁, P_Three, Four corner vertices Q in the deformed image₁, Q_Two, Q
_Three, Q_FourAnd Q to specify the condition of perspective_Five, Q₆Enter
Also specify the size and shape of the small face patch to be divided here
I do. (702) The trace control unit 12 changes the original image area to Kodaira
Divide into surface patches and start tracing the patches. Deformed
Coordinate calculation is not performed for all pixels on the original image.
Target the vertices of the divided patch. (703) The trace control unit 12 further executes the trace
Select facet patches in order. (704) The pixel density changing unit 130 in the deformed coordinate calculating unit 13
2 is the small plane patch selected by the trace control unit 12.
Calculates the coordinates of the vertex coordinates for changing the pixel density.
This is performed according to the equations (1) and (2). (705) The image shape changing unit 1302 calculates
Coordinate calculation for outer shape change processing is performed on the obtained coordinates.
Do. For this calculation, the above equations (3) and (4) are used.
You. (706) In the inverse mapping function calculation unit 14, at step 703
The coordinates of the vertices of the selected facet patch and the total of step 705
From the coordinates obtained by the calculation,
Find the inverse mapping function of the transformation (707) The trace control unit 12 obtains the
Integer coordinate point in the region for the deformed facet patch
Start tracing for. (708) The integer coordinate detection under the control of the trace control unit 12
Protrusion 16 is aligned by traces in the deformed facet patch area.
Select several coordinate points in order. (709) The inverse mapping coordinate calculation unit 16 selects the
Inverse mapping function found in step 706 for the integer coordinate points
Calculate inverse mapping coordinates using
You. (710) The interpolation processing unit 17 calculates the coordinates
The pixel values of one or more integer coordinate points in the vicinity are read. (711) Then, an interpolation process is performed based on the read pixel values.
To calculate a new pixel value. (712) The pixel input / output control unit 18 triggers the calculated new pixel value.
Write to the integer coordinate point in the deformed image selected by the race
No. (713) The trace control unit 12 is configured to generate the deformed small plane patch area.
It is determined whether or not the tracing has been completed. If yes
Proceeds to step 714, and if No, returns to step 708. (714) Trace of small plane patch on original image
It is determined whether or not has been completed. If Yes, end all processing
In the case of No, the process returns to step 703. The shape of the small plane patch to be divided in step 702
There is no limit on the size. Triangles, squares, etc.
I think. Also the interpolation method is not determined in step 711
However, the methods include the nearest neighbor method and four-point linear interpolation. these
Is related to processing speed and image quality,
And select it. According to the apparatus of the present embodiment, a two-dimensional bitmap image
Performs a transformation that adds a parse to the file using relatively simple processing
The perspective can be further specified by the user at the designated point (Q_Five, Q
₆) Allows easy adjustment of berth condition
Can be done. Second Embodiment This embodiment uses a cubic image for an original image having a rectangular shape.
Both horizontal and vertical directions by coordinate calculation using functional expressions
Continuously change the pixel density of horizontal and vertical
Change the width of the image in one direction along the circular shape
The shape of the image can be changed by changing the outer shape by coordinate calculation
It relates to a device for deforming the shape.
The roundness of the sphere can be changed by adjusting the indicated point.
There is a characteristic in the point that it is taken. Glue flat objects without wrinkles on spherical objects
You can't do it. So here like a globe
In the horizontal direction corresponding to the parallels, maintain the linearity, and
Principle of the deformation processing of this embodiment to make the vertical direction curved
Will be described. FIG. 9 is an example of an image after the deformation processing.
FIG. 9 (a) is an original image which is a rectangle before deformation, and FIG.
(B) is an image transformed into a spherical shape. Talk easy
As shown in Fig. 9, the target is the quadrant of the sphere.
In FIG. 9, reference numerals 91, 92, 93 and 94 denote 4 of the original image.
Corner vertex P₁, P_Two, P_Three, P_FourAnd 95,96,97 are the peaks of the original image
Point vertex P₁, P_Two, P_FourVertex Q of the transformed image corresponding to₁, Q_Two, Q_Four
It is. Vertex P_ThreeIs the vertex Q after deformation_TwoIs mapped above. Picture
Lines drawn inside the image represent the density distribution of the image
Things. The same applies to the following figures. First, without using the three-dimensional information of the two-dimensional original image
The principle for three-dimensional deformation is briefly described in ninth
Compare the deformed image in Fig. (B) with the original image in Fig. 9 (a).
Then, it can be seen that the outer shape of the entire image has changed. Ma
The density inside the image has also changed. That is, FIG.
95 points Q in the deformed image of (b)₁Is in three-dimensional space
Point closest to the viewpoint, 96 points Q_TwoAnd 97 dot Q_FourConversely
Can be considered the farthest point,
If the distances are different, the pixel density differs between them.
So we decided to use these two properties to transform
I do. Fig. 10 shows a rectangular element utilizing these two properties.
Demonstrating that the image is transformed into a spherical image in order
It is an example. FIG. 10 (a) shows an original image which is a rectangle before deformation.
And is the same as FIG. 9 (a). FIG. 10 (b)
Continuously changing pixel density in both horizontal and vertical directions
This is to give the difference in pixel density inside the image.
You. In other words, it gives a sense of depth that gives a sense of roundness. Fig. 10
(C) shows the entire outer shape of the image subjected to the pixel density change as a circle.
This is the same as the deformed image in FIG. 9 (b).
The same thing. Next, the operation content for performing image deformation will be described.
You. FIG. 11 shows the image before and after the deformation process of FIG.
It is the one that added the indicated point. Original image of Fig. 11 (a)
In the selection of, two vertices located on the opposite corners of the rectangular area, ie,
Point P₁And P_ThreeOr point P_TwoAnd P_FourInstruct. The change in FIG.
In the shape image, 3 vertices P of the original image₁, P_Two, P_Four3 vertices corresponding to
Q₁, Q_Two, Q_FourAnd two points Q_Five, Q₆And instruct and give. Deformed image
Two sides Q of the image₁Q_TwoAnd Q₁Q_Four2 points Q because the lengths are equal_Two, Q_FourHorse
Only one of the points may be designated.
No. Or vertex P_ThreeIs the vertex Q_TwoMapped onto
Vertex P_ThreeDo not give the designated point corresponding to
No. Point Q_FiveIs two points P of the original image₁And P_TwoCorresponding to a point in the middle of
The side Q of the deformed image₁Q_FourVertex Q from side_TwoOr edge
Q_TwoQ_FourSpecify the degree of change in pixel density in the direction toward
It is. Point Q at the same time₆Is the two points P of the original image₁And P_FourIntermediate position of
Of the deformed image₁Q_TwoVertex Q from side_Four
Or side Q_FourQ_TwoChange in pixel density in the direction toward
It is specified. In other words, the point Q in the world globe_FiveIs latitude
The point Q₆Is related to the longitude direction. 2 points Q_Five, Q₆Is a perspective tool
To adjust the roundness of the sphere.
However, by arranging this point appropriately, the density inside the image can be improved.
The degree distribution can be changed to some extent, and the deformed shape can be diversified.
Can be made. Generally, 2 points Q_Five, Q₆Is the vertex Q₁
Because it is equidistant from, only one point is specified
You can do it. Two directions, horizontal and vertical
2 points only if you want to make the pixel density of
I just want to. Now we will explain the coordinate calculation method for obtaining the deformed image.
You. First, a process of changing the pixel density is performed. Figure 12 shows the pixels
These are images before and after the process of changing the density. FIG. 12 (a)
The image before performing this processing, ie, the original image of FIG. 9 (a)
FIG. 12 (b) is an image after changing the pixel density.
It is a statue. 121, 122, 123, and 124 in FIG. 12 are the four corners of the original image.
Vertex P₁, P_Two, P_Three, P_FourWhere 125, 126, 127 and 128 are
Vertex P₁, P_Two, P_Three, P_FourOf the image after the pixel density change process corresponding to
4 vertex R₁, R_Two, R_Three, R_FourIt is. In this process, the horizontal
-Continuously change the pixel density in the two vertical directions. That
Equations (5) and (5) are used to calculate the coordinates for the conversion process.
It is shown in equation (6). Horizontal direction x '= Ax^Three+ Bx^Two+ Cx + D (5) Vertical direction y '= Ey^Three+ Fy^Two+ Gy + H (6) where x and y are horizontal and vertical coordinates before the pixel density change processing.
X 'and y' are coordinates after processing. Where equation (5)
The coefficients A, B, C, and D in are such that the input coordinate variable x in equation (5) is 0, |
P_1xP_6x|, | P_1xP_4xWhen the value of |, the value of the output coordinate variable x 'is
0, | Q_1xQ_6x|, | Q_1xQ_4x|
The value of the force coordinate variable x is | P_1xP_4xIn the case of |
This is the value when the condition that the first derivative value becomes 0 is satisfied. P
_1xIs the point P₁Represents the horizontal coordinate value of. | P_1xP_4x| Is side P₁P_Four
The length of | Q_1xQ_4x| Is side Q₁Q_FourIs the length of The derivative value 0 and
Is the edge R which is the end of the image_ThreeR_FourHorizontal pixel on the side
Is intended to be as small as possible. This makes the sphere
Pixels at the end become infinitely small when deformed to body shape
You. The coefficients E, F, G, and H in equation (6) are the input values of equation (6).
Force coordinate variable y is 0, | P_1yP_5y|, | P_1yP_2yWhen the value is |
The value of the coordinate variable y 'is 0, | Q_1yQ_5y|, | Q_1yQ_2y|
Is satisfied and the value of the input coordinate variable y is 0, | P_1yP_5y|
Satisfies the condition that the first derivative of the function of equation (6) is 0
This is the value when Before calculating Equations (5) and (6), lower left of the original image
Corner vertex P₁Original image so that is the origin position on the xy plane coordinates
It is necessary to keep the image in place. After performing the coordinate calculation to change the pixel density,
Same coordinate calculation to change. Figure 13 is outside the entire image
Before and after processing to change the shape to a non-rectangular circular shape
You. FIG. 13 (a) shows an image before deformation in this processing.
FIG. 13 (b) is the same as the image in FIG.
It is a spherical image finally obtained. 131, 132, 13 in Fig. 13
3,134 are the vertices R of the four corners before this shape change deformation₁, R_Two, R_Three, R_Fourso
Yes, 135, 136, 137 are vertex R₁, R_Two, R_FourVertex Q corresponding to₁,
Q_Two, Q_FourIt is. To transform from a rectangular shape to a circular shape
Zoom is applied along a circle from one direction of the shape. Paraphrase
Changes the width of the image so that it follows the circle. Total conversion process
The calculations are shown in equations (7) and (8), respectively. X = x '× f₁/ | Q_1x−Q_4x| (7) Y = y '(8) where x' and y 'are horizontal and vertical coordinates before the shape change processing.
X and Y are coordinates after processing. x 'and y' are (5)
The output coordinates of the equation and the equation (6) are used as inputs. f₁Is a circular expression
When the center point of the circle is placed at the origin of the xy plane,
This is expressed as in equation (9). x '= f₁(Y ') = SQRT (| Q_1xQ_4x|^Two-Y '²(9) In this embodiment, compression is performed in the horizontal direction.
Compression may be performed from the direct direction. When the shape change process is completed
Vertex Q of deformed image₁Is the coordinate value Q on the xy plane coordinates₁To the position
The deformed image is arranged as follows. As described above, the process of changing the pixel density and the overall
Deformation is performed by combining with further processing. No.
According to Fig. 10, the image obtained by adding the density
Obtain an image, and then perform shape change processing on the image
However, an intermediate image is not actually generated,
Coordinate calculation of equations (5) to (9) for one pixel continuously
I do. Deformation processing of two-dimensional image of the second embodiment described above
Is the same as that of the first embodiment.
FIG. 8 shows the overall configuration, and the configuration of the image deformation processing unit.
The result is shown in FIG. The processing contents of the deformation coordinate calculation unit 13 are different.
Otherwise, the basic configuration is the same as that of the first embodiment.
The operation of the apparatus according to the second embodiment is similar to the operation shown in FIG.
In the row, the coordinate meter for the pixel density changing process in step 704
The point of using equations (5) and (6) in the calculation, and step 705
Expressions (7) to (9) are used for the coordinate calculation of the outer shape change processing
This is different from the device of the first embodiment. According to this embodiment, the planar image is in a two-dimensional plane
Modified so that it can be pasted onto spherical images with simple operations
Shape and the user can specify the point (Q_Five, Q_FiveAdjust)
With this, the degree of roundness of the sphere can be changed. Third Embodiment In this embodiment, a tertiary image is formed on an original image having a rectangular shape.
Either horizontal or vertical by calculating coordinates using a function expression
Continuously changes the pixel density in one direction and uses a quadratic function.
The other one of horizontal and vertical above was calculated by the coordinates
The pixel density in the direction is continuously changed, and the pixel density is
Two round circles facing each other in the direction of the density change
Change the shape so that it is sandwiched between the curves
Continuous image width in the direction where the pixel density has been changed
By changing the rectangular planar image to a cylindrical
It relates to a device that transforms into a shape.
Is the three-dimensional effect around the cylindrical axis and the direction along the cylindrical axis
The feature is that the perspective can be changed by adjusting the designated point.
It has. FIG. 14 is an example of an image before and after the deformation processing, and FIG.
(A) is an original image that is a rectangle before deformation, and FIG. 14 (b) is a pasted image.
It is a cylindrical image to be attached. This second
14 The shape of the image in FIG.
You. To make the story easier to understand, use a cylinder drawing as shown in Fig. 14.
A description will be given of the right half of the image. Fig. 14
141,142,143,144 are the vertices P at the four corners of the original image₁, P_Two, P_Three, P_Four
145, 146, 147, 148 are each vertex P of the original image₁, P_Two, P_Three, P_Four
Vertices Q at the four corners of the transformed image corresponding to₁, Q_Two, Q_Three, Q_FourIn
You. Lines drawn inside the image represent the density distribution of the image
It is for The same applies to the following figures. First, a two-dimensional original image is created without using three-dimensional information.
The principle for three-dimensional deformation will be briefly described. 14th
Compare the deformed image in Fig. (B) with the original image in Fig. 14 (a).
Then, it can be seen that the outer shape of the entire image has changed. You
That is, the original image has a rectangular shape, and the deformation target image has a curved shape.
It is a rectangle including a shape. Also the density inside the image changes
I have. 145 points Q in the deformed image of FIG. 14 (b)₁Is three
It is the point closest to the viewpoint in the dimensional space.
146 points Q_TwoOf pixels along the cylindrical axis
Gradually gets smaller, 147 points Q_FourOf cylindrical shaft with
The size of pixels in the surrounding direction is decreasing at an accelerated rate
You. Therefore, there are two characteristics: differences in overall shape and differences in pixel density.
We will transform using quality. Figure 3 shows the two
From the original rectangular image to the cylindrical image
FIG. Fig. 15 (a)
Is an original image which is a rectangle before deformation, and is the same as FIG. 14 (a).
The same thing. Fig. 15 (b) shows the horizontal and vertical directions
The pixel density of the image is continuously changed.
This is for giving a difference in pixel density. In other words, cylindrical
It provides perspective along the axis and three-dimensionality around the cylindrical axis.
You. FIG. 15 (c) shows the image of FIG.
Fig. 14 shows the overall shape changed to a cylindrical shape.
This is the same as the deformed image in (b). Next, the operation content for performing image deformation will be described.
You. FIG. 16 shows the image before and after the deformation process of FIG.
It is the one that added the indicated point. Original image of Fig. 16 (a)
In the selection of, two vertices located on the opposite corners of the rectangular area, ie,
Point P1 and P3 or Point P_TwoAnd P_FourInstruct. The change in FIG. 16 (b)
In the shape image, 4 vertices P of the original image₁, P_Two, P_Three, P_FourFour peaks corresponding to
Point Q₁, Q_Two, Q_Three, Q_FourAnd two points Q_Five, Q₆And instruct and give. point
Q_FiveIs two points P of the original image₁And P_TwoCorresponding to a point in the middle of
So the side Q of the deformed image₁Q_FourVertex Q from side_TwoQ_ThreeFrom side Q_TwoQ_ThreeTo the side
Change in pixel density in the direction of orientation, ie along the cylinder axis
It specifies the condition. Also point Q₆Is the two points P of the original image₁
And P_FourOf the deformed image₁
Q_TwoVertex Q from side_FourQ_FiveTo the side, ie around the cylindrical axis
This specifies the degree of change in the pixel density in the direction. 2
Point Q_Five, Q₆Is also used to adjust the three-dimensional appearance of the cylindrical shape.
However, by adjusting this point appropriately,
The pixel distribution of the part can be changed to some extent,
Can be diversified. Now we will explain how to calculate the coordinates to get the deformed image
I do. First, a process of changing the pixel density is performed. Figure 17 shows the original
This shows how an image is transformed into an image with changed pixel density
This is an example. FIG. 17 (a) is an original image, and FIG.
(B) is an image after changing the pixel density in the vertical direction.
FIG. 17 (c) shows the horizontal square pixel density of the image shown in FIG. 17 (b).
Is an image after changing. Fig. 17 (b) shows the cylindrical shaft
Fig. 17 (c) shows the three-dimensional effect along the cylinder axis.
Is to give. Change pixel density in vertical direction
Equation (10) shows the coordinate calculation formula for the calculation. Vertical direction y '= Ay^Two+ By + C (10) where y is the vertical coordinate before the vertical pixel density change processing.
Where y 'is the coordinate after processing. Where in equation (1)
Coefficients A, B, and C are as follows: input coordinate variable y in equation (1) is 0, | P_1yP_5y|,
| P_1yP_2yWhen |, the value of the output coordinate variable y 'is 0, | Q_1YQ_5Y|, |
Q_1YQ_2YIt is a value that satisfies | P_1yIs the point P₁Vertical
Indicates the coordinate value of the direction. | P_1yP_2y| Is side P₁P_TwoThe length of | Q_1Y
Q_2Y| Is side Q₁Q_TwoIs the length of Before calculating equation (10), the vertex P1 at the lower left corner of the original image
The original image so that is the origin position on the xy plane coordinates.
It is necessary to keep it. Next, coordinate calculation formula for horizontal pixel density change processing
This is shown in equation (11). Horizontal direction x '= Dx^Three+ Ex^Two+ Fx + G (11) where x is the horizontal coordinate before the horizontal pixel density change processing.
Where x 'is the coordinate after processing. Where in equation (11)
Coefficients D, E, F, and G are set as follows:_1xP_6x
|, | P_1xP_4xWhen the value of |, the value of the output coordinate variable x 'is 0, | Q_1x
Q_6x|, | Q_1xQ_4x| The condition that satisfies
Variable x is | P_1xP_4xIn the case of |
This is a value when the condition that the minute value is 0 is satisfied. Differential value 0
Condition makes the horizontal edge of the image as small as possible.
It is for. Due to this, it was transformed into a cylindrical shape
When the pixel at the end becomes infinitely small. After performing the coordinate calculation to change the pixel density,
Calculate coordinates to change. Figure 6 shows the rectangular image
Non-rectangular shape with four sides including two curves facing the whole outer shape
It is an example showing a state of change. FIG. 18 (a) is outside the whole
FIG. 17 (c) is an image before deformation in the shape changing process.
Fig. 18 (b) is the same as
Fig. 18 is an image in which the shape of the opposite side of the set is curved.
(C) is a cylindrical image finally obtained. A set of
In order to make the shape of the opposite side a curved shape,
Deform so that it is sandwiched between two meshed curves
You. Formula (12) shows the coordinate calculation formula for this processing. What
Because the horizontal width of the image is not changed at this time
The horizontal coordinates are left as they are. y '= (f₁+ Q_3y−Q_4y−f_Two) / (Q_2y−Q_1y) × y ′ + f_Two …… (12) In the equation, y ′ is the vertical coordinate before the process of sandwiching in the curve shape.
Yes, Y is the coordinates after processing. f₁Is the image in Fig. 18 (b).
Expression y ′ = f for expressing the shape of the upper side of the image₁(X ')
Yes, | Q_1xQ_4x| Is the horizontal axis length | Q_2yQ_3y| Vertical
When placed at the origin on the center point xy plane coordinates as the axis length of the direction
It is a Sui-yen style. f_TwoIs the lower side of the image in FIG. 18 (b).
Expression y '= f for expressing shape_Two(X ') and | Q_1xP_4x|
Is the horizontal axis length | Q_1yQ_4y| As the vertical axis length
This is the elliptical formula when the center point is placed at the origin on the xy plane coordinates.
You. Obtain the final image after performing the deformation sandwiched by the curved shape
Process for In this process, the horizontal width of the image is
Coordinate calculation for processing, changing to change continuously
The equation is shown in equation (13). In this case, the vertical direction of the image
The vertical coordinate is left as it is because the width of
You. X = (1- (1-H) / 1Y) x '(13) where x' is the horizontal coordinate before this processing, and X and Y are
Horizontal and vertical coordinates after processing. The coefficients H and I are
Is a constant expressed as H = ｛(Q_2y−Q_4y) (Q_3x−Q_2x)-(Q_2y−Q_3y) (Q_4x −Q_1x)｝ / (Q_3y−Q_4y) / (Q_4x−Q_1x) …… (14) I = Q_2y−Q_4y ...... (15) As described above, the process of changing the pixel density and the overall
Deformation is performed by combining with further processing. No.
In Fig. 15, density change processing is added by the former processing.
Image, and then apply external shape change processing to that image.
But it does not actually generate an intermediate image
And the positions of equations (10) to (13) for one pixel continuously
Performs a target calculation. Deformation processing of two-dimensional image of the third embodiment described above
The hardware configuration for performing
FIG. 8 shows the overall configuration as in the second embodiment.
The configuration of the image transformation processing section is shown in FIG. Deform coordinate calculation
The basic configuration is the first embodiment except that the processing content of the unit 13 is different.
The configuration is the same as that of the example or the second embodiment. This third
The operation of the apparatus according to the embodiment corresponds to the operation flow shown in FIG.
To the coordinate calculation of the pixel density change processing in step 704 (10)
Point to use expression (11) and change of outer shape of step 705
The first point is that the equations (12) to (13) are used for the coordinate calculation of the processing.
It differs from the device of the embodiment. According to the third embodiment, a two-dimensional plane is displayed in a two-dimensional plane.
Can be pasted on the cylindrical image in
And can be deformed in the direction around the cylindrical axis.
The point where the sensation and the perspective in the direction along the cylinder axis are indicated (Q_Five, Q₆)
Can be easily changed by adjusting.

【The invention's effect】

According to the present invention, the user can easily change and adjust the degree of the three-dimensional effect by a simple operation by providing the input unit for inputting the designated point for adjusting the degree of the three-dimensional effect of the deformed image. Can be.

[Brief description of the drawings]

FIG. 1 is a schematic block diagram of an image deformation processing unit that performs calculations for image deformation, which is a feature of the present invention. FIG. 2 is a diagram showing an example of an image before and after a transformation process of adding a perspective to a plane image, wherein (a) shows an input original image,
(B) is an output deformed image. FIG. 3 is a diagram showing an example of a state in which a rectangular original image is sequentially transformed into a parsed image, where (a) is an input original image, (b) is an intermediate image, and (c) Is an output deformed image. FIG. 4 shows images (a) and (b) before and after the deformation processing in FIG.
FIG. 8 is a diagram in which an indication point given by an operator is added to FIG. FIG. 5 is a diagram showing images (a) and (b) before and after the process of changing the pixel density with respect to the original image. FIG. 6 shows images (a) and (b) before and after the process of changing the entire outer shape of the image into a non-rectangular shape. 7A and 7B are diagrams showing the entire processing flow from reading of pixel values to writing through coordinate calculation and the like. FIG. 8 is a diagram showing an example of a hardware configuration for implementing the present invention. FIG. 9 is a diagram showing an example of an image before and after a process of transforming a planar image into a spherical shape like a globe, where (a) is an input original image and (b) is an output deformed image. FIG. 10 is a diagram showing an example of a state in which a rectangular original image is sequentially transformed into a spherical image, (a) is an input original image, (b) is an intermediate image, and (c) is an image. It is an output deformed image. FIG. 11 shows images (a) and (b) before and after the deformation processing of FIG.
FIG. 8 is a diagram in which an indication point given by an operator is added to FIG. FIG. 12 shows images (a) and (b) before and after the process of changing the pixel density for the original image. FIG. 13 shows images (a) and (b) before and after the process of changing the entire outer shape of the image into a non-rectangular shape. FIG. 14 is a diagram showing an example of an image before and after a transformation process of adding a perspective to a plane image, where (a) shows an input original image,
(B) is an output deformed image. FIG. 15 is a diagram showing an example of a state in which a rectangular original image is sequentially transformed into a cylindrical image, (a) is an input original image, (b) is an intermediate image, and (c) is an image. It is an output deformed image. FIG. 16 shows images (a) and (b) before and after the deformation processing of FIG.
FIG. 8 is a diagram in which an indication point given by an operator is added to FIG. FIG. 17 shows images (a) to (e) for explaining the process of changing the pixel density in the vertical and horizontal directions with respect to the original image.
It is a figure showing (b). FIG. 18 is a diagram showing images (a) to (c) for explaining a process of changing the entire outer shape of the image into a non-rectangular shape. 11 ... deformation initial value input unit, 12 ... trace control unit, 13 ... deformation coordinate calculation unit, 1301 ... pixel density change unit, 1302 ... image shape change unit, 14 ... inverse mapping calculation unit, 15 ... ... Integer coordinate calculation unit, 16 ... Inverse mapping coordinate calculation unit, 17 ... Interpolation processing unit, 18 ... Pixel input / output control unit, 19 ... Image memory, 81 ... Calculator, 82 ... Central processing unit, 83: Image transformation processing unit, 84: Image memory, 85: Input device, 86: Display, 87: Scanner, 88: Printer, 89: Bus.

Claims (1)

(57) [Claims] 1. A pixel density changing means for changing a pixel density of an original image by a predetermined function, an image shape changing means for changing a shape of an image processed by the pixel density changing means, and a pixel density by the pixel density changing means Input means for providing an instruction point for determining the coefficient of the predetermined function in order to adjust the rate of change of the two-dimensional image.