JPH04167076A - Two-dimensional image deformation processor - Google Patents

Abstract

PURPOSE:To deform an image which is a two-dimensional plane so that the image is made simply uneven still in the form of two-dimensional information by giving random distortion to the two-dimensional image and changing its shape. CONSTITUTION:Representative points which are separated at certain intervals are determined by 1st processing means 102 - 104 in the two-dimensional plane image, and the coordinates of those representative points are converted by using a random function to restrict the range wherein the coordinates move so that adjacent representative points interfere. When the coordinate conversion of all the representative points is completed, coordinate converting processes for points other than the representative points are performed by 2nd processing means 102, and 105 - 108. This coordinate conversion is carried out in small plane units consisting of plural adjacent representative points. Namely, the two-dimensional image is distorted at random and deformed. Consequently, the operation is easy, the amount of data which are processed decreases, and the processing speed becomes fast.

Description

[Detailed description of the invention] [Industrial application field]

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

[Conventional technology]

In the world of image processing and computer graphics, there is a need for functions aimed at so-called special expressive effects that give a two-dimensional plane image a sense of unevenness to make it appear three-dimensional. In order to give a flat image a sense of unevenness, it is necessary to change the shape. In order to fully change the shape and create a sense of unevenness, there is texture mapping, which generates a 3D model with varying heights in 3D space and pastes a flat image on top of it (for example, , Angoin et al.: “
"Texture Mapping for Multiple Three-Dimensional Shapes", Graphics and CAD31-7 (1988, 2, 26
)reference). Although this method makes it possible to express realistic unevenness, it is troublesome to create a three-dimensional three-dimensional model. Furthermore, since the data used for processing is three-dimensional information, there is a problem in that the amount of data and the amount of calculation are large. Another method to give a sense of unevenness is to generate a self-contained surface in a two-dimensional plane without using a three-dimensional solid model, and then perform texture mapping on that curved surface.This gives a sense of unevenness to a flat image. Another possible method is to transform it into However, when generating a mountain surface, it is necessary to determine the positions of control points for determining the shape, which is also a troublesome operation.

[Problem to be solved by the invention]

The present invention was proposed in order to solve the above-mentioned problems, and provides a two-dimensional image deformation processing device that easily deforms a two-dimensional plane image so that it has unevenness while retaining two-dimensional information. The purpose is to provide

[Means to solve the problem]

The two-dimensional image deformation processing device of the present invention has two-dimensional
A first processing means converts the coordinates of representative points separated by a certain interval in the image to arbitrary positions using random values within a movement range that does not interfere with adjacent representative points (see Fig. 10).
3 to 104.102), and a second processing means that performs coordinate transformation for points other than the representative points based on a deformation function for each small plane WIi image composed of a plurality of adjacent representative points ( 105-108, 102) in FIG.

[Operation] In the first processing means, first, a two-dimensional plane image (
2), representative points (FIG. 3) separated by a certain interval are determined. Coordinate transformation is performed on this representative point using a random function. Through this transformation, the range in which the coordinates move is limited so that they do not interfere with adjacent representative points (Figure 4).This means that if the moving distance is large, the coordinates will be placed at a position that jumps over adjacent points, and the representative points will This is necessary to avoid the problem of swapping the order of the . This coordinate transformation is calculated, for example, using equations (1) and (2), which will be described later. When the coordinate transformation process is completed for all the representative points, the second processing means performs the coordinate transformation process for points other than the representative points. This coordinate transformation is performed in units of small planes each consisting of a plurality of adjacent representative points. For example, the unit is a triangle connecting three adjacent representative points (FIG. 5). For example, affine transformation is used for the transformation. According to the present invention, since the shape is changed by applying random distortion to a two-dimensional image, the above-mentioned conventional technology that imparts a sense of unevenness by texture mapping to a three-dimensional three-dimensional model or a free-form surface in a two-dimensional plane It has the advantages of being easier to operate, requiring less data to be processed, and faster processing speed.

【Example】

An embodiment of the present invention will be described with reference to the drawings. As shown in FIG. 1, the deformation processing device of this embodiment includes a deformation initial value input section 101 that inputs information such as the size and position of images before and after deformation, and the size of a small plane image that is a unit of deformation processing. , a trace control unit 102 that controls the selection procedure of a Kodaira image to be a unit of transformation processing based on the input initial value, and also controls the selection procedure of a point that is an integer coordinate in the small plane image after the transformation processing; The random value generation unit 103 generates a random value between 1 and 1, and the coordinates of the selected small plane image using the random value generated from the random value generation unit 103 after the transformation process are determined. A deformation coordinate value calculation unit 104 that calculates a value calculates a coordinate calculation formula for deformation from the vertex coordinate values of the selected small plane image and the small plane Ii image after the deformation processing, and further calculates a coordinate calculation formula for inverse deformation from that formula. A coordinate calculation formula generation unit 105 to be obtained, an integer coordinate point detection unit 106 that detects a point with an integer coordinate value in the small plane image after transformation processing, and an inverse transformation of the point detected by the integer coordinate point detection unit 106 an inversely transformed coordinate value calculation unit 107 that calculates coordinate values on the original image using the coordinate calculation formula; and an interpolation processing unit 108 that calculates new pixel values on the transformed image from the calculated coordinate values by interpolation calculations. and a pixel input/output control unit 109 that controls writing and reading of pixel values. 110 is an image memory that stores image data before and after transformation processing. First, an outline of the modification processing of this embodiment will be explained. FIG. 2 is an example of an original image on a two-dimensional plane. The width of this image in the horizontal direction is W, and the height in the vertical direction is Hl. Here, the unit of these sizes is the number of pixels. The coordinates of pixel psLa in the lower left corner of the image are (XaLo, YIILD
). The coordinate values (XsRu, YSRU) of pixel P IIRII in the upper right corner of the image are (Xsto+Ws-1, Ys
to + Hs-l). Transform the two-dimensional plane image. The purpose of the present invention is not to express faithful unevenness, but to deform it so as to give a three-dimensional effect. Therefore, the present invention attempts to create a three-dimensional effect by applying random distortion to a two-dimensional image and changing its shape. FIG. 3 is an example in which pixels located at equal intervals in the original image of FIG. 2 are displayed as representative points. The black circles in the figure are representative points. In the example shown in this figure, the intervals between the representative points in the horizontal direction and the vertical direction are equal, but there is no need to be limited to this. The representative points may be determined by specifying the interval by the number of pixels, or by giving the number of wooden frames in the horizontal and vertical directions of the original image. Now let's perform coordinate transformation on this representative point using a random function. FIG. 4 is an example showing one representative point and a plurality of representative points located around it. Although nine representative points are shown in the figure, we will focus on the representative point located at the center among these points and consider the case where processing is performed on that point. The coordinates of the central representative point Pc are (Xc, Yc). The interval between two adjacent representative points is WP in the horizontal direction and HP in the vertical direction. In this example, since the representative points are arranged at regular intervals in FIG. 3, Wp and H in FIG. 4 are equal. Let us now move the central representative point Pc to an arbitrary position. If this point is moved a long distance, it will be placed at a position that jumps over adjacent points, causing a problem in which the arrangement of representative points will be swapped. Since other representative points also move arbitrarily, it is necessary to limit the range of movement to prevent this problem from occurring. The dotted line surrounding the central representative point PC in FIG. 4 is the moving range of this point. That is, according to this condition, the central representative point P. The moving range of is (xc-Wp/2
~Xc+Wp/2. Yc-Hp/2~Yc+Hp/2)
becomes. Equations (1) and (2) show conversion equations in each horizontal and vertical direction for coordinate conversion within this movement range. In this equation, RAND (X) and RAND (Y) are random functions whose values are output from 0 to I. RAND
(X) and RAND (Y) are one random function RAN
D may also be used. In that case, the function RAND needs to take different values in equations (1) and (2). X and Y are coordinate values in the horizontal and vertical directions before movement transformation, and X l and Y l are coordinate values after movement transformation. X'=Xc-Wp/2+RAND(X)xWp (
1) Y'=Yc-H,/2+RAND(Y) zero HP
(2) Perform movement conversion processing on all representative points using the above equation. Next, a method for performing movement conversion processing on points other than representative points will be described. FIG. 5 is an example showing a method of extracting three adjacent representative points from among the representative points and moving and converting points other than the representative points along with the representative points. FIG. 5(a) shows three adjacent representative points selected from a plurality of representative points. FIG. 5(b) also shows representative points and points between them, where black circles are representative points and white circles are points inside the triangular image formed from the three representative points. FIG. 5(C) shows the situation after the three representative points have been subjected to movement conversion processing within the movement ranges indicated by dotted lines. By the movement conversion process, the right triangular image formed by the representative points in FIG. 5(a) is transformed into the triangular image having an arbitrary shape as shown in FIG. 5(c). This transformation process is performed using the Fli and standard transformations of equations (1) and (2), but an affine transformation will be applied to this transformation. At this time, for example, a point located midway between the two representative points before movement conversion is considered to be converted to a position midway between the two representative points after movement conversion. The coefficients of the coordinate calculation formula for the affine transformation are obtained from the coordinate values of the six vertices of the triangular image before and after the movement transformation. From the coordinate calculation formula obtained from this, find the coordinate calculation formula for the inverse transformation from after the movement transformation to before the transformation, select the points inside the triangular image after the transformation processing by tracing, and use the coordinate calculation formula for the inverse transformation. The coordinate values inside the triangular image before the deformation process are determined by the process, and the deformation process is performed by writing the pixel values calculated by the interpolation process to the selected points. There are several methods of interpolation processing, such as nearest neighbor interpolation and four-point linear interpolation, but since these give differences in processing speed, image quality, etc., the operator can select the method of his choice. In order to transform the entire original image, the aforementioned transformation of the triangular image may be performed on all the triangular images in the original image. A method for sequentially selecting all the triangular images will be described. FIG. 6 shows an example of dividing the original image of FIG. 3 into a plurality of triangular images and tracing them. Arrows represent tracing order. First, select from the triangular image in the lower left corner, proceed to the adjacent triangular image on the right, and when you reach the rightmost triangular image, proceed to the upper stage and proceed with the selection in the same way. Repeat this selection until the last triangular image in the upper right corner. The deformation process is performed when each triangle tm& is selected. Two adjacent triangles contain points in common. Therefore, among the representative points of one triangular image, for a representative point of a triangular image to be selected later, the deformed coordinate values of that point are held until they are processed at the time of later selection. Points on the lower side among the three vertices of the triangular image are used again when the trace advances to the upper stage. Therefore, 1. It is at least necessary to hold the coordinate values after movement transformation for the number of representative points for the line. Now, the overall processing flow including triangular image tracing and deformation processing by the deformation processing device of this embodiment shown in 1st vlJ will be explained. FIG. 7 shows the flow. (701) Initial information necessary for transformation is input using the transformation initial value input unit 101. The information includes the size and position of the image before and after transformation, the interval between representative points, or the number of divisions into small plane images of the transformation processing unit. (702)) Original picture #! by the race control unit 102. Divide the area into triangles mca and start tracing. (703)) Triangular images are sequentially selected by the trace control of the race control unit 102. (704) The deformed coordinate value calculation unit 104 performs coordinate calculation for the movement transformation process on the coordinate values of the three vertices of the selected triangular image using the above equations (1) and (2). Here, calculations are performed only for points for which coordinate calculations have not been performed among the three vertices. For points that have already been calculated, use those values. Also, remember the coordinate values of the points to be reused later (.tt t
d, random values RAND (X) and RAND (Y) are those generated by the random value generation unit 103. (705) The coordinate calculation formula generation unit 105 calculates a coordinate calculation formula for deformation of the triangular image from the coordinate values of the vertices of the triangular image selected in step (703) and the coordinate values obtained in step (704). . (70B) Furthermore, the coordinate calculation formula generation unit 105 obtains a coordinate calculation formula for inverse deformation from the coordinate calculation formula for deformation obtained in step (705). (707) Start tracing points that are integer coordinates within the area of the deformed triangular image obtained in step (704). This trace is processed by an integer coordinate point detection unit 106 that operates under the control of a trace control unit 102.
The river ■ is carried out. The subsequent steps (708) to (713) are transformation processing for one triangular image. (708)) The race control unit 102 sequentially selects points having integer coordinate values by tracing within the deformed triangular image area. (709) The roundabout transformation coordinate value calculation unit 107 calculates the coordinates of the point selected in step (708) using the coordinate calculation formula of the inverse transformation obtained in step (706), and calculates the coordinate value on the original image. seek. (710) The pixel value of a point having one or more integer coordinate values near the coordinate value obtained in step (709) is read from the image memory 110 under the control of the pixel output control unit 109. (711) Interpolation processing 108 performs interpolation processing based on the read pixel values to calculate new pixel values. (712) Write the obtained new pixel value to a point on the image memory 110 that is an integer coordinate value in the deformed triangular image selected by tracing. (713)) The race control unit 102 determines whether tracing within the deformed triangular image area has been completed. If Yes, proceed to step (714);
If so, the process returns to step (708). (714) Determine whether tracing of the triangular image on the original image is completed. Finally, an example of the overall hardware configuration for implementing the present invention is shown in FIG. In the figure, 81 is a computer;
82 is a central processing unit, 83 is an image transformation processing unit, 84
is an image memory, 85 is an input device, 86 is a display,
87 is a scanner, 88 is a printer, and 89 is a bus. The input device 85 is for giving commands etc. to the computer. The display 86 outputs and displays images. The display 86 outputs and displays images. The scanner 87 is for inputting images, and the printer 88 is for outputting images. The path 89 serves to transfer image data and control information. The image memory 84 stores image data, and the central processing unit w182 controls the entire image processing apparatus and performs general calculations. The image transformation processing unit 83 performs processing for image transformation as described above, and has the configuration shown in FIG. 1.

【Effect of the invention】

According to the present invention, the configuration is such that a two-dimensional image is given a random distortion and its shape is changed to give a three-dimensional effect. Compared to the technique of imparting unevenness by muffing, this method has the advantage of being easier to operate, requiring less data to be processed, and having a faster processing speed.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a processing section that performs calculations for image transformation. FIG. 2 is an example showing an original image on a two-dimensional plane. FIG. 3 is an example in which pixels located at equal intervals in the original image are displayed as representative points. FIG. 4 is an example showing one representative point and a plurality of representative points located around it. FIG. 5 is an example showing a method of moving and converting three adjacent representative points and points other than the representative points. FIG. 6 shows an example of tracing a plurality of triangular images divided within the original image. FIG. 7 shows the overall processing flow from reading out pixel values to performing coordinate calculations and writing. FIG. 8 shows an example of a hardware configuration for implementing the present invention. 101... Transformation initial value input section, 102... Trace control section, 103... Random value generation section, 1o4...
Deformed coordinate value calculation unit, 105...Coordinate calculation formula generation unit, 1
0B...Integer coordinate point detection section, 107... Inverse transformation coordinate value calculation section, 108... Interpolation processing section, 109... Pixel input/output control section, 11o... Image memory. Patent applicant Fuji Xerox Co., Ltd. Agent Patent attorney Noboru Iwagami - (
Figure 3 Figure 4 (a) (b) (c) Figure 6 Figure 1 Figure 2 Figure 8

Claims (1)

[Claims] For representative points separated by a certain interval in a two-dimensional plane image, coordinates are transformed to arbitrary positions using random values within a movement range that does not interfere with adjacent representative points. and a second processing means that performs coordinate transformation for points other than the representative points based on a deformation function in units of small plane images made up of a plurality of adjacent representative points. A two-dimensional image transformation processing device characterized by: