JPH0887586A - Hierarchical approximating method for graphic data - Google Patents

Abstract

PURPOSE: To provide a position determining method for the graphic data which gives an observer no feeling of physical disorder when the positions of feature points which are made to correspond to each other are converted in the case of shape transformation. CONSTITUTION: Geometrical model data are inputted and made into a defocused shape through spatial filter processing according to desired resolution S2. Its outline is extracted S3, angle variations of respective parts are calculated, and parts exceeding a certain value are extracted as feature points S4. The processes S2-S4 are repeated as many times as desired layers S5, the extracted feature points are made to correspond on the basis of the positions of the feature points in respective layers and the signs of angle variations on corresponding outlines S6 and adjusted so that the feature point positions smoothly move to other layers S7, and graphic data wherein the positions of the feature points are determined are drawn as CG S8.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION This invention relates to CG (Comput
er Graphics), the present invention relates to an approximation method for reducing the complexity of the shape of a geometric model. In particular, starting with games that use CG, VR (Vi
It is widely used and suitable for design, etc.

[0002]

2. Description of the Related Art When drawing a CG, it is common to always draw using the same model regardless of the position, size, depth, observer's attention point, and model moving speed of the model. . However, this CG drawing does not always require the same model. In other words, depending on the model position, size, depth, observer's attention point, and model moving speed on the screen, it is sufficient to draw using a more simplified model instead of the original detailed model. It is possible to obtain excellent image quality. For example, when the model is arranged at a distance, the size of the model becomes small on the screen, and the minute unevenness change of the original model cannot be seen. In other words, in such a case, it is not necessary to draw using a detailed original model, and drawing may be performed using an approximate model in which the detailed part of the original model is cut.

For that purpose, it is necessary to approximate the shape making use of the feature points of the original model. This characteristic point is a part where a human strongly reacts to a shape.
If the shape is approximated by ignoring the feature points, a person feels a lot of discomfort with the approximation of the shape. However, by approximating while leaving the feature points, it is expected to minimize the discomfort for the shape approximation.

Further, in order to obtain the shape of the intermediate layer between the discrete layers, it is necessary to associate the feature points between the layers and determine the position of each feature point from the correspondence relationship of the feature points. . As a result, the conversion from the original shape to the most approximated and simplified shape becomes continuous, and the discomfort for the conversion is reduced.

Regarding the approximation of this shape, Greg Turk
"Re-Tiling Polygonal Surface" by Computer Graphi
In csVol.26, No.2, July 1992), an attempt is made to hierarchically approximate a polygon model. However, in the method of this paper, shapes that are hierarchically approximated are not associated with each other, so that the shapes in the intermediate layers cannot be obtained. Furthermore, the position of the feature point is not adjusted so that the conversion is smooth depending on the correspondence relationship of the feature points between different layers.

Francis JMSchmitt, Brian A. Bars
"An Adaptive Subdivision Method" by ky, Wen-Hui Du
forSurface-Fitting from Sampled Data "(Computer Gr
In aphics Vol.20, No.4, August 1986), a Bezier patch is attached to a three-dimensional shape to approximate the shape, but there is a problem that general polygons are not targeted. Furthermore, in this research, even if the shapes are approximated hierarchically, the shapes of the layers are not associated with each other, so that interpolation between layers cannot be executed. In other words, since it is not possible to obtain an intermediate layer, the conversion becomes abrupt when layers are switched.

In order to perform continuous smooth conversion from the original shape to the most approximated and simplified shape, feature points between different layers are associated with each other, and the corresponding relationship is used to identify each layer. It is necessary to determine the position of the feature point at. In determining the position, points that are considered to be the same feature point in different layers can be adjusted by adjusting the position.
Differences when transforming shapes can be reduced. However, such a method has not been taken conventionally.

[0008]

Up to now, although the approximation of the geometric model used for CG can be performed, the shape of each layer has a shape by associating it with discrete layers. The process of returning the position of the feature point to the position that further suppresses the discomfort for the observer has not been performed. Also,
There has been no method for performing approximation based on feature points and minimizing discomfort during shape conversion.

Therefore, an object of the present invention is to determine the position of the feature point in each layer from the correspondence in which the feature points are associated with each other, so that the position of the feature point in each layer can be determined so as to suppress discomfort during shape conversion. It is to provide a position determination method. Also, based on the correspondence of feature points between layers, the points that are considered to be the same between layers are judged, and the position is adjusted so as to suppress the discomfort to the observer when converting the layers. Such as having a position determination method of
It is to provide a hierarchical approximation method for graphic data.

[0010]

In order to solve the above-mentioned problems, the present invention provides a step of feature point extraction for hierarchically extracting feature points of graphic data at a plurality of resolutions, and a group of feature points. A hierarchical structure of graphic data, which includes a correspondence determination step for determining the correspondence with the position on the original graphic data that it represents and the correspondence between feature points among multiple layers, and a position determination step for adjusting the positions of the feature points. In the approximation method,
The position determining step is a method of hierarchical approximation of graphic data, which is characterized in that the positions of the characteristic points of a plurality of corresponding layers are adjusted so that the positions change smoothly when the layers are sequentially changed.

[0011]

According to the above-mentioned structure, the characteristic points at the desired resolution of the original graphic data are extracted by the characteristic point extracting section. The extracted individual feature points are associated with each layer in the correspondence determining unit, and the correspondence between these feature points indicates that the position of the feature points changes smoothly when the layers are sequentially changed. Is adjusted. Further, these feature points are set so that their positions do not greatly shift based on the correspondence between the feature points. Furthermore, the positions of feature points that are considered to be the same in each layer are adjusted in position. This adjustment of the feature point position can also be performed by an adjustment method of returning to the point of the original shape.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A first embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a flowchart showing the entire first embodiment of the present invention. The geometric model data used for drawing the CG is input in step S1. This geometric model data is spatially filtered in step S2 according to the desired resolution. In a specific example of this spatial filter processing, convolution of an original geometric model and a Gaussian function is executed to perform low-pass characteristic spatial filter processing. As a result, the model shape becomes a blurred shape. The geometric model that has been blurred in step S2 is
At 3, the contour line is extracted. This example is shown in FIG. FIG. 2A shows the original shape, and FIG. 2B shows the contour line extracted after the low-pass characteristic spatial filter processing is performed on the original shape. Regarding the method of convolution and tracking of contour lines described above, Shigeo Minami, “Waveform Data Processing for Scientific Measurement” (C
Since it is described in books such as "Practical computer graphics" (Nikkan Kogyo Shimbun) supervised by Q Publishing Company) and Fujio Yamaguchi, it is omitted here.

The above-mentioned geometric model whose contour line has been extracted is
In step S4, the angle change in each part of the contour line is calculated, and the part exceeding a certain fixed value is extracted as a feature point. FIG. 3 shows an example of this feature point extraction. In FIG. 3A, with respect to the contour line of FIG. 2B described above, the angle change in each part is calculated, and the part where the extreme value of the angle change exceeds a certain value is shown by a black circle. Further, in this figure, the feature points are extracted even in the region in which the cumulative change in angle exceeds a certain value in the region in which the change in angle changes with the same sign. Here, the half value part of the cumulative value is used as the feature point. The black circles of each part in FIG. 3A are the extracted feature points and correspond to the black circles of FIG. 3B.

FIG. 4 shows an example in which these extracted feature points are connected. Here, FIG. 4A is an example of a case where the shape is greatly blurred in step S2 of FIG. Also, FIG.
B is an example in which the degree of blurring is smaller in step S2 of FIG. 1 than in FIG. 4A. Therefore, FIG. 4B has a shape closer to the original shape. Note that the processing from step S2 to step S4 described above is repeated for the number of desired layers (step S
5).

The feature points extracted as described above are associated between layers in step S6 of FIG. The feature points are associated with each other based on the position of the feature point in each layer and the sign of the angle change on the contour line corresponding to the feature point. As the layers become different and the shape becomes more complex, new feature points appear. FIG. 5 shows an example in which the feature points are associated as described above. A and B in the figure are characteristic points extracted from different layers and connected. The correspondence of the feature points between the layers at this time is indicated by dotted arrows. Although only two layers are shown here, by associating all the layers with the original vertices, the association from the original to the most approximated and simplified layer is executed.

However, the position of the feature point obtained in each layer as described above is the position in the shape subjected to the spatial filter processing of the low pass characteristic. This state is shown in FIG. In FIG. 6A and FIG. 6B, one of the characteristic points is selected and shown by the original shape by the dotted line, the blurred shape by the solid line, and the black circle and the square. Here, the point X is a certain characteristic point in the original shape, and this characteristic point corresponds to the point Y in FIG. 6A and further corresponds to the point Z in FIG. 6B. Further, in FIGS. 6A and 6B, which are different layers, points Y and Z correspond to each other, and this is indicated by an arrow in the figure. The solid line is the shape obtained by performing the low-pass spatial filtering process on the original shape shown by the dotted line.

The shape is blurred due to the spatial low-pass processing, and as a result, the position of the contour line is also changed.
That is, the positions of the feature points extracted from the contour line also differ in each layer. Therefore, as indicated by the dotted arrow, the positions of the associated feature points are shifted for each layer. Further, as can be seen from this figure, in the case of this example, the size of FIG. 6B is different from that of FIG. 6A, and especially the vertical spread is changed. As a result, when moving between layers and displaying based on the position of the characteristic point as it is, the position of each characteristic point fluctuates, and the size changes due to the blurred shape.

In order to prevent these problems, the position of each feature point is adjusted by the correspondence between layers (step S7). In the processing up to step S6, since the correspondence relation of the feature points between the layers has already been obtained, it is possible to determine which feature point is the same in each layer by using the correspondence relation. The feature points that are considered to be the same are set so that the feature point positions move smoothly when moving between layers. In the first embodiment, a method of following the locus of the change in the position of the feature point between layers and setting it so as to be smooth is used. In this method, by constantly adjusting the position to the center of gravity of the locus, there is little variation in feature points, and it is possible to reduce the sense of discomfort felt by the observer when moving between layers.

When actually determining the position of a feature point from the correspondence relation of feature points between layers, for example, a method of making the correspondence relation of feature points between layers into a list structure is conceivable. This list structure is, for example, as shown in FIG. 7, from the original vertex position to the corresponding feature point of the hierarchy N, from the feature point of this hierarchy N to the corresponding feature point of the hierarchy N + 1, and further to this hierarchy. The corresponding feature points can be traced in order between the layers, such as from the N + 1 feature points to the corresponding feature points of the hierarchy N + 2.

From this list structure, it is determined whether or not each feature point is the same between layers such as layer N and layer N + 1. For each characteristic point determined to be the same, the position of each characteristic point is examined from this list structure, the locus of the positional change of the characteristic point described above is followed, and the position of the characteristic point is adjusted to its center of gravity. With the above processing, the position of the characteristic point for each layer can be determined for each characteristic point.

The graphic data in which the positions of the characteristic points are determined by the above processing is drawn as CG in step S8.

Next, a second embodiment of the present invention will be described with reference to the drawings. This second embodiment is the same as the first embodiment described above.
The processing of this embodiment and the processing up to step S6 are exactly the same. Therefore, the description of the processes up to step S6 will be omitted. That is, in the processing up to step S6, since the correspondence relation of the feature points between the layers has already been obtained, it is possible to determine which feature point is the same in each layer by using the correspondence relation.

In the second embodiment, the above-mentioned step S
For the feature points determined to be the same in the processes up to 6, a method of returning the position of each feature point to the position in the original shape is used. This method has the advantage that the position can be easily obtained without the need for complicated calculations. further,
Since the positions of the characteristic points do not change, the discomfort felt by the observer when the shape changes can be further suppressed.

Also in the case of the second embodiment, the list structure described in the first embodiment can be used when the position of the feature point is actually determined from the correspondence of the feature points between the layers. Here, from this list structure, for example, the characteristic points of the layer N are traced from the original vertex position, the characteristic points of the layer N + 1 are traced from the characteristic points of the layer N, and the characteristic points of the layer N + 2 are further traced from the characteristic points of the layer N + 1. By repeating the above procedure, it is determined whether the characteristic points in each layer are the same characteristic point. If it is determined that they are the same, the position of the feature point in each of these layers in the original shape is traced from the list structure, and the position is examined. Since the position of each feature point is determined by the above work, the feature point of each layer is rearranged at that position.

FIG. 8 shows the second embodiment described above, that is, an example in which the position of the characteristic point is determined by using the method of returning the position of the characteristic point to the position in the original shape. FIG. 8A shows an example in which the position of the feature point is left as a blurred contour line, that is, the position of the feature point is not adjusted, and FIG. 8B shows the position of the feature point as the position of the original contour line. It was returned to. In this example showing the figure of Hokkaido, it can be seen that the shape of the part T indicated by the dotted circle in the diagram of the hierarchy N, which corresponds to the peninsula near Hakodate, is different. It can be seen that the part T has a similar shape at the stage of the layer N, but the shape changes when the level of approximation changes due to different layers.

The graphic data in which the positions of the characteristic points are determined by the above processing is drawn as a CG in step S8.

Next, an example in which the above-described first embodiment or second embodiment is executed by a computer having a standard configuration will be described. 10 is a bus. 11 is
It is a CRT. Reference numeral 12 is an input device. 13 is
It is a floppy disk drive. The input device here may be a mouse, a keyboard, a digitizer, an image scanner, or the like. The CRT 11 and the input device 12 are connected to the bus 10. Further, 14 is a CPU. Reference numeral 15 is a RAM. 1
6 is a ROM. Reference numeral 17 is a hard disk. CP
The U 14, the RAM 15, the ROM 16 and the hard disk 17 are connected to the bus 10. Hard disk 1
A geometric model data created in advance, a program described below, and the like are stored in 7.

The previously created image data stored in the hard disk 17 is transferred to the RAM 15 via the bus 10.
Is supplied to and stored in. In addition, as the image data, previously created data may be read from the floppy disk by the floppy disk drive 13. Further, it may be input by the input device 12.
The image data input by these methods and stored in the RAM 15 is the original geometric model. For example, one solid is drawn by a polygon.

This image data is stored in the RAM 15 and, at the same time, subjected to the spatial filter processing, the contour line extraction, and the feature point extraction in accordance with the above-described steps S2 to S4. In addition, these processes are repeated until a desired number of hierarchical approximation models are obtained as described above (step S5). The obtained hierarchical approximation model is stored in the RAM 15 together with the original geometric model. Furthermore, these hierarchical approximation models are
The feature points for each layer in step S6 are associated with each other. In addition, each associated feature point is set so that the feature point position moves smoothly when moving between layers. The method used at this time is either the method of tracking the locus of the change in the position of the feature point in the first embodiment or the method of returning the position of the feature point to the position of the original shape in the second embodiment. To be selected.

The geometric model is a CPU according to a program.
By the processing of 14, it is displayed on the CRT 11. An example of the display is shown in FIG. 18 is this geometric model.
The display is drawn so as to give the viewer a sense of perspective. In FIG. 11, the upper part of the CRT 11 is farther from the viewer and the lower part is near. At this time,
The geometric model 18 is arranged at a certain position in the virtual space displayed on the CRT 11, and is further moved in this space. Further, the placement and movement of the geometric model 18 can be performed not only by a predetermined designation by a program but also by an input from the input device 12.

The CPU 14 checks the position and movement of the geometric model 18 in the above-mentioned virtual space, and stores in the RAM 15 a hierarchical approximation model of this geometric model suitable for the result. It is selected from the displayed hierarchical approximation models and displayed on the CRT 11. If there is no matching hierarchical approximation model, CPU1
4 uses the result of the above-described feature point association between layers to create a geometric model 18 in which data is interpolated, and a CRT is created.
11 Display on top. These procedures are performed every time, for example, the position of the geometric model 18 in the virtual space changes. In this case, since the feature points for each layer are adjusted to appropriate positions, the graphic data can be smoothly converted when the layers are switched, and the viewer's discomfort can be suppressed.

[0032]

According to the present invention, continuous conversion from the original shape to the most approximated and simplified shape can be smoothly performed. Then, by drawing using the model in which the position of the characteristic point in each layer is returned to the original position, it is possible to draw with a model that is hierarchically approximated. Further, in order to approximate the geometric model used for CG in a plurality of layers, the characteristic points of each layer are associated with each other, and the characteristic points considered to be the same between layers are determined. It is possible to make the conversion between layers continuous by adjusting the positions of the characteristic points determined to be the same from the correspondence relation of the characteristic points. This allows
It is possible to suppress the discomfort felt by the observer at the time of conversion.

As described above, the use of the present invention has the effect of greatly improving the drawing speed without degrading the drawing quality in CG. Also, when approximating the model, it is possible to suppress discomfort as much as possible. The positions of the feature points forming the approximated shape are adjusted even when the hierarchy is moved. Therefore, the process of shape approximation is continuous and smooth.

[Brief description of drawings]

FIG. 1 is an overall flowchart of an embodiment of the present invention.

FIG. 2 is a schematic diagram illustrating a result obtained by a spatial filtering process of a low pass characteristic.

FIG. 3 is a schematic diagram illustrating extraction of a feature point based on a contour line and its angle change.

FIG. 4 is a schematic diagram showing an example in which extracted feature points are connected.

FIG. 5 is a schematic diagram showing an example in which feature points between different layers are associated with each other.

FIG. 6 is a schematic diagram showing an example in which feature points are associated with each other as the contour line subjected to the spatial filtering process of the low-pass characteristic.

FIG. 7 is a schematic diagram for explaining a list structure.

FIG. 8 is a schematic diagram showing a result of determining positions of feature points.

FIG. 9 is a block diagram showing a configuration when the present invention is executed by a computer having a standard configuration.

FIG. 10 is a schematic diagram showing a state where a geometric model is displayed on a CRT.

[Explanation of symbols]

 10 ... Bus 11 ... CRT 12 ... Input device 14 ... CPU 15 ... RAM 18 ... Geometric model

Claims (3)

[Claims] 1. A step of feature point extraction for hierarchically extracting feature points of graphic data at a plurality of resolutions, a correspondence between the group of feature points and a position on the original graphic data represented by the feature points, and In the hierarchical approximation method of graphic data, which comprises a step of determining correspondence between feature points between layers and a step of determining position of adjusting the positions of the feature points, A hierarchical approximation method for graphic data, characterized in that the positions of characteristic points of a plurality of layers are adjusted so that the positions change smoothly when the layers are sequentially changed. 2. The hierarchical approximation method for graphic data according to claim 1, wherein the position determining step adjusts the positions of the characteristic points of the corresponding plurality of layers so as not to change even if the layers are changed. A hierarchical approximation method for graphic data, which is characterized by the above. 3. The method for hierarchical approximation of graphic data according to claim 2, wherein the step of determining the position determines the position of the feature point of the corresponding plurality of layers as the position on the original graphic data corresponding to the feature point. A method for hierarchical approximation of graphic data, characterized by using the position of the part.