JP2015158778A - Solid molding data optimization device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a solid molding data optimization device capable of efficiently obtaining solid molding data while reducing polygons from a surface data formed from the polygons without deteriorating the quality of an output molded object.SOLUTION: An area and a circumferential length of each of polygons forming surface data are calculated (S21), a polygon with a minimum area×circumferential length is defined as a deletion target polygon (S22), and a side to be deleted is selected on the basis of lengths of sides of the deletion target polygon (S23). The deletion target polygon is deleted (S24), polygons sharing an apex of the side to be deleted are successively selected (S25), and polygons sharing two apexes are deleted (S26). For the polygons sharing only one apex, the shared apex is corrected into a midpoint, and an area and a circumferential length of that polygon are calculated again (S27). After processing of the polygons sharing the apex, it is determined whether the number of all polygons is equal to or less than a target value (S28) and if it is not equal to or less than the target value, polygon deletion processing is performed iteratively until the number becomes equal to or less than the target value.

Description

  The present invention optimizes data for output to a three-dimensional object forming apparatus when using a three-dimensional object forming apparatus such as a 3D printer that processes a resin or the like to form a three-dimensional object based on data representing the three-dimensional object. Related to technology.

  In recent years, 3D printers that process and model resin, gypsum, and the like based on data representing a three-dimensional object have become widespread as three-dimensional object modeling apparatuses. As a medical application of 3D printers, modeling data for 3D printer output is created based on CT / MRI images (DICOM format 3D voxel data) taken by medical diagnosis, and organs for surgical simulation, medical education, and informed consent Attempts have been made to output models and create artificial organs (blood vessels, bones, joints, etc.) for implantation in the body.

  As a method for creating modeling data for 3D printer output of an organ model from a CT / MRI image, as shown in Patent Document 1 and Patent Document 2, a voxel of a target organ region is extracted and the Marching Cube method (patent In general, a technique for converting the boundary surface of an organ region into a surface approximated by a polyhedron (a set of triangles) and obtaining surface data expressing the surface shape is used.

Japanese Patent No. 3690501 JP 2012-216102 A US Pat. No. 7,209,136 Japanese Patent Laid-Open No. 8-153211 Japanese Patent No. 3350473 Japanese Patent No. 4463597 Japanese Patent No. 5018721 Japanese Patent Laid-Open No. 11-232489

  In the above surface conversion method, a method of generating a triangle three-dimensionally based on a voxel of about the power of 512 is used, so the triangle polygon becomes enormous, and in the CG field, time is required for screen drawing. It has been pointed out that there is a problem that the responsiveness of the interactive operation becomes worse. This also causes a problem in 3D printer output, and it takes time for printer output preprocessing (processing for converting polygons into data for additive manufacturing), and in some cases causes a work memory overflow and disables output.

  On the other hand, in the CG field, a method is used in which polygonal division is roughened in accordance with the resolution of the screen to express a curved surface, or polygons that are not used for visualization are deleted in advance. For example, in a flat object, polygons that represent only a single surface and constitute the back surface are omitted, or polygons that are hidden from the viewpoint direction and are not used for visualization are deleted in advance, as in Patent Document 4, or Patent Document 5 As described above, a method has been proposed in which inner polygons hidden by other polygons are determined and deleted in advance to reduce the drawing target polygons. Further, as disclosed in Patent Document 6, a method of substituting texture or bump mapping instead of using polygons for a fine uneven expression has been proposed.

  However, in 3D printer output, a resolution finer than that of the screen is required, and since it is necessary to form a closed space with polygons, it is not possible to omit the polygons constituting the back surface in a flat object. In addition, since the model is viewed from various angles, there is basically no hidden surface, and another shape can be modeled inside the model, so the occlusion culling method proposed in the CG field means Does not have Even minute irregularities cannot be visually expressed by shading or the like, and must be expressed by polygonal shapes. As in Patent Document 7, there are cases where a deformed expression with extremely reduced polygons is required in an architectural model, but in medical applications, reality is required as such. Therefore, as in Patent Document 8, there is no choice but to take a technique of reducing the triangular polygons themselves to such an extent that the appearance of the shaped object is not impaired.

  In the reduction of triangular polygons, the method of Patent Document 8 is fundamental. This includes (1) processing to search for triangles and ridges that do not affect the overall shape even if they are reduced from a given polygon, and (2) search for neighboring triangles that share vertices with the target ridges while reducing them. The two types of search processing including correction processing are repeated until the data amount is reduced to a desired amount. The technique of Patent Document 8 requires processing for searching for adjacent triangles that share ridgelines and vertices that are candidates for deletion in both processes, and the processing load increases in proportion to the square of the number of polygons given. Was holding.

  As a method for reducing the processing load, it is only necessary to reduce the number of input polygon sets to be reduced. Therefore, the input polygon set may be divided and processed. Japanese Patent Application Laid-Open No. 2004-228620 proposes a method of performing polygon conversion at the stage of a voxel image before conversion into polygons, and performing surface conversion for each divided voxel data to reduce polygons. Although this can certainly increase the speed, a surface is formed at the boundary part divided by voxel, so when polygon data is integrated, discontinuous cracks corresponding to the divided boundary part occur on the modeled object. (However, in CG applications, the boundary portion is not noticeable due to rendering).

  Also, even if it looks as if it is a single closed space, it may be separated into multiple closed spaces in terms of data. If there is a small space in contact with the main closed space, it will be output directly to the printer. However, if polygon reduction is performed according to Patent Document 8 or the like, the small space may be cut away, and the distance from the main closed space may become longer, resulting in a completely free state. If the printer is output in this state, there is a possibility that an error will be stopped in the printer output pre-processing, or that a small space instructed to float in the air may fall in the middle of output and cause an output trouble (however, for CG applications) In particular, there is no problem.)

  Therefore, the present invention is for three-dimensional object modeling that can efficiently reduce polygons and obtain three-dimensional object modeling data from surface data composed of polygons without degrading the quality of the output modeling object. It is an object to provide a data optimization device.

  In order to solve the above-described problem, according to a first aspect of the present invention, there is provided an apparatus for optimizing surface data representing a structure of a three-dimensional object by an aggregate of polygons for modeling a three-dimensional object, Based on the area of each polygon constituting the data, the deletion target polygon selection means for selecting the polygon to be deleted as the deletion target polygon, and the side to be deleted based on the length of the side constituting the deletion target polygon. Deletion vertex shared polygon that is a polygon that shares one of the two vertices of the deletion target side (both end points of the side) from a polygon other than the deletion target polygon, and a deletion target side selection unit that selects as a deletion target side A deleted vertex shared polygon search means for searching for a deleted vertex shared polygon that shares both two vertices of the side to be deleted together with the deleted target polygon. By deleting and correcting one shared vertex to the position of the midpoint of the two vertices of the deletion target side with respect to the deleted vertex sharing polygon that shares only one vertex of the searched deletion target side A surface data updating means for updating, and when the number of polygons constituting the surface data updated by the surface data updating means is larger than a predetermined target value, the deletion target polygon selecting means, the deletion target edge selecting means, There is provided a three-dimensional object modeling data optimizing device comprising: a deleted vertex shared polygon search unit and a control unit that performs control to repeatedly execute processing on a surface data update unit.

  According to the first aspect of the present invention, the deletion target polygon is selected based on the area of each polygon constituting the surface data, and the deletion target side is selected based on the length of the side constituting the deletion target polygon. Then, from the polygons other than the deletion target polygon, a deleted vertex shared polygon that shares one of the two vertices of the deletion target side is searched, and the deleted vertex shared polygon that shares both the two vertices of the deletion target side is deleted. Surface data by deleting together with the target polygon and correcting one shared vertex to the midpoint of the two vertices of the deleted target side for the deleted vertex shared polygon that shares only one vertex of the deleted target side If the number of polygons constituting the updated surface data is greater than the predetermined target value, selection of polygons to be deleted, selection of edges to be deleted, and shared vertex for deletion Since the search for Lygon and the update of the surface data are performed repeatedly, the surface data composed of polygons can be used with less processing load that is simply proportional to the number of polygons without degrading the quality of the output molding. Therefore, it is possible to efficiently reduce polygons and obtain data for modeling a three-dimensional object. In particular, it is possible to efficiently output surface data generated based on medical CT / MRI images to a three-dimensional object forming apparatus such as a 3D printer.

  Further, in the second aspect of the present invention, the deletion target polygon selection means calculates an area and a perimeter for the polygon constituting the surface data, and selects a polygon having a minimum product of the area and the perimeter. It is selected as the polygon to be deleted.

  According to the second aspect of the present invention, the area and perimeter are calculated for the polygons constituting the surface data, and the polygon that minimizes the product of the area and perimeter is selected as the polygon to be deleted. Therefore, it is possible to select an appropriate polygon to be deleted without cutting a polygon that is important for forming a long and small surface.

  The third aspect of the present invention is characterized in that the deletion target side selecting means calculates the length of each side of the deletion target polygon and selects the shortest side as the deletion target side.

  According to the third aspect of the present invention, since the length of each side of the selected polygon to be deleted is calculated and the shortest side is selected as the deletion target side, it is important in forming the surface. It is possible to select an appropriate deletion target edge without cutting the edge.

  Further, in the fourth aspect of the present invention, there are a plurality of sets corresponding to the voxel format 3D medical image data divided into a plurality of regions, and the control means includes each surface data. On the other hand, the deletion target polygon selection means, the deletion target edge selection means, the deletion vertex shared polygon search means, and the surface data update means execute processing to create a plurality of sets of updated surface data. The method further comprises surface data synthesizing means for synthesizing the plurality of sets of updated surface data into a single surface data.

  According to the fourth aspect of the present invention, a plurality of sets exist corresponding to the voxel format 3D medical image data in which the surface data is divided into a plurality of regions, and the polygon to be deleted is selected for each surface data. , Select the deletion target edge, search the deleted vertex shared polygon, update the surface data, create multiple sets of updated surface data, and create multiple sets of updated surface data on a single surface Since it is synthesized with the face data, the series of processes are executed at high speed based on the surface data composed of a small number of polygons divided for each area, and the data processed for each part area of the human body is obtained. Surface data for batch output can be efficiently obtained as three-dimensional object modeling data.

  In the fifth aspect of the present invention, for the surface data to be executed by the deletion target polygon selection means, polygons belonging to each group constitute a polygon with any other polygon in the group. Surface data group classification means for grouping so as to share edges (two vertices), the deletion target polygon selection means for each of the classified groups, the deletion target edge selection means, a deletion vertex shared polygon search means, It further comprises group combining means for combining a plurality of groups of updated surface data obtained by execution of processing by the surface data updating means into a single surface data.

  According to the fifth aspect of the present invention, with respect to the surface data to be deleted polygon selection target, a polygon belonging to each group shares a side constituting the polygon with any other polygon in the group. The updated surface of the multiple groups created by selecting the deletion target polygon, selecting the deletion target edge, searching for the deleted vertex shared polygon, and updating the surface data for each classified group Since the data is combined into a single surface data, the number of polygons to be reduced at a time is reduced in units of groups, so that the overall processing speed can be increased.

  In the sixth aspect of the present invention, the group combining means calculates the volume of a circumscribed cuboid for the updated surface data of the plurality of groups, and the calculated volume is less than a predetermined threshold value. In this case, the surface data of the group is excluded from the synthesis targets.

  According to the sixth aspect of the present invention, the volume of a circumscribed cuboid is calculated for the updated surface data of a plurality of groups, and when the calculated volume is less than a predetermined threshold value, Since the group is synthesized by excluding the surface data from the synthesis target, the released small volume object is output by a 3D object shaping apparatus such as a 3D printer, and is clogged in the 3D object shaping apparatus. It is possible to reduce the occurrence of problems such as end.

  According to the present invention, three-dimensional object modeling data that can be efficiently output by a 3D printer by efficiently reducing polygons from surface data composed of polygons without degrading the quality of the output modeled object. Can be obtained.

It is a hardware block diagram of the data optimization apparatus for solid object modeling which concerns on one Embodiment of this invention. It is a functional block diagram which shows the structure of the data optimization apparatus for solid object shaping which concerns on one Embodiment of this invention. It is a figure which shows the mode of the conversion from DICOM format voxel data to surface data. It is the figure which showed the DICOM format voxel data as several slice images. This is an example in which DICOM format voxel data is displayed in four forms. FIG. 6 is a diagram showing STL data obtained by converting the DICOM format voxel data shown in FIG. 5 by “3D-Slicer”. It is a figure which shows the surface data according to area | region (STL data). It is a figure which shows the rendering image of the surface data (STL data) which synthesize | combined STL data according to area | region. It is a flowchart which shows the process outline | summary of the data optimization apparatus for solid object modeling which concerns on one Embodiment of this invention. 4 is a flowchart showing details of group classification processing by group classification means 20. It is a figure which shows the basic principle of polygon reduction. It is a flowchart which shows the detail of a polygon reduction process. It is a figure for demonstrating the problem which arises by a polygon reduction process. It is a flowchart which shows the detail of the deletion process of the isolated small volume group. It is a flowchart which shows the detail of the calculation process of the minimum space | interval in FIG.14 S44. It is a figure which shows the rendering image from the front of the surface data (STL data) shown in FIG. FIG. 17 is an enlarged view of the rendered image shown in FIG. 16. FIG. 17 is an enlarged view of the rendered image shown in FIG. 16. FIG. 17 is an enlarged view of the rendered image shown in FIG. 16. It is a wire frame figure of the rendering image enlarged view shown in FIG. It is a figure which shows the stepwise reduction example corresponding to a left eyeball among the surface data according to area | region shown in FIG. It is a figure which shows the stepwise reduction example corresponding to a left eyeball among the surface data according to area | region shown in FIG. It is a figure which shows the stepwise reduction example corresponding to a left eyeball among the surface data according to area | region shown in FIG. FIG. 22 is a wire frame diagram of the rendered image shown in FIG. 21. It is a wire frame figure of the rendering image shown in FIG. FIG. 24 is a wire frame diagram of the rendered image shown in FIG. 23.

DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described in detail with reference to the drawings.
<1. Device configuration>
FIG. 1 is a hardware configuration diagram of a three-dimensional object modeling data optimizing device according to an embodiment of the present invention. The three-dimensional object modeling data optimizing device according to the present embodiment can be realized by a general-purpose computer. As shown in FIG. 1, a CPU (Central Processing Unit) 1 and a RAM (Random) which is a main memory of the computer. Access Memory) 2, a large-capacity storage device 3 such as a hard disk or flash memory for storing programs and data executed by the CPU 1, a key input I / F (interface) 4 such as a keyboard and a mouse, and a 3D printer And a data input / output I / F (interface) 5 for data communication with an external device such as a data storage medium and a display unit 6 which is a display device such as a liquid crystal display, and are connected to each other via a bus. Yes. The 3D printer 7 is a general-purpose 3D printer, and is a three-dimensional object forming apparatus that forms a three-dimensional object by processing a material such as resin or gypsum based on data representing the three-dimensional object. The 3D printer 7 is connected to a data input / output I / F 5 and forms a three-dimensional object based on optimized output data received from the data input / output I / F 5.

  FIG. 2 is a functional block diagram showing the configuration of the three-dimensional object formation data optimizing device according to the present embodiment. In FIG. 2, 10 is a control means, 20 is a group classification means, 30 is a deletion target polygon selection means, 40 is a deletion target edge selection means, 50 is a deletion vertex shared polygon search means, 60 is a surface data update means, and 70 is Group combining means, 80 is surface data combining means, 90 is area-specific surface data storage means, and 100 is combined surface data storage means.

  The control means 10 controls the entire three-dimensional object modeling data optimizing device. In the group classification means 20, the polygons belonging to each group share the sides (two vertices) constituting the polygon with any other polygon in the group for the surface data to be selected as the deletion target polygon. Group as follows. The deletion target polygon selection means 30 calculates an area and a circumference for all the polygons constituting the surface data, and selects a polygon having a minimum product of the area and the circumference as a deletion target polygon. The deletion target side selecting means 40 calculates the length of each side of the deletion target polygon and selects the shortest side as the deletion target side. The deleted vertex shared polygon search means 50 searches for a deleted vertex shared polygon that shares one of the two vertices (end points of the side) of the deletion target side from all the polygons other than the deletion target polygon constituting the surface data. To do. The surface data update unit 60 deletes the deleted vertex shared polygon that shares both of the two vertices of the deletion target side together with the deletion target polygon, and applies to the deleted vertex shared polygon that shares only one vertex of the deletion target side. Thus, one shared vertex is corrected to the midpoint position of the two vertices of the deletion target side.

  The group synthesizing unit 70 is created by executing a series of processes by the deletion target polygon selecting unit 30, the deletion target side selecting unit 40, the deleted vertex shared polygon searching unit 50, and the surface data updating unit 60 for each classified group. The updated surface data of a plurality of groups is synthesized into a single area-specific surface data for each area. The surface data synthesizing unit 80 synthesizes a plurality of sets of updated area-specific surface data into a single synthesized surface data. The control means 10, group classification means 20, deletion target polygon selection means 30, deletion target edge selection means 40, deletion vertex shared polygon search means 50, surface data update means 60, group composition means 70, and surface data composition means 80 The CPU 1 is realized by executing a program stored in the storage device 3. The area-specific surface data storage unit 90 is a storage unit that stores area-specific surface data, and is realized by the storage device 3. The area-specific surface data is surface data configured with one area as one set. The combined surface data storage unit 100 is a storage unit that stores combined surface data that is a single surface data obtained by combining a plurality of sets of area-specific surface data, and is realized by the storage device 3. The composite surface data stored in the composite surface data storage unit 100 is solid object modeling data optimized for output to the 3D printer 7.

  Each component shown in FIG. 2 is actually realized by installing a dedicated program in hardware such as a computer and its peripheral devices as shown in FIG. That is, the computer executes the contents of each means according to a dedicated program. Note that in this specification, a computer means a device that has an arithmetic processing unit such as a CPU and is capable of data processing.

  In the storage device 3 shown in FIG. 1, a dedicated program for operating the CPU 1 and causing the computer to function as a three-dimensional object modeling data optimizing device is mounted. By executing this dedicated program, the CPU 1 controls the control means 10, the group classification means 20, the deletion target polygon selection means 30, the deletion target edge selection means 40, the deletion vertex shared polygon search means 50, and the surface data update means 60. The functions as the group synthesizing unit 70 and the surface data synthesizing unit 80 are realized. The storage device 3 not only functions as the area-specific surface data storage unit 90 and the composite surface data storage unit 100, but also stores various data necessary for processing as the three-dimensional object modeling data optimization device. .

<2. Processing action>
<2.1. Pretreatment>
Next, the processing operation of the three-dimensional object modeling data optimizing apparatus shown in FIGS. 1 and 2 will be described. First, pre-processing for creating region-specific surface data from medical voxel data will be described. The surface data is data representing a three-dimensional structure as a surface shape by surface modeling, and is composed of polygons that are polygon data. Preprocessing is performed by a computer executing a known program. Here, a case where “3D-Slicer” which is open source software developed mainly by Harvard University in the United States is used as a known program will be described as an example. First, the computer extracts voxel data of a specific organ region from the whole voxel data. Next, the computer converts the voxel data of the specific organ region into surface data. The area-specific surface data obtained in this manner is stored in the area-specific surface data storage means 90. The pre-processing may be performed by causing the computer that realizes the three-dimensional object modeling data optimizing device shown in FIG. 1 to execute the program, or the three-dimensional object modeling data optimizing device shown in FIG. It may be executed by a computer different from the computer to be realized, and the obtained area-specific surface data may be stored in the area-specific surface data storage means 90 realized by the storage device 3.

  In the present embodiment, voxel data in DICOM format, which is an international standard for medical image data exchange, is used as medical voxel data. Also, STL (Standard Triangulated Language) data is used as the surface data. STL data was originally an industry standard for modeling data exchange in the field of 3D graphics, and recently it has come to be used as a standard in 3D printers. Three-dimensional shapes are represented by a collection of triangular polygons. It is data. Then, when the computer executes “3D-Slicer”, the voxel data for each area as shown in FIG. 3B is extracted from the DICOM format voxel data as shown in FIG. By performing face model conversion, surface data (STL data) as shown in FIG. 3C is obtained. The Marching Cube method described in Patent Document 3 or the like is used for extraction of voxel data for each region and surface model conversion.

  4 to 8 show examples of processing screens by “3D-Slicer”. FIG. 4 is a diagram illustrating the DICOM format voxel data to be processed by “3D-Slicer” as a plurality of slice images. In the example of FIG. 4, the sagittal image of the human head is shown by 256 slice images. Each slice image is composed of 512 × 512 pixels.

  FIG. 5 is an example in which DICOM format voxel data is displayed in four forms. In FIG. 5, a volume rendering image is displayed at the top. The axial image, the sagittal image, and the coronal image are displayed in order from the left side.

  FIG. 6 is a diagram showing STL data obtained by converting the DICOM format voxel data shown in FIG. 5 by “3D-Slicer”. In FIG. 6, what is displayed at the top is a rendering image of STL data. In the example of FIG. 6, the eyeball / brain region threshold value: 626.75 or more (when the voxel value is 10 bits width in MRI imaging), the eyeball connective tissue L threshold value: 298.36-458.74, the eyeball connective tissue R threshold value: 298.36- By setting 431.32, the STL data classified by region classified by region (tissue) is rendered.

  FIG. 7 is a diagram in which the STL data for each region obtained by “3D-Slicer” is represented by shading by shading. FIG. 8 is a diagram showing a rendering image of the surface data (STL data) obtained by combining the STL data for each region. The Front View shown in FIG. 8 is the same as that shown in the upper part of FIG. After the preprocessing, the area-specific surface data as shown in FIG. 7 is stored in the area-specific surface data storage means 90.

<2.2. Process Overview>
Next, the processing operation of the three-dimensional object modeling data optimizing apparatus shown in FIGS. 1 and 2 will be described. FIG. 9 is a flowchart showing a processing outline of the three-dimensional object modeling data optimizing device according to the embodiment of the present invention. First, the group classification means 20 reads the area-specific surface data for one area from the area-specific surface data storage means 90, and classifies it into a plurality of groups (step S10). Next, the deletion target polygon selection means 30, the deletion target edge selection means 40, the deletion vertex shared polygon search means 50, and the surface data update means 60 execute polygon reduction processing on the surface data of each group ( Step S20). The control means 10 determines whether or not the polygon reduction process has been completed for all groups (step S30). If the polygon reduction process has not been completed for all groups, the polygon reduction process of step S20 is executed for the unprocessed group. To do. When the polygon reduction processing for all the groups is completed, the group composition unit 70 deletes the isolated small volume group (step S40). Subsequently, the group combining means 70 combines the surface data of all groups (step S50). The control means 10 determines whether or not the processing has been completed for all the areas (step S60). If the processing has not been completed for all the areas, the processing of steps S10 to S50 is executed for the unprocessed areas. When the processing for all areas is completed, the surface data synthesizing unit 80 synthesizes area-specific surface data for all areas (step S70).

<2.3. Group classification processing>
First, the group classification process by the group classification means 20 in step S10 will be described. FIG. 10 is a flowchart showing details of group classification processing by the group classification means 20. The group classification means 20 first performs an initial setting (step S11). Specifically, the number of polygons P constituting the read area-specific surface data, variables p (p = 0,..., P−1) for specifying polygons, search status S (p), group attribute G ( p) A variable g for specifying the group is set. The search status S (p) takes one of 0, 1, and 2. S (p) = 0 is a state not classified into a group, S (p) = 1 is a state classified into a group, and S (p) = 2 is a polygon adjacent to the polygon is classified into a group. Indicates the state that has been performed. Further, in step S11, S (p) = G (p) = 0 and g = 1 are initially set for all p of p = 0,..., P-1.

  Next, one polygon p satisfying S (p) = G (p) = 0 is extracted from all the polygons (step S12). In the initial state, all S (p) = G (p) = 0, so that the extraction is always performed at the first time. When there are a plurality of polygons p with S (p) = G (p) = 0, for example, the one having the smallest value of p is extracted. In step S12, when there is no polygon p satisfying S (p) = G (p) = 0, that is, when S (p) = 2 for all p = 0,..., P−1. , The group classification process is terminated assuming that it has been separated into g−1 groups. In step S12, since there is no state where S (p) = 1, S (p) is either “0” or “2”. When a polygon p satisfying S (p) = G (p) = 0 can be extracted, the search status S (p) = 1 corresponding to the extracted polygon p is set to G (p) = g.

  In step S12, one polygon p satisfying S (p) = G (p) = 0 is extracted, the search status S (p) = 1 corresponding to the extracted polygon p, and the group attribute G (p) = g are set. If set, one polygon q with S (q) = 1 is extracted from all polygons (step S13). q is a variable for specifying a polygon, and takes an integer value in q = 0,... When there are a plurality of polygons q with S (q) = 1, for example, the one having the smallest q value is extracted. In step S13, when there is no polygon q satisfying S (q) = 1, that is, when all S (q) = 0 or 2, the variable g for specifying the group is incremented by 1 (g ← g + 1). (Step S14). And it returns to step S12 and classify | categorizes into the following group.

  In step S13, when one polygon q satisfying S (q) = 1 is extracted, S (r) = G (r) = 0 from all the polygons, and two vertices are shared with the extracted polygon q. All adjacent polygons r are extracted (step S15). r is a variable for specifying a polygon, and takes an integer value in r = 0,..., P−1, similarly to polygons p and q. “Share two vertices with polygon q” means that one side is shared with polygon q. That is, the adjacent polygon r is a polygon adjacent to the polygon q by sharing one side. The extracted adjacent polygons are set to search status S (r) = 1 and group attribute G (r) = g, respectively.

  Next, the search status S (q) = 2 of the polygon q is set (step S16). That is, the polygon q is in a state after the adjacent polygon is set. Then, returning to step S13, one polygon q satisfying S (q) = 1 is extracted from all the polygons.

  In the flowchart shown in FIG. 10, there are two loop processes: a loop constituted by steps S12, S13, and S14 and a loop constituted by steps S13, S15, and S16. In the loop composed of steps S12, S13, and S14, polygons that are not classified into groups are extracted and classified into groups (step S12), and one of the polygons that have been classified into groups is searched for adjacent polygons. If it cannot be extracted as (step S13), it is assumed that there is no polygon to be classified into the group, and the process proceeds to the next group (g ← g + 1) (step S14). In the loop composed of steps S12, S13, and S14, a group of continuous polygons is classified as one group. Therefore, in step S13, if there is no polygon with S (q) = 1, there is no polygon adjacent to any polygon already classified in the group, so the classification into group g is finished. , The process proceeds to the group g + 1 classification.

  In the loop composed of steps S13, S15, and S16, one of the polygons classified into groups is extracted as a polygon q for searching for adjacent polygons (step S13), and polygon q and two vertices are extracted from all the polygons. Is extracted as the adjacent polygon r (step S15), and the search status S (q) of the search polygon q is set to the search of the polygon adjacent to the polygon q, whether or not it can be extracted. S (q) = 2, which is the state of

  By using two loop processes, a loop composed of steps S12, S13, and S14, and a loop composed of steps S13, S15, and S16, all P polygons are continuous polygons sharing two vertices. Each group is classified into groups. When only polygons that share only one vertex are adjacent to each other, they are classified as another group. By executing the processing according to FIG. 10, the polygons constituting the area-specific surface data are classified into a plurality of groups.

<2.4. Polygon reduction processing>
Next, the polygon reduction processing by the deletion target polygon selection unit 30, the deletion target side selection unit 40, the deletion vertex shared polygon search unit 50, and the surface data update unit 60 in step S20 will be described. FIG. 11 is a diagram showing the basic principle of polygon reduction. The surface data is actually a polygon group having a three-dimensional value in space, but the example of FIG. 11 shows a polygon group having a two-dimensional value for convenience of explanation. In the example of FIG. 11A, a polygon group consisting of 18 polygons before reduction is shown.

  FIG. 12 is a flowchart showing details of polygon reduction processing by the deletion target polygon selection unit 30, the deletion target side selection unit 40, the deletion vertex shared polygon search unit 50, and the surface data update unit 60. First, the deletion target polygon selection means 30 calculates the area and perimeter of each polygon for all the polygons constituting the surface data (step S21). Next, the deletion target polygon selection means 30 selects the polygon having the smallest value obtained by multiplying the calculated area and the circumference as the deletion target polygon (step S22). In the example of FIG. 11, as shown in FIG. 11B, a polygon EJK composed of three vertices E, J, and K is selected as a deletion target polygon.

  Subsequently, the deletion target side selecting means 40 selects the shortest one of the three sides of the deletion target polygon as the deletion target side, and calculates the midpoint of the selected deletion target side (step S23). In the example of FIG. 11, as shown in FIG. 11C, the side JK of the deletion target polygon EJK is selected as the deletion target side. Further, the midpoint O of the side JK is calculated.

  Next, the surface data update unit 60 deletes the deletion target polygon (step S24). As a result, the number of polygons is reduced by one. Next, the deleted vertex shared polygon search means 50 sequentially selects the deleted vertex shared polygon that shared the vertex with the deletion target side of the deletion target polygon (step S25). When a deleted vertex shared polygon that shares two vertices with the deletion target side of the polygon to be deleted (that is, the side is shared) is selected, the surface data update unit 60 deletes the deleted vertex shared polygon ( Step S26). In the example of FIG. 11, the polygon JKN composed of three vertices J, K, and N is deleted. This further reduces the number of polygons by one. When a deleted vertex shared polygon that shares one vertex with the deletion target side of the deletion target polygon is selected, the surface data update unit 60 corrects the shared vertex to the midpoint calculated in step S23, The area and circumference are recalculated (step S27). In the example of FIG. 11, as can be seen by comparing FIG. 11C and FIG. 11D, the polygon EHJ → polygon EHO, the polygon EFK → polygon EFO, the polygon HJM → polygon HOM, the polygon JMN → polygon OMN, and the polygon FKN. → Changed to polygon FON. As a result, in the example of FIG. 11, the 18 polygons at the time of FIG. 11A are reduced to 16 at the time of FIG. 11D.

  When the processing of step S25 to step S27 is completed, the control means 10 determines whether or not the total number of polygons after the reduction is equal to or less than the set target value (step S28). As the target value, an arbitrary value can be set in advance. As a result of the determination, if the total number of polygons is not less than or equal to the target value, the process returns to step S22, and the deletion target polygon selecting means 30 has the smallest value obtained by multiplying the remaining polygons by the area and the circumference. Is selected as a polygon to be deleted. If the result of determination in step S28 is that the total number of polygons is less than or equal to the target value, the polygon reduction process is terminated. By executing the processing according to the flowchart of FIG. 12, the number of polygons is reduced so as to be equal to or less than the target value set for each group.

<2.5. Small volume group deletion processing>
Next, the process of deleting the isolated small volume group by the group combining means 70 in step S40 will be described. Here, prior to the description of the processing for deleting the isolated small volume group, the reason for performing the processing for deleting the isolated small volume group will be described. FIG. 13 is a diagram for explaining a problem caused by the polygon reduction process. FIG. 13A shows two groups of polygon groups immediately after the group classification processing in step S10 is completed. In the example of FIG. 13A, there are a left group including vertices D, E, and K and a right group including vertices D ′, E ′, and K ′. If the vertices D and D ′, the vertices E and E ′, and the vertices K and K ′ are less than the resolution of the 3D printer, respectively, when these two groups of polygons are output by the 3D printer, the vertices D and D ′, E ′, vertices K and K ′ are connected by resin or the like, and the two groups are output as connected objects. That is, the two groups are physically bonded.

  When the triangle polygon reduction process in step S20 is executed on the right side group of polygons shown in FIG. 13A, the polygon E'FK 'is selected as the deletion target polygon as shown in FIG. 13B. Is done. Further, as shown in FIG. 13C, the shortest side E′F of the deletion target polygon E′FK ′ is selected as the deletion target side, and the midpoint O of the side E′F is calculated. Then, as shown in FIG. 13 (d), the deletion target polygon E'FK 'and the deletion vertex shared polygon D'E'F are deleted.

  As can be seen by comparing FIG. 13A and FIG. 13D, the two groups on the left and right sides are further separated after the polygon reduction processing than before the polygon reduction processing. When the distance between the two groups shown in FIG. 13D exceeds the resolution of the 3D printer, when the two groups of polygons are output by the 3D printer, the two groups are output as different objects. If the two groups are both large objects, there is no problem. However, when one becomes a very small object, for example, when one group is a main body and the other group is like a beard attached to the main body, the whiskers output separately from the main body are 3D. There is a risk that it may get into the printer and cause problems such as machine clogging. In order to eliminate such a problem, an isolated small volume group such as the above-mentioned beard is deleted.

  FIG. 14 is a flowchart showing details of the processing for deleting the isolated small volume group by the group combining means 70. First, the group composition means 70 calculates the maximum value and the minimum value in the X, Y, and Z directions for all vertices of all polygons for each group (step S41). As a result, for group g, minimum value Xmin (g), maximum value Xmax (g), minimum value Ymin (g), maximum value Ymax (g), minimum value Zmin (g), maximum value Zmax (g) Is obtained. Next, the group composition means 70 calculates the volume of the circumscribed cuboid for each group (step S42). Specifically, the volume V (g) of the circumscribed cuboid of the group g is calculated by executing the processing according to the following [Formula 1].

[Formula 1]
V (g) = (Xmax (g) −Xmin (g)) × (Ymax (g) −Ymin (g)) × (Zmax (g) −Zmin (g))

  The group synthesizing unit 70 executes the processes of steps S41 and S42 for all groups. Thereby, the volume of a circumscribed rectangular parallelepiped is obtained for all groups. Next, the group synthesizing unit 70 specifies a group in which the volume of the circumscribed rectangular parallelepiped is the smallest among all the groups (Step S43). Then, the group synthesizing unit 70 calculates the minimum interval between the identified group and all other groups (step S44). Details of the minimum interval calculation process in step S44 will be described later.

  Subsequently, the group composition means 70 determines whether or not the minimum interval calculated in step S44 is equal to or greater than a predetermined threshold value (step S45). The predetermined threshold for comparison with the minimum interval is preferably a numerical value corresponding to 0.2 mm at the time of 3D printer output. As a result of the determination, if the minimum interval is equal to or greater than a predetermined threshold, the group with the smallest volume is separated from the nearest group by a predetermined distance or more. Is deleted (step S46). If the result of determination in step S45 is that the minimum interval is less than the predetermined threshold, the group with the smallest volume is close to the closest group and below the resolution of the 3D printer. Does not perform the process of deleting the smallest volume group.

  When the group is deleted in step S46, the group is processed in any case where the deletion is not performed. And it returns to step S43 and specifies the group from which the volume of a circumscribed rectangular parallelepiped becomes the smallest among unprocessed groups. When the processing of step S43 to step S46 is repeated and the processing for all the groups is completed, the processing for deleting the isolated small volume group is ended.

  Details of the minimum interval calculation processing in step S44 will be described. FIG. 15 is a flowchart showing details of the minimum interval calculation processing in step S44 of FIG. The group synthesizing unit 70 sets a predetermined value or more as a variable dmin indicating the minimum interval as an initial value. As an initial value given to dmin, a large value that cannot be set as a minimum interval between groups may be set. Then, the group synthesizing unit 70 extracts one group h having a volume larger than that of the group g from all the groups (Step S47). As a result, the minimum value Xmin (h), the maximum value Xmax (h), the minimum value Ymin (h), the maximum value Ymax (h), the minimum value Zmin (h), and the maximum value Zmax (h) are obtained for the group h. It is done. Next, the group synthesizing unit 70 calculates the minimum value of the difference between the two groups for each of the X, Y, and Z directions (step S48). Specifically, the minimum value dx of the difference in the X direction, the minimum value dy of the difference in the Y direction, and the minimum value dz of the difference in the Z direction are calculated by executing processing according to the following [Equation 2]. .

[Formula 2]
dx = Min {| Xmax (g) -Xmin (h) |, | Xmin (g) -Xmax (h) |}
dy = Min {| Ymax (g) -Ymin (h) |, | Ymin (g) -Ymax (h) |}
dz = Min {| Zmax (g) -Zmin (h) |, | Zmin (g) -Zmax (h) |}

  Subsequently, the group synthesizing unit 70 sets the maximum among the calculated direction-specific minimum values as the inter-group minimum value, and compares the inter-group minimum value with the all-group minimum value among other groups that have already been processed. Then, the process of setting the smaller one as the minimum value between all groups is performed (step S49). Specifically, among the three direction-specific minimum values dx, dy, and dz calculated in step S48, the inter-group minimum value Max (dx, dy, dz) that takes the maximum value is compared with dmin, and Max ( When dx, dy, dz) <dmin, dmin = Max (dx, dy, dz). As a result, the minimum value between all groups, which is the minimum value among the minimum values between groups, is recorded in dmin.

  When the processes of steps S47, S48, and S49 are repeatedly executed and the processes for all the groups having a volume larger than the group g are finished, the minimum interval calculation process shown in FIG. 15 is terminated, and the process proceeds to step S45 of FIG. Thus, the minimum interval dmin is compared with the threshold value.

<2.6. Composition of surface data>
As shown in the processing outline of FIG. 9, in step S50, the group synthesizing unit 70 synthesizes the surface data of all the groups to obtain updated surface data by area. In step S70, the surface data synthesizing unit 80 synthesizes the area-specific surface data of all areas to obtain synthesized surface data. Any combination processing is performed by simply combining the surface data for each area excluding the header portion and recording the total number of polygons in the header portion. By outputting the composite surface data stored in the composite surface data storage unit 100 to the 3D printer 7, the 3D printer 7 can efficiently form a three-dimensional object. For the three-dimensional object that was shaped, there was almost no deterioration.

<3. Polygon reduction example>
FIGS. 16 to 26 show examples of surface data obtained by reducing polygons using the three-dimensional object modeling data optimizing apparatus according to the present embodiment. FIG. 16 is a diagram showing a rendering image from the front of the surface data (STL data) shown in FIG. FIG. 16A shows a rendered image of surface data before reduction, and FIG. 16B shows a rendered image of surface data after reduction. In the data after reduction shown in FIG. 16B, the number of polygons and the data amount are reduced to about 1/10 compared to the data before reduction in FIG. 16A, but there is no significant change in appearance. Similarly, there is no big change also about the solid thing modeled. There is substantially no deterioration in the three-dimensional object to be shaped.

  17 to 19 are enlarged views of the rendered image shown in FIG. FIG. 17 shows the smallest enlargement ratio, and FIG. 19 shows the largest enlargement ratio. 20 is a wire frame diagram of the enlarged rendering image shown in FIG. In FIG. 19, there is no significant change before and after the reduction, but comparing the wire frame diagrams of FIG. 20, it can be seen that the number of polygons is greatly reduced in FIG. 20B.

  FIG. 21 to FIG. 23 are diagrams showing a gradual reduction example corresponding to the left eyeball in the surface data by area (STL data) shown in FIG. 21 shows the state before reduction and 1/2 reduction, FIG. 22 shows the state reduced to 1/4 and 1/8, and FIG. 23 shows the state reduced to 1/10 and 1/20, respectively. ing. In the state reduced to 1/4 shown in FIG. 22A, an isolated small area is generated, and in the state reduced to 1/8 shown in FIG. 22B, the isolated small area disappears. 24 to 26 are wire frame diagrams of the rendered images illustrated in FIGS. 21 to 23.

<4. Modified example>
The preferred embodiments of the present invention have been described above. However, the present invention is not limited to the above embodiments, and various modifications can be made. For example, in the above embodiment, the processing target polygon is a triangle, but it may be a quadrilateral or more.

  Further, in the above embodiment, the area and perimeter are calculated for the polygon, and the polygon having the smallest product of the area and perimeter is selected as the polygon to be deleted, but only the area is used without using the perimeter. You may make it use.

  Further, in the above embodiment, the area and perimeter are calculated for the polygon, and the polygon having the smallest product of the area and perimeter is selected as the polygon to be deleted, but only the area is used without using the perimeter. You may make it use.

  Further, in the above embodiment, the circumscribed cuboid volume whose volume is less than the predetermined threshold is deleted from the updated surface data of a plurality of groups. Is possible.

1 ... CPU (Central Processing Unit)
2 ... RAM (Random Access Memory)
3 ... Storage device 4 ... Key input I / F
5. Data input / output I / F
6 ... Display unit 7 ... 3D printer (three-dimensional object modeling device)
DESCRIPTION OF SYMBOLS 10 ... Control means 20 ... Group classification | category means 30 ... Deletion target polygon selection means 40 ... Deletion target edge selection means 50 ... Deletion vertex shared polygon search means 60 ... Surface data update means 70... Group synthesizing means 80... Surface data synthesizing means 90... Area-specific surface data storing means 100.

Claims (7)

A device that optimizes the surface data representing the structure of a three-dimensional object as a collection of polygons for modeling a three-dimensional object,
A deletion target polygon selection means for selecting a polygon to be deleted as a deletion target polygon based on the area of each polygon constituting the surface data;
Based on the length of the sides constituting the deletion target polygon, a deletion target side selection means for selecting a side to be deleted as a deletion target side;
A deleted vertex shared polygon search means for searching for a deleted vertex shared polygon which is a polygon sharing one of the two vertices of the deletion target side from polygons other than the deleted target polygon;
The deleted vertex shared polygon that shares both two vertices of the deletion target side is deleted together with the deletion target polygon, and the deleted vertex shared polygon that shares only one vertex of the searched deletion target side, Surface data update means for updating the shared one vertex by correcting it to the position of the midpoint of the two vertices of the side to be deleted;
When the number of polygons constituting the surface data updated by the surface data update unit is larger than a predetermined target value, the deletion target polygon selection unit, the deletion target side selection unit, the deleted vertex shared polygon search unit, the surface Control means for controlling the data updating means to repeatedly execute processing;
A data optimizing device for modeling a three-dimensional object.   The deletion target polygon selection means calculates an area and a perimeter for the polygon constituting the surface data, and selects a polygon having a minimum product of the area and the perimeter as the deletion target polygon. The data optimization apparatus for modeling a three-dimensional object according to claim 1.   3. The three-dimensional object according to claim 1, wherein the deletion target side selecting unit calculates a length of each side of the deletion target polygon and selects the shortest side as the deletion target side. Modeling data optimization device. There are a plurality of sets corresponding to the voxel format 3D medical image data in which the surface data is divided into a plurality of regions,
The control means causes the deletion target polygon selection means, the deletion target edge selection means, the deletion vertex shared polygon search means, and the surface data update means to execute processing for each surface data, and sets a plurality of sets. To create updated surface data,
Surface data synthesizing means for synthesizing the plurality of sets of updated surface data into a single surface data;
The three-dimensional object modeling data optimizing device according to any one of claims 1 to 3, further comprising: For the surface data to be executed by the deletion target polygon selection means, the group classification is performed so that the polygons belonging to each group share the sides constituting the polygon with any other polygon in the group. A grouping means for face data;
Updated surfaces of a plurality of groups obtained by executing processing by the deletion target polygon selection means, the deletion target edge selection means, the deletion vertex shared polygon search means, and the surface data update means for each of the classified groups. A group combining means for combining data with a single surface data;
The data optimization device for three-dimensional object modeling according to any one of claims 1 to 4, further comprising:   The group synthesizing unit calculates a volume of a circumscribed rectangular parallelepiped for the updated surface data of the plurality of groups, and when the calculated volume is less than a predetermined threshold value, the surface data of the group The data optimizing device for three-dimensional object modeling according to claim 5, wherein   A program for causing a computer to function as the three-dimensional object modeling data optimizing device according to any one of claims 1 to 6.