JP7807475B2 - Search device, search method, and information processing system - Google Patents

Description

The present invention relates to a search device, a search method, and an information processing system.

For example, companies that own manufacturing equipment such as machine tools may search their own drawings for ones similar to those requested by a customer, and create estimates based on these drawings. For this reason, a device that searches for drawings similar to those requested by a customer has been proposed (Patent Document 1). Furthermore, a device has also been proposed that iconizes component drawing areas corresponding to keywords specified by the user and places them in a virtual space to provide the user with search clues (Patent Document 2).

Japanese Patent Application Laid-Open No. 2023-55219 Japanese Patent Application Laid-Open No. 2004-30122

The technology described in Patent Document 1 allows users to search for similar company drawings based on a 3D or plan view provided by a customer, but it is difficult to use because it does not support data other than drawings. The technology described in Patent Document 2 requires users to specify an appropriate key image, making it difficult to use.

The present invention was made in consideration of the above-mentioned problems, and its purpose is to provide a highly user-friendly search device, search method, and information processing system that can search image data based on three-dimensional information about the search target.

To solve the above problem, a search device according to one aspect of the present invention is a search device that outputs search results in response to a search request, and includes an image data registration unit that vectorizes feature quantities based on the characteristics of image data and registers them in a search database, and an image search unit that vectorizes feature quantities based on characteristics extracted from three-dimensional information of a search target object to generate a search request, and searches the search database using the generated search request.

According to the present invention, image data registered in a search database can be searched based on three-dimensional information about the search target.

FIG. 1 is an overall configuration diagram of a search system. FIG. 2 is a hardware configuration diagram of the search system. This is a conceptual diagram showing how various search requests are input into the search system to search for similar product drawings. FIG. 10 is an explanatory diagram showing how features of a search object are extracted from three-dimensional data of the search object. FIG. 10 is an explanatory diagram showing how features appear differently depending on the viewpoint. This is an explanatory diagram showing how the same search object is imaged from multiple viewpoints, the difference in feature amount between each image and image data in a database is calculated as a distance based on the features of each image, and the search results are output in order of shortest distance. FIG. 10 is a sequence diagram showing the flow of processing in the search system. FIG. 10 is an explanatory diagram of an information processing system according to a second embodiment. FIG. 10 is an explanatory diagram of a search system according to a third embodiment.

The following describes an embodiment of the present invention based on the drawings. In this embodiment, drawing data such as plan views or three-dimensional views is used as an example of "image data." Feature quantities based on the characteristics of the drawing data are calculated, vectorized, and registered in a search database. The search target is an object for which drawing data similar to its shape is being sought. Examples of objects include various products or product models. Products may be prototypes, custom products, or mass-produced products. Product models include exterior models, miniatures, and samples. Products include finished products and parts. For example, three-dimensional information of the search target is acquired using a three-dimensional information acquisition device such as SfM (Structure from Motion), MVS (Multi-View Stereo), or LiDAR (Light Detection and Ranging). Features are extracted from the acquired three-dimensional information, feature quantities are calculated, and the calculated feature quantities are vectorized. By searching the search database using a search request based on the three-dimensional information of the search target, drawing data similar to the search target is found.

The device that acquires three-dimensional information about an object (search target) may be a portable device, or it may be a device that is difficult to transport and is installed in a facility such as a factory or laboratory.

In the search device of this embodiment, feature quantities based on the characteristics of the drawing data (image data) are vectorized and registered in advance in a search database. When searching for drawing data related to a search target, three-dimensional information is acquired for the search target itself or a model of the search target. The search database is then searched using a search request based on the acquired three-dimensional information. Therefore, even if drawing data for the search target does not exist or only a model of the search target is available, it is possible to search for drawing data similar to the search target. Furthermore, by linking the drawing data with a manufacturing device capable of manufacturing the object represented by the drawing data, it is possible to easily identify the manufacturing device capable of manufacturing the search target, and to calculate an estimate for manufacturing the object (product) related to the search target using the identified manufacturing device.

The first embodiment will be described using Figures 1 to 7. Figure 1 is an overall configuration diagram of search system 1 as an example of a "search device." In this embodiment, a case will be described in which drawing data previously registered in a search database is searched for by a search request based on three-dimensional information about the search target. In the following, user terminal U1 used by a user who registers drawing data in search system 1 and user terminal U2 used by a user who searches for drawing data similar to the search target will be described as separate terminals, but user terminal U1 and user terminal U2 may be the same. In other words, the user who registers drawing data and the user who searches for drawing data similar to the search target may belong to the same organization. Hereinafter, user terminal U1 that registers drawing data in search system 1 may be referred to as registration user terminal U1, and user terminal U2 used to search for registered drawing data may be referred to as search user terminal U2.

The search system 1 includes, for example, an image data collection unit 11, an image data registration unit 12, an index database 13, a project information database 14, a three-dimensional information reception unit 15, and an image search unit 16. The image data collection unit 11 and the image data registration unit 12 are responsible for indexing the drawing data. The three-dimensional information reception unit 15 and the image search unit 16 are responsible for searching the indexed drawing data.

The image data collection unit 11 collects drawing data as "image data." The image data collection unit 11 collects drawing data as image data from the web server 111 or the file server 112. A person who registers drawing data can use a registration user terminal U1 to send the drawing data to the image data collection unit 11 via the web server 111. For example, a person who registers drawing data may input image data obtained by photographing the drawing data from the web server 111 to the image data collection unit 11. The drawing data may also be input to the image data collection unit 11 via a file server 112 connected to a CAD (Computer Aided Design) system (not shown), for example.

The image data registration unit 12 extracts feature quantities from the image data collected by the image data collection unit 11 using the feature point extraction unit 121, vectorizes the extracted feature quantities using the vectorization processing unit 122, and registers the vectorized quantities in the index database 13. The image data registration unit 12 registers project information, including drawing data, information on manufacturing equipment capable of manufacturing the object shown in the drawing data, a parts list, a cost list, etc., in the project information database 14.

The index database 13 is an example of a "search database." As described above, the index database 13 stores drawing data DW (see Figure 3) of products (search targets, objects) that can be manufactured by the manufacturing equipment 3 (see Figure 3) indexed based on feature quantities. The project information database 14 stores information related to the manufacture and sale of products indicated by the drawing data. Information related to the manufacture and sale of products includes, for example, a parts list, a materials list, a cost list, a sales price, and a delivery date.

The three-dimensional information receiving unit 15 receives three-dimensional information of the search object 2 as an image that serves as the basis for the search. The three-dimensional information of the search object 2 can be generated using a device that acquires three-dimensional information about objects. Known technologies used in such devices include, for example, SfM (Structure from Motion), MVS (Multi-View Stereo), and LiDAR (Light Detection and Ranging).

The device for acquiring three-dimensional information about an object may be built into the search user terminal U2 or connected to the search user terminal U2. The three-dimensional information about the search object 2 can be input from the user terminal U2 to the three-dimensional information accepting unit 15. Alternatively, if the creator of the search object 2 (e.g., the end user) is provided with a device for creating three-dimensional information about objects, the three-dimensional information about the search object 2 can be acquired at the creator of the search object 2, the acquired three-dimensional information can be fixed on a storage medium, brought into the search system 1, and input to the three-dimensional information accepting unit 15. As in the example described below, the three-dimensional information about the search object 2 can also be input to the three-dimensional information accepting unit 15 via the Internet.

The image search unit 16 generates a search request based on the three-dimensional information of the search object 2, and searches the index database 13 to extract drawing data similar to the search object 2. The image search unit 16 includes, for example, a 3D preprocessing unit 161, an image preprocessing unit 162, and a search unit 163.

The 3D preprocessing unit 161 performs preprocessing of three-dimensional information. The 3D preprocessing unit 161 includes, for example, an opacity processing unit 1611 and a multi-viewpoint processing unit 1612. Hereinafter, the opacity processing unit 1611 may be abbreviated as "opacity 1611," and the multi-viewpoint processing unit 1612 may be abbreviated as "multi-viewpoint 1612."

The opacity processing unit 1611 performs opacity processing on the three-dimensional information of the search object 2 so that only the surface as seen from the viewpoint is displayed. In other words, the three-dimensional information that has been opacified makes it impossible to see the point cloud inside or behind it.

The multi-viewpoint processing unit 1612 generates an image when the opaque-processed three-dimensional information is viewed from multiple viewpoints. Examples of processing by the opaque processing unit 1611 and the multi-viewpoint processing unit 1612 will be described later.

The image preprocessing unit 162 preprocesses the image data processed by the 3D preprocessing unit 161 in order to generate a search request. Image preprocessing includes, for example, the extraction of feature points and the vectorization of the extracted feature points. By extracting feature points from the three-dimensional information of the search object 2 and vectorizing these feature points, it is possible to search for vectorized drawing data in the index database 13.

The search unit 163 generates a search request based on the vectorized data and searches the index database 13. The search results for the drawing data are sent to the search user terminal U2 along with related information read from the project information database 14.

Figure 2 is a hardware configuration diagram of the search system 1. The search system 1 can be generated using a single computer 1000. Alternatively, the search system 1 can be generated by linking multiple computers.

The computer 1000 includes, for example, a processor 1100, memory 1200, an auxiliary storage device 1300, a communication unit 1400, and a user interface unit 1500. In the figure, the user interface is abbreviated as UI.

The processor 1100 executes computer programs stored in the auxiliary storage device 1300 to realize the various functions of the search system 1. The processor 1100 is not limited to a CPU (Central Processing Unit), but may also include a GPU (Graphics Processing Unit) or an ASIC (Application Specific Integrated Circuit).

Memory 1200 includes RAM (Random Access Memory) and ROM (Read Only Memory).

The communication unit 1400 is a device for communicating with user terminals U1 and U2 and other systems not shown in the figure via the communication network CN. Examples of other systems not shown in the figure include a design support system, a production management system, a sales management system, an estimate creation system, and a document management system. The type of communication network CN is not important. The communication network CN may be a public line or a dedicated line.

The user interface unit 1500 is a device for exchanging information between users of the search system 1 (those who register drawing data and those who search for drawing data) and the search system 1. The user interface unit 1500 may be equipped with an information input device and an information output device (neither of which are shown). Information input devices include, for example, keyboards, voice input devices, pointing devices, and touch panels. Information output devices include, for example, monitor displays, printers, and synthesized voice output devices. Users and the search system 1 may also exchange information using technologies such as Virtual Reality (VR), Augmented Reality (AR), Mixed Reality (MR), and Extended Reality (XR).

The auxiliary storage device 1300 is a relatively large-capacity storage device such as a semiconductor memory device, hard disk device, optical disk device, or magnetic tape device. The auxiliary storage device 1300 stores specific computer programs for implementing the functions of the image data collection unit 11, image data registration unit 12, three-dimensional information reception unit 15, and image search unit 16. The auxiliary storage device 1300 can also store an index database 13 and a project information database 14. These databases 13 and 14 may be located in a computer not shown.

Search system 1 can be connected to storage medium MM. Storage medium MM is configured, for example, as a memory device, hard disk device, optical disk device, magneto-optical disk device, magnetic tape device, etc., and stores computer programs and data non-temporarily. Storage medium MM can transfer computer programs and data to auxiliary storage device 1300 of search system 1 for storage. Computer programs and data can also be transferred from auxiliary storage device 1300 to storage medium MM for storage. A computer program that realizes the functions of search system 1 can be stored on storage medium MM, and by connecting storage medium MM to another computer and installing the computer program stored on storage medium MM on the other computer, the other computer can function as search system 1.

Figure 3 is a conceptual diagram showing how various search requests are input into the search system 1 to search for drawing data similar to the search target object 2.

As mentioned above, the index database 13 is pre-registered with drawing data DW for products that can be manufactured by the manufacturing equipment 3.

The manufacturing equipment 3 is, for example, a processing or production device such as an injection molding machine, NC machine, or electric discharge machine. The manufacturing equipment 3 is used to manufacture finished products or parts thereof, such as automobiles, trains, aircraft, ships, robots, home appliances, and transportation equipment. The manufacturing equipment 3 can be used in any industry. The multiple manufacturing equipment 3(1) to 3(n) shown in the figure will be referred to as manufacturing equipment 3 unless otherwise distinguished.

The three-dimensional information receiving unit 15 receives three-dimensional information of the search object 2, such as paper drawings, digitalized drawings, samples such as specimens and models, photographs, and point cloud data. The image search unit 16 extracts features from the three-dimensional information received by the three-dimensional information receiving unit 15 and generates vectorized data from them. The searched product drawing data is then sent to the user terminal U2, which is the source of the search, along with information obtained from the project information database 14. The project information database 14 is not shown in Figure 3.

Figure 4 is an explanatory diagram showing how features of the search object 2 are extracted from the three-dimensional information 21 of the search object 2. The upper part of Figure 4 shows point cloud data of the search object 2 as three-dimensional information 21. The three-dimensional information 21 of the search object 2 has a shape similar to, for example, a chair, with a protrusion 211 formed on the underside of the backrest. The three-dimensional information 21 includes many features, such as intersections. The protrusion 211 is shown as one of these features.

The lower part of Figure 4 shows a 3D image 22 that has been opaque-processed. The 3D image 22 also has a protrusion 221 formed on the underside of the backrest.

Figure 5 is an explanatory diagram showing how features appear differently depending on the viewpoint. The lower left side of Figure 5 shows a 3D image 22(3) of the search object 2 viewed from the back. The lower right side of Figure 5 shows a 3D image 22(1) of the search object 2 viewed from the right side. As shown in the rear image 22(3), protrusions 221L and 221R protrude from the left and right sides of the underside of the backrest. However, in the image 22(1) viewed from the right side, protrusion 221R blends into the background and is difficult to distinguish. Thus, even for the same 3D information 21 of the search object 2, the appearance of the features varies depending on the viewing angle. Therefore, in this embodiment, the features of the search object 2 are captured from multiple viewpoints and vectorized, making it possible to search for similar drawing data from registered drawing data.

Figure 6 is an explanatory diagram showing how the same search object 2 is imaged from multiple viewpoints, the distance between each image 22 and the image data (drawing data) in the index database 13 is calculated based on the characteristics of each image 22, and the search results are output in order of shortest distance. Here, 3D images are shown as seen from a total of four viewpoints: right side image 22(1), left side image 22(2), back image 22(3), and front image 22(4). Additionally, top and bottom images (neither of which are shown) can also be added.

In addition to images 22 from predetermined viewpoints such as front, back, top, bottom, left, and right, or instead of all or part of the images 22 from predetermined viewpoints, images 22 from any viewpoint may be used. A searcher operating the search user terminal U2 can also generate images 22 from an angle that best shows the characteristics of the search object 2.

The central part of Figure 6 shows the relationship between image 22 of the search object 2 and image data (drawing data) registered in the index database 13 that is similar to image 22.

For example, when focusing on the right side image 22(1) of the search target object 2, image data (drawing data) DW11, DW12, and DW13 similar to the right side image 22(1) are extracted. The distances indicating the similarity between the right side image 22(1) and the similar image data DW11, DW12, and DW13 are d11, d12, and d13.

Similarly, for left side view image 22(2) of search object 2, image data DW21, DW22, and DW23 similar to left side view image 22(2) are extracted. The distances indicating the similarity between left side view image 22(2) and similar image data DW21, DW22, and DW23 are d21, d22, and d23.

Similarly, when focusing on rear image 22(3) of search target object 2, image data DW31, DW32, and DW33 similar to rear image 22(3) are extracted. The distances indicating the similarity between rear image 22(3) and similar image data DW31, DW32, and DW33 are d31, d32, and d33.

Focusing on the left side image 22(2) of the search target object 2, image data DW41, DW42, and DW43 similar to the front image 22(4) are extracted. The distances indicating the similarity between the front image 22(4) and the similar image data DW41, DW42, and DW43 are d41, d42, and d43.

A search request is generated for each of the images 22(1) to 22(4) obtained from these multiple viewpoints, and the index database 13 is searched. Images DW similar to the search target object 2 are then extracted for each of the multiple viewpoints. The image search unit 16 extracts similar images that are close in distance (images with a high degree of similarity) from the results of the simultaneous search for similar images from multiple viewpoints, and provides these to the search user terminal U2.

Figure 7 is a sequence diagram showing the processing of the search system 1. First, image data registration processing S1 to S7 is performed to register image data (drawing data) in the index database 13. Next, image search processing S11 to S17 is performed to search the index database 13 based on the three-dimensional information of the search target object 2.

In the image data registration process, the registration user terminal U1 instructs the image data registration unit 12 to register image data (drawing data) via the image data collection unit 11 (S1). The feature point extraction unit 121 of the image data registration unit 12 extracts feature points contained in the image data and calculates feature quantities (S2). The vectorization processing unit 122 converts the calculated feature quantities into n-dimensional vector data (S3) and generates an index (S4).

The image data registration unit 12 registers the generated index in the index database 13 (S5). The image data registration unit 12 links the image data registered in the index database 13 with information in the project information database 14 related to this image data (S6), and stores this in the project information database 14 (S7).

In the image search process, a search instruction is input from the search user terminal U2 to the image search unit 16 via the three-dimensional information receiving unit 15 (S11). This search instruction includes three-dimensional information about the search target object 2.

The searching user terminal U2 can also clearly indicate important features of the search object 2 to the image search unit 16. For example, the searching user terminal U2 can mark the most distinctive part of the three-dimensional information of the search object 2, thereby notifying the image search unit 16 that this feature is important.

The image search unit 16 can set the priority of search results for search requests from a viewpoint that includes a feature with a high degree of importance among multiple viewpoints higher than the priority of search results from other viewpoints, and output search results according to similarity and priority. The searching user terminal U2 may instruct the image search unit 16 to search without explicitly specifying important features.

The 3D preprocessing unit 161 of the image search unit 16 performs 3D preprocessing, such as opacity processing and multi-viewpoint extraction processing, on the three-dimensional information of the search target object 2 (S12). The image preprocessing unit 162 of the image search unit 16 extracts feature points from the 3D preprocessed image, calculates feature amounts of the extracted feature points (S13), and generates n-dimensional vector data from the calculated feature amounts (S14). Note that when generating vector data, the image preprocessing unit 162 may set a high priority for feature amounts with high importance (feature amounts of feature points with high importance).

The search unit 163 generates a search request based on the vectorized data and searches for image data registered in the index database 13 (S15).

When the image search unit 16 obtains the search results from the index database 13, it obtains information related to the search results from the project information database 14 (S16). The image search unit 16 associates the search results from the index database 13 with the related information obtained from the project information database 14, transmits them to the search user terminal U2, and presents them to the user (S17).

With the search system 1 of this embodiment configured as described above, when searching for drawing data related to the search object 2, three-dimensional information is acquired about the search object itself or a model of the search object, and the index database 13 is searched using a search request based on the acquired three-dimensional information. Therefore, even if no drawing data for the search object 2 exists or if only a model of the search object 2 is available, it is possible to search for image data (drawing data) similar to the search object 2.

In the search system 1 of this embodiment, a search request is generated for each image of the search object 2 viewed from multiple viewpoints, and the index database 13 is searched simultaneously using the multiple search requests. Therefore, the search system 1 can search the index database 13 by capturing the characteristics (feature points) of the search object 2, and can find image data similar to the search object 2.

The search system 1 of this embodiment can easily identify a manufacturing device 3 capable of manufacturing the search target 2 by linking the manufacturing device capable of manufacturing the object indicated in the drawing data with that drawing data. Therefore, the search system 1 can also calculate an estimate for manufacturing the object (product) related to the search target using the identified manufacturing device 3.

The search system 1 of this embodiment can clearly indicate important feature points when three-dimensional information about the search target object 2 is input into the search system 1. Therefore, the search system 1 can search for image data (drawing data) that is similar based on important feature points. This makes it possible to search the index database 13 with an emphasis on key manufacturing aspects (such as how to handle cutting tools, how easy it is to remove a mold, and how easy it is to inject molten resin), further improving usability for users looking for drawing data.

Example 2 will be described using Figure 8. In the following examples, including this example, differences from Example 1 will be mainly described. Figure 8 is an explanatory diagram of an information processing system according to Example 2.

In this embodiment, search system 1A is connected to embedded system 4. In other words, search system 1A is used as one of the functions of embedded system 4. Embedded system 4 is, for example, a sales support system, quotation creation support system, production management system, design support system, etc.

This embodiment, configured in this manner, also achieves the same effects as embodiment 1. Because the search system 1A of this embodiment is incorporated as one of the functions of the embedded system 4, the convenience of the embedded system 4 can be improved.

Figure 9 is an explanatory diagram of a search system 1B according to a third embodiment. The search system 1B of this embodiment further includes an external publication unit 17 in addition to the configuration of the search system 1 described in the first embodiment.

The external publishing unit 17 has the function of publishing the search system 1B on the communications network CN. For example, the registration user terminal U1 and the search user terminal U2 can access the search system 1B via the communications network CN and register image data (drawing data) in the search system 1B or retrieve image data similar to the search target 2 from the search system 1B. The search system 1B may be made public 24 hours a day, 365 days a year via the external publishing unit 17. This improves user convenience.

Furthermore, in the search system 1B of this embodiment, the material providing unit 181 and the estimating unit 182 are connected to the project information database 14. The material providing unit 181 provides materials such as catalogs, explanatory videos, and specifications for the manufacturing equipment 3 to the project information database 14. The estimating unit 182 creates or assists in the creation of an estimate when the manufacturing equipment 3 manufactures an object shown in the drawing data.

This embodiment, configured in this manner, also achieves the same effects as embodiment 1. The search system 1B of this embodiment is made public on the communications network CN via the external publication unit 17, improving usability for users. Furthermore, in this embodiment, document data and estimates for the manufacturing equipment 3 related to the drawing data can be provided to the search user terminal U2, improving convenience for the user who is the searcher.

The present invention is not limited to the above-described embodiments and includes various modifications. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and are not necessarily limited to those including all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, or to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is possible to add, delete, or replace part of the configuration of each embodiment with the configuration of another embodiment. Furthermore, each embodiment can be combined as appropriate, as long as there is no obvious contradiction.

In this embodiment, the apparatus or method described below is clearly described in an operable manner.

(Representation 1) A search device that outputs search results in response to a search request, comprising: an image data registration unit that vectorizes feature quantities based on the characteristics of image data and registers them in a search database; and an image search unit that vectorizes feature quantities based on characteristics extracted from three-dimensional information of a search target object to generate a search request, and searches the search database using the generated search request.

(Representation 2) The image search unit creates search requests corresponding to images of three-dimensional information of the search object viewed from multiple viewpoints, and searches the search database using the search requests for each of the multiple viewpoints.

(Representation 3) The search device described in Representation 1 or 2, wherein the image search unit performs opacity processing on the three-dimensional information of the search object, creates search requests corresponding to images of the opacified three-dimensional information of the search object viewed from the multiple viewpoints, and searches the search database using the search requests for each of the multiple viewpoints.

(Representation 4) A search device according to any one of Representations 1-3, wherein the image search unit outputs the search results for each of the multiple search requests according to the degree of similarity between the search results and the three-dimensional information of the search target.

(Representation 5) A search device according to any one of Representations 1-4, wherein the image search unit sets a higher priority for search results for a search request from a viewpoint that includes a feature of high importance among the multiple viewpoints than for search results from other viewpoints, and outputs the search results according to the similarity and the priority.

(Representation 6) A search device described in any one of Representations 1-5, in which the three-dimensional information of the search object is generated using a three-dimensional information acquisition method for acquiring three-dimensional information.

(Representation 7) A search device according to any one of Representations 1-6, in which the search object is a physical sample of an object.

(Representation 8) A search device according to any one of Representations 1-7, wherein the image data registered in the search database is two-dimensional image data.

(Representation 9) A search device described in any one of Representations 1-8, in which the image data registered in the search database is associated with a manufacturing device capable of manufacturing an object related to the image data.

(Representation 10) A search method for outputting search results in response to a search request from a search device, the search method comprising: an image data registration step in which the search device vectorizes feature quantities based on the characteristics of image data and registers the vectorized feature quantities in a search database; and an image search step in which the search device vectorizes feature quantities based on characteristics extracted from three-dimensional information of a search target to generate a search request, and searches the search database using the generated search request.

(Representation 11) An information processing system equipped with the search device described in any one of Representations 1-9, which outputs information regarding the production or sales of the search target based on the search results output by the search device.

1, 1A, 1B: Search system, 2: Search object, 3: Manufacturing equipment, 4: Embedded system, 11: Image data collection unit, 12: Image data registration unit, 13: Index database, 14: Project information database, 15: Three-dimensional information reception unit, 16: Image search unit, U1: User terminal for registration, U2: User terminal for search

Claims (8)

A search device that outputs search results in response to a search request,
an image data registration unit that vectorizes feature quantities based on the features of the image data and registers the vectorized feature quantities in a search database;
an image search unit that generates a search request by vectorizing a feature quantity based on features extracted from three-dimensional information of a search target object, and searches the search database using the generated search request;
the image search unit creates the search requests corresponding to images of three-dimensional information of the search object viewed from a plurality of viewpoints, and searches the search database using the search requests for each of the plurality of viewpoints;
a degree of similarity between the search results for the search requests for each of the plurality of viewpoints and the three-dimensional information of the search object;
a priority set so that a search result in response to a search request from a viewpoint including a feature quantity with high importance among the plurality of viewpoints is higher than a search result from other viewpoints;
and outputting the search results according to the
Search device. The search device described in claim 1, wherein the image search unit performs opacity processing on the three-dimensional information of the search object, creates search requests corresponding to images of the opacified three-dimensional information of the search object viewed from the multiple viewpoints, and searches the search database using the search requests for each of the multiple viewpoints. The search device according to claim 1 , wherein the three-dimensional information of the object to be searched is generated using a three-dimensional information acquisition method for acquiring three-dimensional information. 4. The search device according to claim 3 , wherein the search object is a physical specimen of an object. 5. The search device according to claim 4 , wherein the image data registered in the search database is two-dimensional image data. 6. The search device according to claim 5 , wherein the image data registered in the search database is associated with a manufacturing device capable of manufacturing an object related to the image data. A search method for outputting search results in response to a search request from a search device, comprising:
The search device includes:
an image data registration step of vectorizing feature quantities based on the features of the image data and registering the vectorized feature quantities in a search database;
an image search step of vectorizing feature quantities based on features extracted from the three-dimensional information of the object to be searched to generate a search request, and searching the search database using the generated search request ;
the image search step generates the search requests corresponding to images of three-dimensional information of the search object viewed from a plurality of viewpoints, and searches the search database using the search requests for each of the plurality of viewpoints;
a degree of similarity between the search results for the search requests for each of the plurality of viewpoints and the three-dimensional information of the search object;
a priority set so that a search result in response to a search request from a viewpoint including a feature quantity with high importance among the plurality of viewpoints is higher than a search result from other viewpoints;
The search results are output in accordance with the
How to search. An information processing system comprising the search device according to any one of claims 1 to 6 ,
An information processing system that outputs information relating to the production or sales of the search object based on the search results output by the search device.