Similar drawing search system, similar drawing search method, and program

Abstract

A similar drawing search system, a similar drawing search method, and a program are provided that can improve search accuracy. [Solution] The similar drawing search system disclosed herein is a similar drawing search system that searches for drawings similar to a target drawing from among a plurality of pre-stored candidate drawings, and is equipped with a control unit that executes a feature estimation process that estimates feature information of a target object included in the target drawing, an individual similarity calculation process that compares the feature information of the target object with the feature information of all candidate objects included in each of the plurality of candidate drawings to calculate individual similarities for each of the candidate objects, an overall similarity calculation process that calculates overall similarities for each candidate drawing using the individual similarities of all candidate objects included in each candidate drawing, and an output data generation process that generates output data based on the overall similarities.

Description

[Technical Field]

[0001] The present disclosure relates to a similar drawing search system, a similar drawing search method, and a program Regarding. [Background technology]

[0002] In recent years, methods have been proposed for searching past drawings for drawings similar to a target drawing in order to improve work efficiency by utilizing large amounts of drawing data created in the past (for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2024-000702 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the above-mentioned Patent Document 1 aims to search for similar drawings using a simple method, and there is room for improvement in terms of search accuracy.

[0005] Therefore, the present disclosure has been made in consideration of the above-mentioned problems, and its purpose is to provide a similar drawing search system, a similar drawing search method, and a program that can improve search accuracy. [Means for solving the problem]

[0006] According to the present disclosure, there is provided a similar drawing search system for searching for a drawing similar to a target drawing from among a plurality of pre-stored candidate drawings, the system comprising: a feature estimation process for estimating feature information of a target object included in the target drawing; an individual similarity calculation process for calculating an individual similarity for each of the candidate objects by comparing feature information of the target object with feature information of all of the candidate objects included in each of the plurality of candidate drawings; a comprehensive similarity calculation process for calculating a comprehensive similarity for each candidate drawing using the individual similarities of all candidate objects included in each candidate drawing; an output data generation process for generating output data based on the overall similarity; A similar drawing search system is provided, which includes a control unit that executes the above.

[0007] According to the present disclosure, there is also provided a similar drawing search method for searching for a drawing similar to a target drawing from among a plurality of pre-stored candidate drawings, the method comprising: a feature estimation process for estimating feature information of a target object included in the target drawing; an individual similarity calculation process for calculating an individual similarity for each of the candidate objects by comparing feature information of the target object with feature information of all of the candidate objects included in each of the plurality of candidate drawings; a comprehensive similarity calculation process for calculating a comprehensive similarity for each candidate drawing using the individual similarities of all candidate objects included in each candidate drawing; an output data generation process for generating output data based on the overall similarity; A similar drawing search method is provided in which the control unit of the similar drawing search system executes the above.

[0008] According to the present disclosure, there is also provided a program for a similar drawing search system that searches for a drawing similar to a target drawing from among a plurality of pre-stored candidate drawings, the program comprising: a feature estimation process for estimating feature information of a target object included in the target drawing; an individual similarity calculation process for calculating an individual similarity for each of the candidate objects by comparing feature information of the target object with feature information of all of the candidate objects included in each of the plurality of candidate drawings; a comprehensive similarity calculation process for calculating a comprehensive similarity for each candidate drawing using the individual similarities of all candidate objects included in each candidate drawing; an output data generation process for generating output data based on the overall similarity; A program is provided that causes a control unit of a similar drawing search system to execute the above. [Effects of the Invention]

[0009] According to the present disclosure, it is possible to provide a similar drawing search system, a similar drawing search method, and a program that can improve search accuracy. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a diagram illustrating an example of the configuration of a similar drawing search system according to an embodiment of the present disclosure. [Figure 2] FIG. 2 is a diagram illustrating an example of the configuration of a similar drawing search device according to the embodiment. [Figure 3] FIG. 2 is a diagram illustrating an example of the configuration of a similar drawing search device according to the embodiment. [Figure 4] FIG. 4 is a flowchart illustrating a series of controls in the system according to the embodiment. [Figure 5] FIG. 10 is a diagram illustrating an example of calculation of individual similarity of an object according to the embodiment. [Figure 6] FIG. 10 is a diagram illustrating an example of calculation of individual similarity of an object according to the embodiment. [Figure 7] FIG. 10 is a diagram showing an example of output data according to the embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0011] Preferred embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings. In this specification and drawings, components having substantially the same functional configurations are designated by the same reference numerals, and redundant description will be omitted.

[0012] FIG. 1 shows an example of an information processing system (similar drawing search system) 1 according to this embodiment. This system 1 is a system that calculates the similarity of each drawing in order to search for a drawing similar to a target drawing from among a plurality of pre-stored candidate drawings. This system 1 includes an information processing device (similar drawing search device) 10 and a user terminal 20, and is capable of communication via a network NW. Note that the target "drawing" may be, for example, a drawing of a part used in various devices, and the target "object" may be, for example, a front view, side view, plan view, oblique view, cross-sectional view, etc. of the part, but is not limited thereto.

[0013] The information processing device 10 is a server device used by a system administrator or the like when operating and managing various services, and may be, for example, a general-purpose computer such as a workstation or personal computer, or may be logically realized using cloud computing technology.

[0014] The information processing device 10 includes a control unit 11, a storage unit 12, an input unit 13, an output unit 14, and a communication unit 15, as shown in FIG.

[0015] As shown in Fig. 3, the user terminal 20 includes a control unit 21, a storage unit 22, an input unit 23, an output unit 24, and a communication unit 25. The user terminal 20 is a computer operated by a user for inputting and outputting various types of information. For example, the user terminal 20 is a smartphone, a tablet computer, a personal computer, or the like. The user can access the information processing device 10 by, for example, an application or a web browser executed on the user terminal.

[0016] When the information processing device 10 receives various commands (requests) from other information processing devices or the like via the input unit 13 or the communication unit 15, the control unit 11 executes processing according to a program, and the program processing results (e.g., images, sounds, etc.) are sent to the output unit 14 or other information processing devices or the like. Alternatively, the information processing device 10 receives various commands (requests) from other information processing devices or the like via the communication unit 15, and transmits the program processing results executed by the control unit 11 to the other information processing devices or the like. Note that part of the program may be transmitted to the other information processing devices and executed on the other information processing devices. In this case, the other information processing devices (user terminals 2) may be, for example, smartphones, mobile phone terminals, tablet terminals, personal computers, etc., and are connected to the information processing device 10 wirelessly or via a wire via a network such as the Internet.

[0017] The control units 11 and 21 transfer data between each unit and control the entire device, and are realized, for example, by a CPU (Central Processing Unit), MPU (Micro Processing Unit), or GPU (Graphics Processing Unit) executing a program stored in a specified memory.

[0018] The storage units 12 and 22 store various data and programs, and are non-volatile or volatile semiconductor memories such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable Read Only Memory), and EEPROM (Electrically Erasable Programmable Read-Only Memory), as well as magnetic disks, flexible disks, optical disks, compact disks, minidisks, DVDs (Digital Versatile Discs), etc.

[0019] The input units 13 and 23 are used by users and system administrators to input various data, and are realized by, for example, a keyboard, a mouse, a touch panel, buttons, a microphone, and the like.

[0020] The output units 14 and 24 output various information generated by the control unit, etc. The output units are, for example, a liquid crystal display (LCD), a touch panel, a printer, a speaker, etc.

[0021] The communication units 15 and 25 are for communicating with other information processing devices, and have a function as a receiving unit that receives various data and signals transmitted from other information processing devices, etc., and a function as a transmitting unit that transmits various data and signals to other information processing devices, etc. in response to commands from the control unit. The communication units are realized by, for example, a NIC (Network Interface Card), an adapter for connecting to Ethernet (registered trademark), a modem for connecting to a public telephone line network, a wireless communication device for wireless communication, a USB (Universal Serial Bus) connector or an RS232C connector for serial communication, etc.

[0022] The control unit 11 can execute a feature estimation process (step S1) that estimates feature information of a target object included in a target drawing, an individual similarity calculation process (step S2) that compares the feature information of the target object with the feature information of all candidate objects included in each of a plurality of candidate drawings to calculate individual similarities for each candidate object, an overall similarity calculation process (step S3) that calculates overall similarities for each candidate drawing using the individual similarities of all candidate objects included in each candidate drawing, and an output data generation process (step S4) that generates output data based on the overall similarities.

[0023] In this way, by calculating the overall similarity for each candidate drawing based on individual comparisons between the target object and all candidate objects, the accuracy of similar drawing searches can be improved.

[0024] FIG. 4 shows an example of the flow of information processing executed in this system.

[0025] For example, user A shown in FIG. 1 may access the information processing device 10 via the user terminal 20 and select any target drawing data from multiple drawing data pre-stored in the storage unit 12. Alternatively, the information processing device 10 may accept an input from user A specifying a specific area in the target drawing, and only objects included in the specific area may be recognized as target objects. The specific area may be specified, for example, by user A inputting a frame of a specific shape (e.g., rectangle, circle, etc.) that surrounds any object in the target drawing. By user A specifying the target object in this way, the convenience of the similar drawing search function can be improved, and since objects unnecessary for user A are not selected as target objects, the processing load of similarity calculation can be reduced. Alternatively, user A may access the information processing device 10 via the user terminal 20 and select any multiple candidate drawings from multiple drawing data pre-stored in the storage unit 12, or the control unit 11 may automatically select multiple candidate drawings. When the control unit 11 automatically selects, all drawings stored in the storage unit 12 may be used as candidate drawings, or drawings with common information associated with the target drawing may be selected. Specifically, for example, drawings that share "part name" information with the target drawing may be selected as candidate drawings. Similarly, drawings that share one or more items of information, such as "company name," "product name," and "direction attribute (front view, plan view, etc.)," ​​associated with each drawing may be selected. In this case, annotation information (attribute information) indicating the item may be associated with each drawing data (image data).

[0026] The method of acquiring or selecting the target drawing data is not particularly limited, and may be, for example, image data of the target drawing photographed by a camera of the user terminal 20, or data of the target drawing captured by a user A or the like using a scanner or the like. The control unit 11 can store data of the target drawing or the like sent from the user terminal 20 in the storage unit 12.

[0027] In the feature estimation process (step S1), the control unit 11 estimates feature information of a target object included in a target drawing. The feature information can be, for example, a feature amount (feature vector), but is not limited to this. The feature information is data that is uniquely determined according to at least the shape of the target object.

[0028] The method for calculating the feature information is not particularly limited. For example, the feature information can be calculated (inferred) by inputting data of the connected region forming the target object into a feature inference model, and the resulting data can be output. The inference model may be, for example, one that applies a neural network or the like, but any machine learning model can be used. Such an inference model is pre-stored in a storage unit or stored in an external information processing system that can communicate via a communication unit. The number of inference models is not limited to one. For example, multiple inference models with different conditions, such as differences in machine learning methods or data, may be stored and used selectively or in parallel. The feature inference model is machine-learned so that, for example, the more similar the target object (target connected region) and the candidate object (candidate connected region) are, the higher the similarity when comparing the features. The similarity is defined, for example, by the distance when comparing the feature values ​​of images. For example, a distance index such as Euclidean distance or Manhattan distance, or a similarity index such as cosine similarity may be used.

[0029] The feature (feature information) is output as vector data of a fixed-length numeric array, but is not limited to vector format and may be output in other data formats. The feature may be, for example, SIFT feature, SURF feature, ORB feature, AKAZE feature, etc.

[0030] Here, the target drawing data may include or be associated in advance with connected areas and feature information of the target objects included in the target drawing, or the control unit 11 may be able to detect the target objects included in the target drawing.

[0031] When the control unit 11 detects a target object, for example, it may extract a connected area (which can be, for example, a circular or polygonal annular area, but may also be a shape with some discontinuous parts rather than a completely continuous shape) defined by multiple pixels included in each drawing whose brightness values ​​are consecutive pixels that have a brightness value equal to or greater than a predetermined value (i.e., multiple pixels that form a continuous line), and detect the connected area as an object.

[0032] In addition, in the process of detecting such a target object, the brightness values ​​of multiple pixels included in each drawing may be binarized. "Binarization" refers to the process of converting, for example, pixels in each drawing that are below a predetermined brightness threshold into white and pixels that exceed the threshold into black.

[0033] Furthermore, the control unit 11 may perform a line thickening (dilation) process on each drawing. For example, the control unit 11 may thicken lines by converting or maintaining all pixels adjacent to pixels with a luminance value of "gray to black, intermediate between white and black" (pixels whose luminance exceeds a threshold) before binarization to a luminance value of "black," or by converting or maintaining pixels adjacent to pixels with a luminance value of "black" after binarization to a luminance value of "black." Furthermore, in addition to the adjacent pixels, pixels close to pixels with a predetermined luminance value before or after binarization may be thickened to a predetermined line thickness by converting them to black. By thickening lines, it is possible to connect unintentionally broken (disconnected) lines on the drawing, for example, because the original drawing's color is too light or the lines are too thin, thereby improving the accuracy of detecting the target object (the accuracy of extracting connected regions). The control unit 11 may perform the various processes described above on either or both of the target drawing and the candidate drawing. Condition information and other information required for executing each process are stored in advance in the storage unit 12.

[0034] In the individual similarity calculation process (step S2), the control unit 11 compares the feature information of the target object with the feature information of all candidate objects included in each of the plurality of candidate drawings, and calculates the individual similarity for each candidate object.

[0035] 5, the target drawing 30A includes two target objects 41 and 42. The target object 41 may be, for example, a plan view of a part, and the target object 42 may be a front view of the part. The target drawing may include only one target object, or three or more target objects.

[0036] 5 shows a case where similarity is calculated for three candidate drawings 30B, 30C, and 30D. Candidate drawing 30B includes one candidate object 43, candidate drawing 30C includes two candidate objects 44 and 45, and candidate drawing 30D includes two candidate objects 46 and 47. Note that target drawing 30A and candidate drawings 30B, 30C, and 30D each include a table element T in which text is surrounded by a rectangular frame, but annotation information prevents control unit 11 from recognizing table element T as a target object.

[0037] The control unit 11 compares the feature amount of the target object 41 with the feature amounts of all the target objects 43 to 47 in all the candidate drawings, and calculates the individual similarity of each of the target objects 43 to 47 with respect to the target object 41. The individual similarity is expressed, for example, as a value between 0 and 1, with the higher the similarity being expressed as a numerical value approaching 1. The value of the individual similarity is not limited to this, and may be set between a lower limit value and an upper limit value so that the higher the similarity is, the closer it is to the upper limit value, or the closer it is to the lower limit value.

[0038] Furthermore, the control unit 11 compares the feature amount of the target object 42 with the feature amounts of all the target objects 43 to 47 in all the candidate drawings, and calculates the individual similarity of each of the target objects 43 to 47 to the target object 42. In this example, since there are two target objects 41 and 42 in the target drawing 30A, two individual similarity values ​​are assigned to each of the candidate objects 43 to 47. If the number of candidate objects differs for each candidate drawing, the number of individual similarities will differ for each candidate drawing.

[0039] Then, in the overall similarity calculation process (step S3), the control unit 11 calculates the overall similarity for each candidate drawing using the individual similarities of all the candidate objects included in each candidate drawing.

[0040] Information such as conditions for calculating the overall similarity is stored in advance in the storage unit. The method for calculating the overall similarity is not particularly limited as long as it uses information about individual similarities. For example, the control unit 11 calculates the overall similarity by applying individual similarity information for all candidate objects included in each candidate drawing to a predetermined calculation formula. Specifically, the control unit 11 may calculate the average value of all individual similarities for all candidate objects included in the candidate drawing to obtain the overall similarity, or may further apply the average value to a predetermined formula to calculate an index to obtain the overall similarity. For example, the average value is preferably the harmonic mean, but arithmetic mean, geometric mean, etc. may also be used. The overall similarity may be calculated using any similarity evaluation mechanism. The overall similarity may not necessarily be calculated using the average value of the individual similarities. For example, a model for predicting similarity may be constructed by combining the complexity of the shape of the drawing or vectorized data with metadata (data about the data). More specifically, the objects in each drawing may be structured (graphed) and the structures may be compared to determine the partial similarity or the overall similarity of the structures.

[0041] The overall similarity for each candidate drawing is determined by the overall similarity calculation process (step S3).

[0042] Then, in the output data generation process (step S4), the control unit 11 generates output data based on the overall similarity.

[0043] FIG. 7 is an example of a screen displayed on the user terminal 20. The control unit 11 generates a screen as output data in which the candidate drawings are arranged in descending order of overall similarity. The control unit may extract and output only a predetermined number of candidate drawings with high overall similarity, rather than all candidate drawings. As shown in FIG. 7, the control unit 11 may display the ranking of overall similarity or the overall similarity. For each candidate object included in each candidate drawing, all or part of the individual similarity (e.g., the highest similarity value for each object) may be displayed. The control unit may also calculate the difference between each candidate object and the target object in the target drawing and display the difference as an image. The individual and overall similarity values ​​shown in FIGS. 5, 6, and 7 are illustrative and do not strictly correspond to the object shapes shown.

[0044] In this embodiment, the overall similarity calculation process may calculate the overall similarity for each candidate drawing by a harmonic mean process using the individual similarities of all candidate objects included in each candidate drawing. Using the harmonic mean can improve the accuracy of detecting drawings showing highly similar solids (e.g., components) compared to using the arithmetic mean. For example, if there are three candidate objects (Drawings 1 to 3, e.g., side views, plan views, etc.), and only one of them (e.g., Drawing 1) is similar and the others are dissimilar, using the arithmetic mean will have a greater influence (individual similarities) on the values ​​of the dissimilar drawings (e.g., Drawings 2 and 3) and will likely be determined to be dissimilar. However, using the harmonic mean will have a greater influence on the similar drawing (Drawing 1), making it more likely to be determined to be similar. This allows for accurate retrieval of drawings with high similarity in solid form (shape) even when the candidate objects include drawings (candidate objects) that are the same (or similar) as solids but viewed from different directions.

[0045] In the present embodiment, the control unit may execute a ranking process for ranking the plurality of candidate drawings based on the overall similarity, and generate the output data based on the ranking, thereby making it possible to present candidate drawings with high similarity to the user in an easy-to-understand manner.

[0046] In this embodiment, the output data may include information indicating the individual similarity for each candidate object. This allows the similarity for each candidate object to be presented to the user. The output data may include all of the similarities (individual similarities) for each candidate object, or may present only some of the numerical values, such as the highest individual similarity. For example, in the examples of FIGS. 5 and 6, information on two individual similarities (an individual similarity with respect to the target object 41 and an individual similarity with respect to the target object 42) is associated with one candidate object 43, but the highest individual similarity may be displayed for each candidate object 43.

[0047] In this embodiment, the control unit may perform the following process in the feature estimation process: binarizing the brightness values ​​of multiple pixels included in the target drawing; and recognizing the target object by extracting a connected area defined by pixels in which the line drawing brightness values ​​used to draw the line drawing are consecutive, out of the two binarized brightness values.

[0048] In this embodiment, the control unit may perform a thickening process to thicken the lines included in the target drawing before or after the binarization process, and recognize the target object by extracting the connected area after the thickening process, thereby improving the detection accuracy of the target object (the extraction accuracy of the connected area).

[0049] The control unit 11 may perform annotation processing on each drawing. For example, the control unit 11 may perform image analysis on each drawing data to associate attribute information of the areas of figures, lines, and characters drawn within the frame of the drawing (for example, inside the outermost rectangular frame along the outer edge of the drawing) and store the associated information in the storage unit. The attribute information may include, but is not limited to, round parts, non-round parts (such as squares and polygons), tables, annotation text, seals, etc. Furthermore, handwritten characters may be flagged as "handwritten," and if the drawing is unclear or distorted, a "low quality" flag may be flagged, and the information may be associated with the drawing and stored in the storage unit.

[0050] The control unit 11 may output the annotation results and display them on the user terminal 20. Furthermore, the control unit 11 may accept editing instructions from the user via the user terminal 20, allowing the user to modify the range of the area of ​​the drawing to which attribute information is assigned or the type of attribute. This can improve the recognition accuracy of elements included in the drawing. It is preferable that such processing be performed before the control unit calculates the feature information of each drawing. The annotation processing may be performed using a predetermined trained model. A trained model may be used that has previously trained a dataset (combination) of image data of elements that may be included in the drawing (round-shaped components, non-round-shaped components, tables, annotation text, stamps, etc.) and correct attribute information. Furthermore, the user may input the dataset again to perform re-training to improve accuracy.

[0051] Although the preferred embodiments of the present disclosure have been described in detail above with reference to the accompanying drawings, the technical scope of the present disclosure is not limited to such examples. It is clear that a person skilled in the art of the present disclosure can conceive of various modified or altered examples within the scope of the technical idea described in the claims, and it is understood that these also naturally fall within the technical scope of the present disclosure.

[0052] The devices described in this specification may be realized as a single device, or may be realized by a plurality of devices (e.g., cloud servers) partly or entirely connected via a network. For example, the control unit 11 and the storage unit 12 of the information processing device 10 may be realized by different servers connected to each other via a network.

[0053] The series of processes performed by the device described in this specification may be realized using software, hardware, or a combination of software and hardware. A computer program for realizing each function of the device and terminal according to this embodiment may be created and installed on a PC or the like. A computer-readable recording medium on which such a computer program is stored may also be provided. Examples of the recording medium include a magnetic disk, an optical disk, a magneto-optical disk, and a flash memory. The computer program may also be distributed, for example, via a network, without using a recording medium.

[0054] Furthermore, the processes described herein using flowchart diagrams do not necessarily have to be performed in the order shown. Some process steps may be performed in parallel. Additional process steps may be employed, and some process steps may be omitted.

[0055] Furthermore, the effects described herein are merely descriptive or exemplary and are not limiting. In other words, the technology according to the present disclosure may achieve other effects that will be apparent to those skilled in the art from the description of this specification, in addition to or in place of the above-described effects.

[0056] The following configurations also fall within the technical scope of the present disclosure. (Item 1) A similar drawing search system that searches for a drawing similar to a target drawing from among a plurality of pre-stored candidate drawings, a feature estimation process for estimating feature information of a target object included in the target drawing; an individual similarity calculation process for calculating an individual similarity for each of the candidate objects by comparing feature information of the target object with feature information of all of the candidate objects included in each of the plurality of candidate drawings; a comprehensive similarity calculation process for calculating a comprehensive similarity for each candidate drawing using the individual similarities of all candidate objects included in each candidate drawing; an output data generation process for generating output data based on the overall similarity; A similar drawing search system comprising a control unit that executes the above. (Item 2) Item 1: A similar drawing search system according to item 1, wherein the overall similarity calculation process calculates the overall similarity for each candidate drawing by performing a harmonic mean process using the individual similarities of all candidate objects included in each candidate drawing. (Item 3) The control unit executes a ranking determination process to determine the ranking of the plurality of candidate drawings based on the overall similarity, and generates the output data based on the ranking. (Item 4) 3. The similar drawing search system according to item 1 or 2, wherein the output data includes information indicating an individual similarity for each of the candidate objects. (Item 5) The control unit, in the feature estimation processing, A process of binarizing the luminance values ​​of a plurality of pixels included in the target drawing; and (3) extracting a connected area defined by pixels in which one of the two binarized luminance values ​​used for drawing a line drawing is continuous, thereby recognizing the target object. (Item 6) the control unit performs a thickening process to thicken lines included in the target drawing before or after the binarization process, Item 6. The similar drawing search system according to item 5, wherein the target object is recognized by extracting the connected area after the thickening process. (Item 7) A similar drawing search method for searching for a drawing similar to a target drawing from among a plurality of pre-stored candidate drawings, comprising: a feature estimation process for estimating feature information of a target object included in the target drawing; an individual similarity calculation process for calculating an individual similarity for each of the candidate objects by comparing feature information of the target object with feature information of all of the candidate objects included in each of the plurality of candidate drawings; a comprehensive similarity calculation process for calculating a comprehensive similarity for each candidate drawing using the individual similarities of all candidate objects included in each candidate drawing; an output data generation process for generating output data based on the overall similarity; A similar drawing search method, in which the control unit of a similar drawing search system executes the above. (Item 8) A program for a similar drawing search system that searches for a drawing similar to a target drawing from among a plurality of pre-stored candidate drawings, a feature estimation process for estimating feature information of a target object included in the target drawing; an individual similarity calculation process for calculating an individual similarity for each of the candidate objects by comparing feature information of the target object with feature information of all of the candidate objects included in each of the plurality of candidate drawings; a comprehensive similarity calculation process for calculating a comprehensive similarity for each candidate drawing using the individual similarities of all candidate objects included in each candidate drawing; an output data generation process for generating output data based on the overall similarity; A program that causes a control unit of a similar drawing search system to execute the above. [Explanation of symbols]

[0057] 1 Similar drawing search system 10 Similar drawing search device (server) 11 Control section 12 Storage section 20 User terminal

Claims

1. A similar drawing search system that searches for a drawing similar to a target drawing from among a plurality of pre-stored candidate drawings, a feature estimation process for estimating feature information of a plurality of target objects included in the target drawing; an individual similarity calculation process for comparing feature information of the target object with feature information of all of a plurality of candidate objects included in each of the plurality of candidate drawings, and calculating an individual similarity of each of the candidate objects with respect to each of the target objects; a comprehensive similarity calculation process for calculating a comprehensive similarity for each candidate drawing using all individual similarities of all candidate objects included in each candidate drawing; an output data generation process for generating output data based on the overall similarity; a control unit that executes the the overall similarity calculation process calculates an overall similarity for each candidate drawing by a harmonic mean process using individual similarities of all candidate objects included in each candidate drawing; The control unit selects and presents the highest individual similarity from among a plurality of individual similarities for each of the candidate objects for the generated output data.

2. The similar drawing search system according to claim 1 , wherein the control unit executes a ranking determination process for determining a ranking of the plurality of candidate drawings based on the overall similarity, and generates the output data based on the ranking.

3. The control unit, in the feature estimation processing, A process of binarizing the luminance values ​​of a plurality of pixels included in the target drawing; The similar drawing search system of claim 1 or 2 executes a process of recognizing the target object by extracting a connected area defined by pixels in which the line drawing brightness value used to draw the line drawing is consecutive, out of the two binarized brightness values.

4. the control unit performs a thickening process to thicken lines included in the target drawing before or after the binarization process, The similar drawing search system according to claim 3 , wherein the target object is recognized by extracting the connected area after the thickening process.

5. A similar drawing search method for an information processing device that searches for a drawing similar to a target drawing from among a plurality of pre-stored candidate drawings, comprising: a feature estimation process for estimating feature information of a plurality of target objects included in the target drawing; an individual similarity calculation process for comparing feature information of the target object with feature information of all of a plurality of candidate objects included in each of the plurality of candidate drawings, and calculating an individual similarity of each of the candidate objects with respect to each of the target objects; a comprehensive similarity calculation process for calculating a comprehensive similarity for each candidate drawing using all individual similarities of all candidate objects included in each candidate drawing; an output data generation process for generating output data based on the overall similarity; is executed by a control unit of the information processing device, the overall similarity calculation process calculates an overall similarity for each candidate drawing by a harmonic mean process using all individual similarities of all candidate objects included in each candidate drawing; The control unit selects and presents the highest individual similarity from among a plurality of individual similarities for each of the candidate objects for the generated output data.

6. A program for a similar drawing search system that searches for a drawing similar to a target drawing from among a plurality of pre-stored candidate drawings, a feature estimation process for estimating feature information of a plurality of target objects included in the target drawing; an individual similarity calculation process for comparing feature information of the target object with feature information of all of a plurality of candidate objects included in each of the plurality of candidate drawings, and calculating an individual similarity of each of the candidate objects with respect to each of the target objects; a comprehensive similarity calculation process for calculating a comprehensive similarity for each candidate drawing using all individual similarities of all candidate objects included in each candidate drawing; an output data generation process for generating output data based on the overall similarity; The control unit of the similar drawing search system executes the above. the overall similarity calculation process calculates an overall similarity for each candidate drawing by a harmonic mean process using all individual similarities of all candidate objects included in each candidate drawing; The control unit selects and presents the highest individual similarity from among a plurality of individual similarities for each of the candidate objects for the generated output data.

Images