Object recognition system, object recognition program and object recognition method

JP2023095748A5Active Publication Date: 2024-12-05MEIJI UNIV +2
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
JP2022112137
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2022-07-13
Publication Date
2024-12-05
Estimated Expiration
2041-12-24

AI Technical Summary

Technical Problem

Existing object recognition systems struggle to accurately identify similar objects based solely on their appearance from a single direction, especially when products or parts have similar shapes, leading to inefficiencies in customer identification and purchase processes.

Method used

An object recognition system that registers and utilizes multiple appearance images of an object from different characteristic angles, incorporating three-dimensional shape information and feature images to enhance identification accuracy, and provides user instructions for optimal image capture.

Benefits of technology

Enhances object recognition accuracy by leveraging multifaceted appearance images and user-guided image capture, enabling precise identification of similar objects and facilitating efficient customer interactions, such as purchases or repairs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

To provide a novel object recognition system capable of further appropriately identifying an object even when it is assumed that there are similar objects therein.SOLUTION: An object recognition system for estimating which one of registered objects registered in a database is a target object included in an object image captured by the user, includes acquisition means for acquiring the object image, registration means for registering a plurality of external appearance images representing external appearances of one object when viewed from different feature angles, in the database for every registered object, and estimation means for estimating which one of the registered objects is the target object based on the object images and the external appearance images.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0005] , ,

[0001] The present invention relates to an object recognition system, an object recognition program, and an object recognition method for recognizing, from an image of an object, which of the registered objects the object is among the registered objects registered in advance.

Background Art

[0002] For example, in the case of industrial products and their parts, there may be a need to reorder due to damage, loss, etc. However, especially for specialized products and parts, even if one has the actual item, in many cases, the product name, model number, etc. are unknown, and since there are many similar products and parts, there has been a problem that consumers cannot identify the necessary products. Especially recently, many products are sold through the Internet, but when consumers themselves cannot identify the target product, ultimately, they have to go to the store with the actual item for individual consultations or exchange multiple times with attached images, and sufficient convenience has not been ensured.

[0003] Regarding such problems, a technique for identifying products registered in a dictionary in advance from the external appearance image of an article is known. For example, Patent Document 1 discloses a technique for capturing information indicating the external appearance of an article, extracting a plurality of feature amounts thereof, and comparing them with the feature amounts of each product to recognize product candidates.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] For example, as in Patent Document 1, if it is sufficient to select candidates from the target products, product candidates can be identified relatively easily from the appearance of the items, and even if there are multiple candidates, the product can be appropriately identified by having the user make a selection. However, when identifying products or parts that often have similar shapes, there is a problem that simply registering the appearance viewed from one direction may not be sufficient, as there are many similar objects and identification may not be possible.

[0006] Therefore, the object of the present invention is to provide a novel object recognition system that can more appropriately identify objects even when similar objects are anticipated. [Means for solving the problem]

[0007] To solve the above problems, the present invention provides an object recognition system for estimating which of the registered objects registered in a database is a target object included in an object image taken by a user, comprising: acquisition means for acquiring the object image; registration means for registering a plurality of appearance images, each representing the appearance of a single object when viewed from a different characteristic angle, as appearance information for each registered object in the database; and estimation means for estimating which of the registered objects the target object is based on the object image and the appearance information.

[0008] This configuration allows for more accurate object recognition based on multi-angle images of each object's appearance.

[0009] In a preferred embodiment of the present invention, the registration means includes means for generating a plurality of appearance images based on information indicating the three-dimensional shape of the registered object, and registering the appearance information based on the generated appearance images in the database for each registered object.

[0010] This configuration allows for the registration of external images based on 3D shape information, making it possible to easily create a database of registered objects without the need to take numerous external images for each registered object.

[0011] In a preferred embodiment of the present invention, the registration means includes means for acquiring an image of the registered object and registering the image as the appearance image for each registered object in the database.

[0012] This configuration allows for the registration of images obtained by actually photographing an object as an appearance image, and enables matching with similar appearance images of the object acquired by the user.

[0013] In a preferred embodiment of the present invention, the registration means generates or acquires an image of the appearance of the registered object with attached material on its surface, and registers the appearance information based on the generated image in the database.

[0014] When users actually photograph objects, the subjects may include objects with dirt, scratches, rust, or other contaminants. By registering images of objects with such contaminants in the database, it is expected that the system will be able to properly recognize objects even when they have such contaminants.

[0015] In a preferred embodiment of the present invention, the registration means registers the appearance information in association with the information indicating the characteristic angle.

[0016] With this configuration, for example, if information about the shooting angle is obtained from the user along with multiple object images, it becomes possible to identify objects based on changes in appearance according to the angle, thereby improving accuracy.

[0017] In a preferred embodiment of the present invention, the registration means is configured to register a feature image containing information for distinguishing the registered object from other objects in association with the registered object, and the estimation means identifies candidates for the registered object that may be the target object and generates shooting instruction information to instruct the user to photograph the target object based on the feature image of the registered object.

[0018] This configuration allows users to register images containing distinctive shapes or unique markings not found in other registered objects as feature images, enabling them to take photos that capture those specific features. This allows for more accurate identification of the target object, even when multiple similar objects exist.

[0019] In a preferred embodiment of the present invention, if the estimation means cannot identify a candidate of the registered object that may be the target object, it generates a shooting instruction information to instruct the user to remove any attached substances from the target object.

[0020] With this configuration, even if the appearance of the object differs from the registered appearance image due to attached substances, it becomes possible to obtain an object image that can accurately recognize the object by giving appropriate instructions to the user.

[0021] In a preferred embodiment of the present invention, the estimation means uses an estimation model trained with training data that takes the object image as input and outputs registered objects identical to the target object included in the object image to estimate which of the registered objects the target object is.

[0022] By adopting this configuration, it becomes possible to improve estimation accuracy by appropriately selecting the dataset.

[0023] The present invention is an object recognition program for estimating which of the registered objects registered in a database is the target object included in the object image captured by a user, and includes an acquisition means for acquiring the object image, and for a plurality of appearance images representing the appearance of one object viewed from different characteristic angles, registering the appearance image or its feature amount in the database as appearance information for each registered object, and an estimation means for estimating which of the registered objects the target object is based on the object image and the appearance information, and causing a computer to function as such.

[0024] The present invention is an object recognition method for estimating which of the registered objects registered in a database is the target object included in the object image captured by a user, and includes an acquisition step of acquiring the object image, a registration step of registering the appearance image or its feature amount in the database as appearance information for each registered object for a plurality of appearance images representing the appearance of one object viewed from different characteristic angles, and an estimation step of estimating which of the registered objects the target object is based on the object image and the appearance information, and causing a computer to execute such.

Effect of the Invention

[0025] According to the present invention, a novel object recognition system can be provided that can more appropriately identify an object even when similar objects are assumed.

Brief Description of the Drawings

[0026] [Figure 1] It is a functional block diagram of an object recognition system according to an embodiment of the present invention. [Figure 2] It is a diagram for explaining an example of an object according to an embodiment of the present invention. [Figure 3] It is a diagram showing the shooting direction of an appearance image of an object recognition system according to an embodiment of the present invention. [Figure 4] It is a diagram showing an example of an appearance image of an object recognition system according to an embodiment of the present invention. [Figure 5]This figure shows an example of object information stored in a database in an object recognition system according to an embodiment of the present invention. [Figure 6] This is a flowchart illustrating the object recognition process of an object recognition system according to an embodiment of the present invention. [Figure 7] This is an example of a display of a shooting instruction for an object recognition system according to an embodiment of the present invention. [Modes for carrying out the invention]

[0027] The object recognition system of the present invention will be described below with reference to the drawings. Note that the embodiments shown below are examples of the present invention, and the present invention is not limited to these embodiments; various configurations can be adopted.

[0028] For example, in this embodiment, the configuration and operation of the object recognition system will be described, but similar configurations, devices, computer programs, etc., can achieve similar effects. Furthermore, the program may be stored on a recording medium. Using this recording medium, the program can be installed on a computer, for example. Here, the recording medium on which the program is stored may be a non-transient recording medium such as a CD-ROM.

[0029] In this invention, "object" refers to any object that exists as a physical object, and the subject is not limited. In this embodiment, industrial products, the parts that make up such products, subassemblies formed by combining multiple parts, etc., are assumed to be objects.

[0030] In this invention, multiple objects are registered in advance along with their appearance, and these registered objects are referred to as "registered objects." The system then recognizes which of the registered objects the object photographed by the user is based on its appearance.

[0031] In this invention, an image showing the appearance of a registered object when viewed from a specific angle is called an appearance image. This specific angle is hereafter referred to as the "feature angle," and in this embodiment, appearance information indicating the appearance image or the feature quantities obtained from the appearance image is stored in the database in association with the feature angle for each appearance image. By storing appearance information for multiple appearance images, each associated with a different feature angle, for each registered object, it becomes possible to recognize objects based on their appearance.

[0032] In this embodiment, we assume that the appearance image itself is the appearance information, and we simplify the explanation by saying "register the appearance image" when referring to the registration of appearance information. However, it is also possible to use feature quantities of the appearance image instead of the appearance image itself. Here, as feature quantities, vector values ​​obtained by calculation based on the pixel values ​​of the appearance image can be used. Preferably, the feature quantities used are information that combines local and global features of the image. Furthermore, the feature quantities to be used may be determined by machine learning. For example, it is conceivable that a neural network model could be trained to perform the task of identifying the type of object from a captured image of an object, allowing the model to extract feature quantities and determine the feature quantities to be used in order to improve the identification accuracy.

[0033] Furthermore, it is possible to register multiple appearance images for the same registered object at the same characteristic angle. For example, in this embodiment, multiple appearance images are registered that are associated with the same characteristic angle and linked to the object information of the same registered object, including multiple appearance images with deposits such as rust, dirt, and scratches, and appearance images without such deposits. In addition, in the present invention, appearance information does not necessarily need to be registered in association with the characteristic angle; it may also be registered in a form that links object information and appearance images.

[0034] Next, the functional configuration of the object recognition system of this embodiment will be described with reference to Figure 1. In this embodiment, the object recognition system is configured such that an object recognition device 1 and a terminal device 2 can communicate via a network NW, and the object recognition device 1 is connected to a database DB by wired or wireless connection. The object recognition device 1 receives an object image taken by the user from the terminal device 2 and performs recognition processing to recognize which of the registered objects registered in the database DB the target object included in the object image is.

[0035] The object recognition device 1 can be a general computer device such as a server, which is equipped with various input / output devices including a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit), a main memory such as RAM (Random Access Memory), an auxiliary storage device such as an HDD (Hard Disk Drive), an SSD (Solid State Drive), or flash memory, and means for connecting to a network NW. Alternatively, the object recognition device 1 may be configured using multiple computers.

[0036] As terminal device 2, any computer device such as a smartphone or tablet terminal can be used, equipped with various input / output devices including a camera, processing unit, storage device, and means for connecting to a network NW. A dedicated application for transmitting object images to object recognition device 1 and a browser application for accessing a dedicated web page are stored in the storage device, and the processing unit executes various processes, so that any computer device functions as terminal device 2 of the present invention.

[0037] In this embodiment, a computer equipped with a camera is used as the terminal device 2 to transmit object images to the object recognition device 1 via the network NW. However, it is also possible to configure the device to transmit object images directly to the object recognition device 1 via wired or wireless connection from, for example, a digital camera. Alternatively, the object recognition device 1 may acquire object images that have been previously stored in any database or the like.

[0038] The object recognition device 1 includes a registration means 11 for registering an external image for each registered object, an acquisition means 12 for acquiring an object image transmitted from a terminal device 2, and an estimation means 13 for estimating which registered object the target object included in the object image belongs to. These means may be distributed across multiple computers.

[0039] The registration means 11 registers object information for each registered object and registers multiple appearance images linked to the object information. Each appearance image shows the appearance of the object when viewed from a specific characteristic angle and is registered in association with the characteristic angle. Below, an example of appearance image registration will be explained using Figures 2 to 4.

[0040] To illustrate an example of registering an external image, an example of an object is shown in Figure 2. As mentioned above, the type of object is not limited, but here we will assume a differential pressure transmitter as shown in Figure 2. This differential pressure transmitter has a constriction between the upstream and downstream pipes and measures the difference between the upstream and downstream pressures to measure the flow rate of the fluid passing through the pipe. As shown in Figure 2, the differential pressure transmitter can be divided into multiple parts, and in this example, we will assume that the conduit, which is one of the components that make up the differential pressure transmitter, will be registered as the registered object.

[0041] In this example, the conduit has a low-pressure side and a high-pressure side, which are distinguishable, but their shapes are very similar, making it difficult to tell them apart. In cases like this, where similar parts or objects are assumed, the external images are registered in association with multiple feature angles so that each object can be distinguished. It is preferable to register multiple external images at multiple feature angles, but association with feature angles is not mandatory; it is also acceptable to simply register multiple external images at different feature angles.

[0042] External images can be obtained, for example, by taking photographs from above, horizontally, and below the object, shifting the observation angle (feature angle) around the vertical axis, as shown in Figure 3. Here, in obtaining external images, the object or camera may be actually moved to photograph the real object for each feature angle, or computer graphics (CG) technology may be used to obtain information showing the 3D shape and generate external images as seen from each feature angle based on that information.

[0043] Figure 4 shows an example of the appearance image obtained for each feature angle in this manner. In this embodiment, the feature angle is associated with the horizontal and vertical angles from a reference position, and images showing the respective appearances are registered. Here, the vertical angle is set to three levels: upward, horizontal, and downward. However, this can also be set to, for example, horizontal as 0 degrees and changed within the range of -90 degrees to 90 degrees at predetermined intervals (e.g., 30 degrees). Furthermore, the feature angles do not necessarily have to be at fixed intervals.

[0044] Furthermore, the registration means 11 of this embodiment registers a feature image, which includes information to distinguish the registered object from other objects, separately from the external image, or a part of the external image, and links it to the object information. For example, in the example of the conduit in Figure 3, images of the parts where different characters are engraved on the low-pressure side and the high-pressure side can be registered as feature images. In this embodiment, the registration means 11 also registers the feature image in association with its observation angle, similar to the feature angle of the external image. Here, the observation angle of the feature image is an angle that makes it easy to distinguish the registered object from other similar objects.

[0045] The acquisition means 12 acquires the object image captured by the terminal device 2. The method of acquiring the object image is not limited; for example, it may be in the form of uploading a previously captured image as the object image. However, in this embodiment, it is assumed that the camera provided by the terminal device 2 is pointed at the target object, and an object image is captured as it occurs, and transmitted to the object recognition device 1 each time or at predetermined intervals, and acquired by the acquisition means 12. In this embodiment, the acquisition means 12 also acquires information regarding the shooting angle along with the object image.

[0046] The estimation means 13 estimates which registered object the target object included in the object image belongs to, based on the object image acquired by the acquisition means 12 and the appearance images of registered objects registered in the database DB. In this embodiment, the estimation means 13 estimates which registered object the target object belongs to by using an estimation model that has been trained with training data that takes an object image as input and outputs registered objects identical to the target object included in the object image. That is, the estimation means 13 estimates which registered object the target object belongs to by inputting the object image into an estimation model that has been trained to take an appearance image as input and estimate the registered object that corresponds to it.

[0047] The database (DB) stores various information, such as object information for each registered object and appearance images associated with that object information. Figure 5 shows an example of object information for each registered object registered in the database (DB). The object information includes an object ID that uniquely identifies the registered object, a model number, an object name, a manufacturer name, a feature image ID, and a shooting instruction. The feature image ID is information that specifies a feature image that shows the appearance of a part that is particularly important for distinguishing the registered object from other similar registered objects. In this embodiment, a feature image is registered for each registered object in this way.

[0048] A feature image is an image that shows the appearance of an object from a shooting direction that is useful for distinguishing it from other objects, such as its characteristic shape, markings, or printing. For example, an image of a characteristic part of an object taken in close-up may be used as a feature image. Any of the appearance images associated with the feature image can be used as the feature image, but an image of a specific part rather than the entire object may also be used separately from the appearance image. In other words, the feature image may or may not be associated with information indicating the observation angle of the object. The shooting instruction is text that prompts the user to take a picture of the part indicated by the feature image, and is registered as a string of characters. Note that the shooting instruction may also be in other forms, such as an illustration (image) of an arrow indicating the shooting direction.

[0049] In this way, for each registered object, appearance images and feature images for each characteristic angle are registered in the database DB. The present invention then infers which of the registered objects the target object included in the object image is, based on the registered information and the object image taken by the user. The procedure for object recognition processing will be explained below using a flowchart.

[0050] Figure 6 is a flowchart of the recognition process in this embodiment. In this embodiment, the imaging means 21 in the terminal device 2 captures an image of the target object and generates an object image. In the object recognition device 1, first in step S1, the acquisition means 12 acquires the object image transmitted from the terminal device 2. In this embodiment, along with the object image, information regarding the shooting angle obtained by the acceleration sensor, gyro sensor, etc., provided by the terminal device 2 is also acquired.

[0051] In capturing object images, it is possible to take each image individually after receiving a capture instruction, but it is preferable that the user points the camera at the object, and the imaging means 21 continuously acquires object images and transmits them to the object recognition device 1. As a result, the acquisition means 12 continuously acquires multiple object images, and estimation processing is performed based on these images as described later, which is expected to improve recognition accuracy. Furthermore, the shooting angle does not need to be acquired as an absolute value each time an object image is acquired; it may be acquired as a relative change, indicating how much and in which direction the object has moved, based on the shooting angle of a given object image.

[0052] Next, in step S2, the estimation means 13 identifies candidate target objects based on the object image and information registered in the database DB. Specifically, for example, a method is envisioned in which candidate target objects are identified by inputting information on the object image and the shooting angle into a neural network model that has been trained to identify objects based on appearance images and their characteristic angles registered in the database DB.

[0053] In step S3, the estimation means 13 determines whether it has succeeded in identifying a candidate. For example, in step S2, the probability that the target object is one of the registered objects is calculated for each registered object, and the success or failure of candidate identification can be determined by whether or not there is a registered object for which a probability of a predetermined value or higher has been calculated. In this case, if there is no registered object that satisfies a value of a predetermined value or higher, it is determined that candidate identification has failed, and the process proceeds to step S4.

[0054] If a candidate cannot be identified, it is possible that there is a problem with the object image. For example, the appearance of the object, such as its shape or color, may have changed due to dirt, rust, or other deposits. Therefore, in this embodiment, in step S4, the estimation means 13 generates shooting instruction information to instruct the user to remove the deposits from the target object and transmits it to the terminal device 2. The terminal device 2 then outputs based on the shooting instruction information, thereby instructing the user to remove the deposits. In addition, if a large movement is detected when capturing the object image, it is possible that the object image may be damaged due to blurring, so other shooting instruction information may be generated, such as generating shooting instruction information that instructs the user to "do not move" for a certain period of time.

[0055] If a candidate is successfully identified in step S3, the estimation means 13 proceeds to step S5 to check whether there are multiple candidates. If there are multiple candidates, the process proceeds to step S6, where the estimation means 13 generates shooting instruction information to instruct the user to photograph the target object based on the feature image of one of the registered objects among the identified candidates. The estimation means 13 transmits the generated shooting instruction information to the terminal device 2, and the output means 23 outputs based on the shooting instruction information.

[0056] Figure 7 shows an example of the display of the instruction screen based on the shooting instruction information. In this embodiment, shooting instruction information is generated based on the feature image and shooting instruction of any candidate object information identified by the estimation means 13. Here, the feature image is displayed, and the words "Please photograph the engraving as shown in the image." are also displayed as the shooting instruction. Alternatively, for example, the feature image may be superimposed on the preview of the captured image, and instructions such as "Align the object with the guide" may be displayed. The choice of which registered object's feature image and shooting instruction to use from the candidates can be arbitrarily determined. For example, one possible method is to use the object information of the candidate with the higher probability calculated in step S3.

[0057] Furthermore, while this example shows the display of feature images and shooting instructions as they are, it is also possible to generate shooting instruction information that specifically specifies the shooting angle, etc. For example, it is conceivable to register the observation angle in association with the feature image in a database DB and generate shooting instruction information based on that.

[0058] More specifically, the acquisition means 12 first acquires the object image along with its shooting angle. Then, the estimation means 13 compares the shooting angle with the observation angle associated with the feature image and generates shooting instruction information that specifically indicates the amount and direction of movement, i.e., how much and in which direction to move from the current shooting angle. In this case, the shooting instruction output by the output means 23 can display, for example, text such as "Move 30 degrees to the right and take the picture," or an arrow indicating the direction in which to move the camera.

[0059] In this embodiment, shooting instruction information is generated based on the identified candidate feature images. However, it is also possible to generate shooting instruction information that simply instructs the user to change the shooting angle, such as "Please take a picture from a different direction" or "Please rotate the object," without using feature images or their observation angles.

[0060] As described above, in this embodiment, object images are continuously acquired, so recognition is performed again when the user moves the camera according to the instructions. Here, since the candidates have already been narrowed down in step S3, the estimation means 13 identifies a more appropriate registered object from among the candidates based on the object image and feature image acquired according to the instructions in step S6, and outputs the identified registered object as a candidate to the terminal device 2 in step S7, and receives approval from the user regarding the recognition result. Note that here as well, the same processing as in steps S2 to S5 may be performed on all registered objects without narrowing down the options.

[0061] If the user approves the displayed candidate (Yes in step S8), the process proceeds to step S9 to identify the object and terminate. On the other hand, if the displayed candidate is considered to be different from the photographed object, the user chooses not to approve it (No in step S8), and the process returns to step S1 to repeat the recognition process.

[0062] As described above, the object recognition system of this embodiment allows the user to take a photograph of the target object using the terminal device 2, enabling suitable object recognition even when similar objects exist. This makes it possible, for example, to display the sales page for the identified registered object or to make inquiries based on the registered object, and even if the user does not know the model number of the target object, they can easily perform procedures such as ordering the object, repairing it, or making inquiries to the manufacturer.

[0063] Specifically, the present invention is envisioned to be applied to a product's online shopping page, where, after identifying an object, a purchase button is displayed to add the identified registered object to the shopping cart. This allows users to easily purchase the correct product even if they do not know the model number of the object they have. Alternatively, for example, the present invention could be applied to a product manufacturer's website, where, after identifying an object, a manual display button is displayed to show links to a set of manuals for the identified registered object. These manuals could include repair procedures, instruction manuals, installation instructions, specifications, etc.

[0064] In the above application examples, if an object is a part of a product and cannot be purchased separately, or if there is no manual for the part itself, the object information may be linked to the product, and a purchase button or manual button for the corresponding product may be displayed. [Explanation of symbols]

[0065] 1:Object recognition device 2: Terminal device 11: Registration method 12: Acquisition method 13: Estimation means 21: Imaging means 22: Input method 23: Output means DB: Database NW: Network

Claims

1. An object recognition system that estimates which of registered objects registered in a database a target object included in an object image captured by a user is, An acquisition means for acquiring the object image; a registration means for rotating a three-dimensional model of an object to generate a plurality of appearance images observed from each characteristic angle, and registering the plurality of appearance images or their respective characteristic amounts in the database as appearance information of the registered object; and an estimation means for estimating which of the registered objects the target object is based on the object image and the appearance information.

2. 2. The object recognition system according to claim 1, wherein said registration means comprises means for acquiring an image of said registered object, and registering said appearance information in said database as said appearance image for each of said registered objects.

3. 3. The object recognition system according to claim 1, wherein the registration means generates or acquires the appearance image of the registered object having an attachment on its appearance, and registers the appearance information based on the generated appearance image in the database.

4. 4. The object recognition system according to claim 1, wherein the registration means registers the appearance information in association with information indicating the characteristic angle.

5. the registration means is configured to be able to register a feature image including information for distinguishing the registered object from other objects in association with the registered object; The object recognition system according to any one of claims 1 to 4, wherein the estimation means identifies candidates of the registered object that may be the target object, and generates shooting instruction information for instructing a user to photograph the target object based on the feature image of the registered object.

6. The object recognition system according to any one of claims 1 to 5, wherein the estimation means generates shooting instruction information for instructing a user to remove any attachments from the target object when it is unable to identify a candidate for the registered object that may be the target object.

7. 7. The object recognition system according to claim 1, wherein the estimation means estimates which of the registered objects the target object is, using an estimation model trained with teacher data that receives the object image as input and outputs a registered object that is identical to the target object included in the object image.

8. An object recognition program for estimating which of registered objects registered in a database a target object included in an object image captured by a user is, An acquisition means for acquiring the object image; a registration means for rotating a three-dimensional model of an object to generate a plurality of appearance images observed from each characteristic angle, and registering the plurality of appearance images or their respective characteristic amounts in the database as appearance information of the registered object; an estimation unit that estimates which of the registered objects the target object is based on the object image and the appearance information.

9. An object recognition method for estimating which of registered objects registered in a database a target object included in an object image captured by a user is, comprising: acquiring the object image; a registration step of rotating a three-dimensional model of the object to generate a plurality of appearance images observed from each characteristic angle, and registering the plurality of appearance images or their respective characteristic amounts in the database as appearance information of the registered object; an estimation step of estimating which of the registered objects the target object is based on the object image and the appearance information.