State determination device, state determination system, state determination method and program

The state determination device divides captured images into grids, estimates section states using a learning model, and determines overall object states, addressing the need for extensive learning data, enabling efficient and real-time object state assessment.

JP2024158680AActive Publication Date: 2024-11-08STEERETAIL CO LTD
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
JP2023074072
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-04-28
Publication Date
2024-11-08
Estimated Expiration
2043-04-28

AI Technical Summary

Technical Problem

Existing image recognition systems for determining the state of objects require a large amount of labeled learning data to create an image recognition model, which is cumbersome and resource-intensive.

Method used

A state determination device that acquires a captured image of an object, divides it into a grid, labels each section, estimates the state of each section using a learning model, and determines the overall state based on section states, reducing the need for extensive learning data.

Benefits of technology

Enables state determination without requiring a large amount of learning data, allowing for real-time object state assessment and efficient model generation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024158680000001_ABST
    Figure 2024158680000001_ABST
Patent Text Reader

Abstract

To provide a state determination device that can determine a state of an object without preparing a large amount of learning data.SOLUTION: A state determination device according to the present disclosure comprises: image acquisition means for acquiring a photographed image of an entire object whose state is to be determined; division means for dividing the image into pieces in a grid manner; estimation means for estimating a state of each section based on a learning model generated using, as learning data, divided images of each section to which a label is given, indicating at a prescribed stage the state of each section divided in the grid manner; determination means for determining a state of the entire object based on the state of each section; and output means for outputting state information related to the state of the entire object.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present disclosure relates to a state determination device, a state determination system, a state determination method, a computer program, and the like. [Background technology]

[0002] In buffet-style restaurants, there is technology that determines how much food is left by recognizing images of food captured by a camera.

[0003] For example, Patent Document 1 discloses a food remaining amount detection device that uses a camera to capture an image of food and calculates the amount of food remaining. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2015-138452 A Summary of the Invention [Problem to be solved by the invention]

[0005] By the way, there are many methods that use image recognition by AI (Artificial Intelligence) to determine the state of an object. However, this method requires images with labels indicating each state of the object as learning data when creating an image recognition model, and it is necessary to prepare a large number of images to be used for learning.

[0006] An example of an objective of the present disclosure is to provide a state determination device that can determine the state of an object without preparing a large amount of learning data. [Means for solving the problem]

[0007] A condition determination device according to one aspect of the present disclosure includes an image acquisition means for acquiring an image of the entire object whose condition is to be determined, a division means for dividing the captured image into a grid, an estimation means for estimating the condition of each partition based on a learning model generated using as learning data the divided images of each partition, to which a label indicating the condition of each partition divided into the grid at a predetermined stage has been assigned, a determination means for determining the condition of the entire object based on the condition of each partition, and an output means for outputting condition information relating to the condition of the entire object.

[0008] A condition determination system according to one aspect of the present disclosure includes a camera and the above-described condition determination device, and an image acquisition means acquires an image of the entire target captured by the camera.

[0009] In one aspect of the present disclosure, a condition determination method includes a computer acquiring an image of an entire object whose condition is to be determined, dividing the acquired image into a grid, and assigning labels to each of the grid-shaped sections indicating their condition at a predetermined stage. The divided images of each section are used as learning data to generate a learning model, which estimates the condition of each section, determines the condition of the entire object based on the condition of each section, and outputs condition information regarding the condition of the entire object.

[0010] A program in one aspect of the present disclosure causes a computer to obtain an image of the entire object whose status is to be determined, divide the image into a grid, and assign labels indicating the status of each grid-like section at a predetermined stage. The divided images of each section are used as learning data to generate a learning model, which estimates the status of each section, determine the status of the entire object based on the status of each section, and output status information regarding the status of the entire object. Effect of the Invention

[0011] One example of the effect of the present invention is to provide a state determination device that can determine the state of an object without preparing a large amount of learning data. [Brief description of the drawings]

[0012] [Figure 1]FIG. 1 is a block diagram showing a configuration including a state determination device according to the present disclosure. [Diagram 2] FIG. 2 is a diagram showing a hardware configuration in which the state determination device according to the present disclosure is realized by a computer device and its peripheral devices. [Diagram 3] FIG. 3 is a diagram illustrating a situation in which the state determination device of the present disclosure is applied. [Figure 4] FIG. 4 is a diagram for explaining an example of determining the overall state of a dish in the present disclosure. [Diagram 5] FIG. 5 is a flowchart showing the state determination operation in the present disclosure. [Figure 6] FIG. 6 is a block diagram showing a configuration including a state determination device in a modified example of the present disclosure. [Figure 7] FIG. 7 is a flowchart showing the state determination operation in the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] [First embodiment] Fig. 1 is a block diagram including the configuration of a condition determination system 10 in a first embodiment. Referring to Fig. 1, the condition determination system 10 includes a condition determination device 100, a camera 200, and an employee terminal 300. The condition determination device 100 is connected to the camera 200 and the employee terminal 300 via a network.

[0014] The status determination device 100 determines the status of an object from a captured image of the entire object, and notifies an employee terminal 300. The camera 200 captures an image of the entire object whose status is to be determined. The employee terminal 300 is a terminal carried by a store employee or the like, and receives information regarding the status of the object. As shown in FIG. 1, the status determination device 100 includes an image acquisition unit 101, a division unit 102, an estimation unit 103, a determination unit 104, and an output unit 105. In this embodiment, a situation is assumed in which a customer determines the amount of food remaining in a self-service buffet-style restaurant, but the scope of application of the present disclosure is not limited thereto.

[0015] Fig. 2 is a diagram showing an example of a hardware configuration in which the state determination device 100 in the first embodiment of the present disclosure is realized by a computer device 500 including a processor. As shown in Fig. 2, the state determination device 100 includes a central processing unit (CPU) 501, memories such as a read only memory (ROM) 502 and a random access memory (RAM) 503, a storage device 505 such as a hard disk for storing a program 504, a communication interface 508 for network connection, and an input / output interface 509 for inputting and outputting data. In this embodiment, a learning model described later is stored in the storage device 505.

[0016] In the first embodiment, the components of the state determination device 100 are connected to each other, for example, via a bus 510. The state determination device 100 in the first embodiment shown in FIG. 1 can also be configured using cloud computing or the like.

[0017] CPU 501 operates an operating system to control the entire state determining device 100 according to the first embodiment of the present invention. CPU 501 also reads programs and data from a recording medium 506 mounted in a drive device 507 or the like into a memory. CPU 501 also functions as image acquisition unit 101, division unit 102, estimation unit 103, determination unit 104, and output unit 105 in the first embodiment, or as part of these, and executes processes or commands in the flowchart shown in Fig. 5, which will be described later, based on the programs.

[0018] The recording medium 506 is, for example, an optical disk, a flexible disk, a magneto-optical disk, an external hard disk, or a semiconductor memory. The semiconductor memory or the like that is a part of the recording medium is a non-volatile storage device, and stores the program therein. The program may also be downloaded from an external computer (not shown) that is connected to a communication network.

[0019] As described above, the first embodiment shown in FIG. 1 is realized by computer hardware shown in FIG. 2. However, the means for realizing each unit of the status determination device 100 in FIG. 1 is not limited to the configuration described above. The status determination device 100 may be realized by one physically combined device, or may be realized by two or more physically separated devices connected by wire or wirelessly. The status determination device 100 may further include a display device such as a display that displays information on a plurality of ordered items. The status determination device 100 and the display device may be realized as a system by separate devices.

[0020] Fig. 3 is a diagram illustrating a scene in which the status determination device 100 of the present disclosure is applied. In the example of Fig. 3, the camera 200 captures an image of the entire dish from above. The status determination device 100 receives the image of the dish captured by the camera 200, determines the amount of food remaining, and transmits the determination result to the employee terminal 300.

[0021] Next, a detailed description will be given of each component of the status determination device 100. In this embodiment, an example of determining the remaining amount of food will be mainly described, but the application of the present disclosure is not limited to this as long as the status of the object can be determined based on a captured image.

[0022] The image acquisition unit 101 acquires a captured image of the entire object whose state is to be determined. In this embodiment, the state is, for example, the amount of food remaining. The captured image is, for example, an image captured at an angle that allows the remaining amount of food in total to be grasped. The image acquisition unit 101 acquires captured images captured by the camera 200 at predetermined time intervals. When the image acquisition unit 101 acquires a captured image of the object whose state is to be determined, the image acquisition unit 101 outputs the captured image to the division unit 102.

[0023] The division unit 102 divides the captured image into a grid pattern. For example, the division unit 102 divides the captured image input from the image acquisition unit 101 equally in the horizontal and vertical directions. The division unit 102 outputs the divided images to the estimation unit 103.

[0024] The estimation unit 103 estimates the state of each partition based on a learning model generated using a divided image of each partition, to which a label indicating the state of each partition divided into a grid at a predetermined stage is added, as learning data. The estimation unit 103 inputs the divided image of each partition into the learning model and estimates the state of each partition. Each divided image of a partition for learning is associated with any of the labels indicating the state of each partition at a predetermined stage that are prepared in advance. For example, when the object of state determination is the remaining amount of food, the divided image of each partition for learning is assigned any of the remaining amount labels indicating the degree of the remaining amount of food, such as a state in which the divided image is filled with food ("FULL"), a state in which about half of the food can be confirmed in the divided image ("HALF"), and a state in which no food is present in the divided image ("EMPTY"). In the present disclosure, by indicating the degree of the state of each partition at fewer stages and reducing the number of types of labels added to the divided images for learning, the amount of learning data to be prepared can be further reduced. The estimation unit 103 outputs state information of each partition to the determination unit 104.

[0025] The determination unit 104 determines the overall state of the object based on the state information of each section. The determination unit 104 determines the overall state of the object by combining the state information of each section. For example, if the state information of each section is information expressed by a numerical value, the determination unit 104 determines the overall state of the object by calculating the sum or average value of the state information of each section.

[0026] The output unit 105 outputs status information regarding the status of the entire object. The output unit 105 notifies the employee terminal 300 of the status information, for example. The output unit 105 may output the status information when the status satisfies a predetermined condition. When determining the remaining amount of food, the output unit 105 may output the remaining amount of the entire food when the remaining amount of the entire food is less than a predetermined amount. In this case, it is possible to inform the employee that food needs to be replenished.

[0027] FIG. 4 is a diagram for explaining an example of determining the state of the entire dish in the present disclosure. The example of FIG. 4 shows an example of determining the remaining amount of pizza. As shown in FIG. 4, the division unit 102 divides an image of an entire pizza A into 16 sections in a grid pattern. In FIG. 4, the remaining amount label indicating the degree of remaining amount of pizza is given in three stages: "EMPTY" as in A1, when there is no pizza in one section; "HALF" as in A2, when about half of the pizza can be seen in one section; and "FULL" as in A3, when the section is filled with pizza. The estimation unit 103 inputs the divided images of each section into a learning model to estimate the remaining amount of pizza in each section.

[0028] The determination unit 104 quantifies the remaining amount of each section by setting the estimated remaining amount of each section to 1 when the section is "FULL", 0.5 when the section is "HALF", and 0 when the section is "EMPTY", and calculates a remaining amount score from the total remaining amount of all sections. The remaining amount score is a value indicating the percentage of the remaining amount of food when the remaining amount of food when the food is replenished is 100%. In the example of FIG. 4, since there are 4 sections of "EMPTY", 3 sections of "HALF", and 9 sections of "FULL", the determination unit 104 calculates the remaining amount of all sections as 10.5 (0×4+0.5×3+1×9). Note that when the remaining amount of the pizza shown in FIG. 4 is 100%, since there are 4 sections of "HALF" and 12 sections of "FULL", the remaining amount can be quantified as 14 (0.5×4+1×12). Therefore, the determination unit 104 determines the remaining amount score of the pizza shown in FIG. 4 to be 75% (10.5 / 14×100).

[0029] The operation of the state determining device 100 configured as above will be described with reference to the flowchart of FIG.

[0030] 5 is a flowchart showing an outline of the operation of the state determining device 100 according to the present disclosure. Note that the processing according to this flowchart may be executed based on program control by the above-mentioned processor.

[0031] As shown in FIG. 5, first, the image acquisition unit 101 acquires a captured image of the entire target whose state is to be determined (step S101). Next, the division unit 102 divides the captured image into a grid (step S102). Next, the estimation unit 103 estimates the state of each section based on a learning model generated using the divided images of each section, to which a label indicating the state of each section divided into a grid at a predetermined stage is added, as learning data (step S103). If the estimation unit 103 has not completed the estimation of the state of all sections (S104; NO), it repeats S103. Next, if the estimation of the state of all sections is completed (S104; YES), the determination unit 104 acquires the state of the entire target based on the state of each section (step S105). Finally, the output unit 105 outputs state information regarding the state of the entire target (step S106). The state determination device 100 repeats the process of this flowchart every time it receives a new captured image from the camera 200, for example. With this, the state determining device 100 ends the operation of the state determining process.

[0032] In the state determination device 100 in the present embodiment described above, the division unit 102 divides the captured image into a grid, and the estimation unit 103 estimates the state of each section based on a learning model generated using the divided images of each section, to which a label indicating the state of each section divided into the grid at a predetermined stage is attached, as learning data. Then, the determination unit 104 determines the state of the entire object based on the state of each section. In a method of determining the state of the entire object using the captured image of the entire object, labeling is required for each state, and there is a problem that a large amount of images to be used for learning must be prepared. On the other hand, in the present embodiment, a learning model is generated using the divided images of each section, to which a label indicating the state of each section divided into the grid at a predetermined stage is attached, as learning data. This makes it possible to determine the state of the object without preparing a large amount of learning data.

[0033] In addition, in this embodiment, the captured image acquired by the camera 200 is divided into a grid and input directly to the learning model, so that the state of the entire target can be output in real time. This shortens the time required to generate the learning model, and enables employees to check the state of the entire target in real time.

[0034] [Variations] Next, a modified example of this embodiment will be described in detail with reference to the drawings. Below, the description of the present embodiment will be omitted to the extent that the description of this embodiment is not unclear. Each component in each embodiment of the present disclosure can be realized not only by hardware but also by a computer device or software based on program control, as in the computer device shown in FIG. 2.

[0035] Fig. 6 is a diagram showing a configuration including a state determination device 110 according to a modified example of the first embodiment of the present disclosure. With reference to Fig. 6, the state determination device 110 according to the modified example of the first embodiment will be described, focusing on the parts that differ from the state determination device 100 according to the first embodiment.

[0036] In this modified example, the target is food, and it is assumed that in a buffet-style restaurant, the restocking of food is detected when the remaining amount of food is greater than the most recently determined remaining amount. However, this modified example is not limited to food, and can also be applied to situations where it is necessary to detect the restocking of a specific product.

[0037] Status determination device 110 according to the modified example of the first embodiment includes an image acquisition unit 111, a division unit 112, an estimation unit 113, a determination unit 114, an output unit 115, a storage unit 116, and a detection unit 117. Note that the configurations of image acquisition unit 111, division unit 112, estimation unit 113, determination unit 114, and output unit 115 are similar to the corresponding components in the first embodiment, except that the target of status determination is limited to determining the amount of remaining food.

[0038] The storage unit 116 stores the remaining amount of the entire dish. The storage unit 116 stores the remaining amount of the entire dish determined by the determination unit 114 in the storage device 505, for example.

[0039] The detection unit 117 detects that food has been replenished when the remaining amount of the entire most recent object is greater than the remaining amount of the entire object before a predetermined time. When the determination unit 114 determines the remaining amount of the entire food, the detection unit 117 refers to the remaining amount of the entire most recent food stored in the storage device 505, and detects that food has been replenished when the determined remaining amount is greater than the remaining amount of the most recent food. Furthermore, when the detection unit 117 determines that food has been replenished, it outputs the determination information to the storage unit 116, and the storage unit 116 may update the remaining amount score of the food to 100% and store it in the storage device 505. On the other hand, when the detection unit 117 does not determine that food has been replenished, it outputs the determination information to the storage unit 116, and the storage unit 116 stores the remaining amount of the food in the storage device 505 with the remaining amount score determined by the determination unit 114.

[0040] When it is determined that a dish has been replenished, the output unit 115 may further notify the employee terminal 300 of the replenishment. For example, the output unit 115 may notify the employee terminal 300 of the person in charge of replenishing the dish of the name of the replenished dish. In this modification, the output unit 115 may also notify the employee terminal 300 of the remaining amount of a specific dish stored in the storage device 505, based on an operation from the employee terminal 300.

[0041] The operation of the state determining device 110 configured as above will be described with reference to the flowchart of FIG.

[0042] 7 is a flowchart showing an outline of the operation of the state determining device 110 according to the present disclosure. Note that the processing according to this flowchart may be executed based on program control by the above-mentioned processor.

[0043] As shown in FIG. 7, first, the image acquisition unit 111 acquires a captured image of the entire dish (step S111). Next, the division unit 112 divides the captured image into a grid (step S112). Next, the estimation unit 113 estimates the remaining amount of each section based on a learning model generated using a divided image of each section, to which a label indicating the remaining amount of food in each section divided into a grid at a predetermined stage is added, as learning data (step S113). Next, the determination unit 114 determines the remaining amount of the entire dish based on the remaining amount of each section (step S114). Next, the output unit 115 outputs the remaining amount of the entire dish (step S115). Next, the storage unit 116 stores the remaining amount of the entire dish (step S116). If the most recent remaining amount of the entire dish is stored in the storage device 505 (S117; YES), the flow proceeds to S118. Next, if the determined remaining amount of the entire dish is greater than the most recent remaining amount (S118; YES), the detection unit 117 determines that the dish has been replenished (step S119), and the output unit 115 notifies the employee terminal 300 that the dish has been replenished (step S120). On the other hand, if the most recent remaining amount of the entire dish is not stored in the storage device 505 in S117 (S117; NO) and if the remaining amount of the entire dish determined in S118 is equal to or less than the most recent remaining amount (S118; NO), this flow ends. With this, the status determination device 110 ends the operation of the status determination processing.

[0044] In the status determination device 110 in the variation of the first embodiment described above, when the determined remaining amount of the entire dish is greater than the most recent remaining amount, the detection unit 117 detects that the dish has been replenished, and the output unit 115 notifies the employee terminal 300 that the dish has been replenished. This makes it possible to prevent, for example, other employees from replenishing the same dish.

[0045] Although the present invention has been described above with reference to the embodiments, the present invention is not limited to the above-mentioned embodiments. Various modifications that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

[0046] For example, although a plurality of operations are described in order in the form of a flow chart, the order of description does not limit the order in which the plurality of operations are performed. Therefore, when implementing each embodiment, the order of the plurality of operations can be changed to the extent that does not interfere with the content. In addition, although the embodiment and the modified example have been described with respect to a case in which the food is pizza, the state determination device of the present disclosure can also be applied to other foods. In this case, the learning data can be replaced with an image of the food for which the remaining amount is to be determined, and a learning model can be generated, making it possible to apply the state determination device to other foods.

[0047] The state determination device 100 of the present disclosure can also be applied to determining the congestion rate of a space. That is, the image acquisition unit 101 acquires a captured image of the entire space in which the congestion rate is to be determined, captured from above. The division unit 102 divides the captured image into a lattice shape. The estimation unit 103 estimates the congestion rate of each section based on a learning model generated using the divided images of each section, to which a label indicating the congestion rate of each section divided into a lattice shape is attached at a predetermined level, as learning data. The determination unit 104 determines the congestion rate of the entire space using the congestion rate of each section. The output unit 105 outputs the congestion rate of the entire space. In this case, the state determination device 100 generates a learning model using the divided images of each section, to which a label indicating the congestion rate of each section divided into a lattice shape is attached at a predetermined level, as learning data. This makes it possible to determine the congestion rate in the space without preparing a large amount of learning data.

[0048] The state determination device 100 of the present disclosure can also be applied to determining the stock rate of products displayed on a display shelf. That is, the image acquisition unit 101 acquires an image of the entire display shelf on which the products to be determined for stock rate are displayed, the division unit 102 divides the image into a grid, and the estimation unit 103 estimates the stock rate of each section based on a learning model generated using the divided images of each section, to which a label indicating the stock rate of each section divided into a grid is attached at a predetermined stage, as learning data. The determination unit 104 determines the stock rate of the entire display shelf using the stock rate of each section. The output unit 105 outputs the stock rate of the entire display shelf. In this case, the state determination device 100 generates a learning model using the divided images of each section, to which a label indicating the stock rate of each section divided into a grid is attached at a predetermined stage, as learning data. This makes it possible to determine the stock rate of the display shelf without preparing a large amount of learning data. [Explanation of symbols]

[0049] 10, 11 Status Judgment System 100, 110 Status determination device 101, 111 Image acquisition unit 102, 112 division part 103, 113 Estimation part 104, 114 Judgment section 105, 115 Output section 116 Holding part 117 Detection Unit 200, 210 Camera 300,310 employee terminals

Claims

1. An image acquisition means for acquiring an image of an entire target whose state is to be determined; A division means for dividing the captured image into a grid pattern; An estimation means for estimating a state of each partition based on a learning model generated using, as learning data, divided images of each partition, to which a label indicating a state of each partition divided into a lattice shape is added at a predetermined stage; A determination means for determining a state of the entire object based on a state of each section; An output means for outputting status information regarding the status of the entire object.

2. The object is a dish, The image acquisition means acquires an image of an entire dish, The dividing means divides the captured image into a grid pattern, The estimation means estimates the remaining amount of each partition based on a learning model generated by using a divided image of each partition as learning data, to which a remaining amount label indicating the remaining amount of each partition divided into a lattice shape is added at a predetermined level; The determining means determines the remaining amount of the entire dish based on the state of each section, The state determination device according to claim 1 , wherein the output means outputs the remaining amount of the entire dish.

3. The state determination device according to claim 2 , wherein the output means outputs the remaining amount of the entire dish when the remaining amount of the entire dish is less than a predetermined amount.

4. A holding means for holding the remaining amount of the entire dish; 3. The state determination device according to claim 2, further comprising a detection means for detecting that the food has been replenished when the most recent remaining amount of the entire food is greater than the remaining amount of the entire food a predetermined time before.

5. 3. The state determination device according to claim 2, wherein the remaining amount label is a label indicating any one of a state where no food is present in the divided image, a state where about half of the food can be seen in the divided image, and a state where the divided image is filled with food.

6. The image acquisition means acquires an image of an entire space for which a congestion rate is to be determined, The dividing means divides the captured image into a grid pattern, The estimation means estimates the congestion rate of each section based on a learning model generated using a divided image of each section as learning data, to which a label indicating the congestion rate of each section divided into a lattice shape is added in a predetermined stage, The determination means determines a congestion rate of the entire space using the congestion rate of each section, The state determination device according to claim 1 , wherein the output means outputs a congestion rate of the entire space.

7. The image acquisition means acquires an image of an entire display shelf of the product for which the stock rate is to be determined, The dividing means divides the captured image into a grid pattern, The estimation means estimates the stock rate of each partition based on a learning model generated using a divided image of each partition, to which a label indicating the stock rate of each partition divided into a lattice shape is added at a predetermined stage, as learning data; The determination means determines an inventory rate of the entire display shelf using the inventory rate of each section, The state determination device according to claim 1 , wherein the output means outputs an inventory rate of the entire display shelf.

8. A camera and a state determination device according to any one of claims 1 to 7, The image acquisition means acquires an image of the entire target captured by the camera.

9. The computer Acquire a captured image of the entire target whose state is to be determined; Dividing the captured image into a grid pattern; A state of each partition is estimated based on a learning model generated using divided images of each partition as learning data, to which labels indicating the state of each partition divided into a lattice shape are attached at a predetermined stage; determining an overall state of the object based on the state of each compartment; A status determination method, which outputs status information regarding the status of the entire object.

10. Acquire a captured image of the entire target whose state is to be determined; Dividing the captured image into a grid pattern; A state of each partition is estimated based on a learning model generated using divided images of each partition as learning data, to which labels indicating the state of each partition divided into a lattice shape are attached at a predetermined stage; determining an overall state of the object based on the state of each compartment; A program that causes a computer to execute the steps of: outputting status information regarding the status of the entire object.

Citation Information

Patent Citations

  • Device and program for cuisine residual quantity detection

    JP2015138452A

  • Congestion status estimation device, method, and program

    JP2022053126A

  • Image processing apparatus and image processing system and control program for the same

    JP2022053849A

  • Learning model generation device, learning model generation system, learning model generation method, and learning model generation program

    JP7396499B2

  • Image recognition device, image recognition program, and image recognition method

    WO2020145085A1