Search device, search method, and computer program
Patent Information
- Application Number
- JP2024530781
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Filing Date
- 2024-12-25
- Publication Date
- 2025-03-07
Abstract
Description
Searching device, searching method, and recording medium
[0001] The present disclosure relates to the technical fields of a search device, a search method, and a recording medium.
[0002] Patent Literature 1 describes a technology that provides suitable support for investigations so that even non-experienced investigators can solve criminal cases by including an estimation unit 32 that estimates the type of a second incident based on behavioral history information (representing time-series changes in the behavior of participants) and human relationship information (representing time-series changes in the human relationships of participants) related to a first incident, an estimation model that represents the relationship between the type of the first incident, and the behavioral history information and human relationship information related to a second incident. Patent Literature 2 describes a technology that acquires trajectory information representing the movement trajectory of a subject within time-series images, acquires situation information representing the situation around the subject within the time-series images, acquires a search query including the movement trajectory and the situation around the subject, and searches for images included in the time-series images that match the search query based on the trajectory information and the situation information, thereby more appropriately searching for a desired scene. Patent document 3 describes a technique for converting a non-RFID tagged evidence collection into RFID tagged evidence, which includes the steps of providing evidence containing a non-RFID tag, obtaining information from the non-RFID tag, inputting the information into a processor, writing the information to the RFID tag, and associating a database record with the RFID tag.
[0003] International Publication No. 2021 / 199101 Japanese Patent Application Laid-Open No. 2018-081630 Special Publication No. 2004-531796
[0004] The present disclosure aims to provide a search device, a search method, and a recording medium that aim to improve upon the techniques described in prior art documents.
[0005] One aspect of the search device comprises an acquisition means for acquiring first situation information indicating the situation in a first predetermined range including the space-time in which a first search target occurred, and an output means for outputting first location information indicating the location of the first search target located within the first predetermined range, wherein the output means inputs the first situation information acquired by the acquisition means into a search model and causes the search model to output the first location information, and the search model is generated by machine learning using training data including second situation information indicating the situation in a second predetermined range including the space-time in which a second search target occurred and second location information indicating the location of the second search target located within the second predetermined range.
[0006] One aspect of the search method is a search method that acquires first situation information indicating the situation in a first predetermined range including the space-time in which a first search target occurred, and outputs first location information indicating the location of the first search target that exists within the first predetermined range, wherein the output is performed by inputting the acquired first situation information into a search model and having the search model output the first location information, and the search model is generated by machine learning using training data that includes second situation information indicating the situation in a second predetermined range including the space-time in which a second search target occurred and second location information indicating the location of the second search target that exists within the second predetermined range.
[0007] One aspect of the recording medium is a search method in which a computer acquires first situation information indicating the situation in a first predetermined range including the space-time in which a first search target occurred, and outputs first location information indicating the location of the first search target located within the first predetermined range, wherein the output is performed by inputting the acquired first situation information into a search model and having the search model output the first location information, and the search model has recorded thereon a computer program for causing the computer to execute a search method generated by machine learning using training data including second situation information indicating the situation in a second predetermined range including the space-time in which a second search target occurred and second location information indicating the location of the second search target located within the second predetermined range.
[0008] FIG. 1 is a block diagram showing the configuration of a search device in a first embodiment. FIG. 2 is a block diagram showing the configuration of a search device in a second embodiment. FIG. 3 is a flowchart showing the flow of a model generation operation performed by the search device in the second embodiment. FIG. 4 is a flowchart showing the flow of a search assist operation performed by the search device in the second embodiment. FIG. 5 is a diagram showing examples of input and output by the search device in the second embodiment. FIG. 6 is a diagram showing examples of input and output by the search device in the third embodiment. FIG. 7 is a block diagram showing the configuration of a search device in a fourth embodiment. FIG. 8 is a flowchart showing the flow of a search assist operation performed by the search device in the fourth embodiment. FIG. 9 is a diagram showing an example of presentation presented by the search device in the fourth embodiment. FIG. 10 is a flowchart showing the flow of a search assist operation performed by the search device in the fifth embodiment. FIG. 11 is a diagram showing an example of presentation presented by the search device in the fifth embodiment. FIG. 12 is a flowchart showing the flow of a search assist operation performed by the search device in the sixth embodiment. FIG. 13 is a diagram showing an example of output output by the search device in the sixth embodiment.
[0009] Hereinafter, embodiments of a searching device, a searching method, and a recording medium will be described with reference to the drawings. [1: First Embodiment]
[0010] A first embodiment of a search device, a search method, and a recording medium will be described below. Hereinafter, the first embodiment of the search device, the search method, and the recording medium will be described using a search device 1 to which the first embodiment of the search device, the search method, and the recording medium is applied. [1-1: Configuration of search device 1]
[0011] The configuration of a search device 1 in the first embodiment will be described with reference to Fig. 1. Fig. 1 is a block diagram showing the configuration of the search device 1 in the first embodiment.
[0012] As shown in FIG. 1, the search device 1 includes an acquisition unit 11 and an output unit 12. The acquisition unit 11 acquires first situation information indicating the situation of a first predetermined range including the time-space where a first search target has occurred. The output unit 12 outputs first location information indicating the location of the first search target present within the first predetermined range. The output unit 12 inputs the first situation information acquired by the acquisition unit 11 to a search model and causes the search model to output the first location information. The search model is generated by machine learning using training data including second situation information indicating the situation of a second predetermined range including the time-space where a second search target has occurred and second location information indicating the location of the second search target present within the second predetermined range. [1-2: Technical Effects of Search Device 1]
[0013] The search device 1 in the first embodiment uses a search model generated by machine learning to output first location information indicating the location of the first search target that exists in a first predetermined range that includes the space-time in which the first search target occurred, thereby reducing the workload of searching for the first search target. [2: Second Embodiment]
[0014] A second embodiment of the search device, the search method, and the recording medium will be described below. Hereinafter, the second embodiment of the search device, the search method, and the recording medium will be described using a search device 2 to which the second embodiment of the search device, the search method, and the recording medium is applied. [2-1: Configuration of the search device 2]
[0015] The configuration of the search device 2 in the second embodiment will be described with reference to Fig. 2. Fig. 2 is a block diagram showing the configuration of the search device 2 in the second embodiment.
[0016] 2 , the search device 2 includes a calculation device 21 and a storage device 22. The search device 2 may further include a communication device 23, an input device 24, and an output device 25. However, the search device 2 does not necessarily include at least one of the communication device 23, the input device 24, and the output device 25. The calculation device 21, the storage device 22, the communication device 23, the input device 24, and the output device 25 may be connected via a data bus 26.
[0017] The arithmetic device 21 includes, for example, at least one of a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), and an FPGA (Field Programmable Gate Array). The arithmetic device 21 loads a computer program. For example, the arithmetic device 21 may load a computer program stored in the storage device 22. For example, the arithmetic device 21 may load a computer program stored in a computer-readable, non-transitory storage medium using a storage medium reading device (e.g., the input device 24 described below) not shown in the drawings that the search device 2 includes. The arithmetic device 21 may acquire (i.e., download or load) the computer program from a device (not shown) located outside the search device 2 via the communication device 23 (or another communication device). The arithmetic device 21 executes the loaded computer program. As a result, logical functional blocks for executing the operations to be performed by the search device 2 are realized within the calculation device 21. In other words, the calculation device 21 can function as a controller for realizing logical functional blocks for executing the operations (in other words, processing) to be performed by the search device 2.
[0018] 2 shows an example of logical functional blocks implemented within the computing device 21 to perform a search assist operation. As shown in FIG. 2, the computing device 21 implements an acquisition unit 211, which is a specific example of an "acquisition means" described in the appendix, which will be described later; an output unit 212, which is a specific example of an "output means" described in the appendix, which will be described later; and a model generation unit 213, which is a specific example of a "model generation means" described in the appendix, which will be described later. However, the model generation unit 213 does not necessarily have to be implemented within the computing device 21. The respective operations of the acquisition unit 211, the output unit 212, and the model generation unit 213 will be described later with reference to FIGS. 3 to 5.
[0019] The storage device 22 can store desired data. For example, the storage device 22 may temporarily store a computer program executed by the arithmetic device 21. The storage device 22 may temporarily store data that the arithmetic device 21 temporarily uses when the arithmetic device 21 is executing a computer program. The storage device 22 may store data that the search device 2 stores long-term. The storage device 22 may include at least one of a RAM (Random Access Memory), a ROM (Read Only Memory), a hard disk device, a magneto-optical disk device, an SSD (Solid State Drive), and a disk array device. In other words, the storage device 22 may include a non-transitory recording medium. The storage device 22 may store at least one of a parameter MP and training data D, which will be described later.
[0020] The communication device 23 is capable of communicating with devices external to the search device 2 via a communication network (not shown). The communication device 23 may be a communication interface based on standards such as Ethernet (registered trademark), Wi-Fi (registered trademark), Bluetooth (registered trademark), or USB (Universal Serial Bus).
[0021] The input device 24 is a device that accepts information input to the search device 2 from outside the search device 2. For example, the input device 24 may include an operation device (e.g., at least one of a keyboard, a mouse trackball, a touch panel, a pointing device such as a pen tablet, a button, etc.) that can be operated by an operator of the search device 2. For example, the input device 24 may include a reading device that can read information recorded as data on a recording medium that can be externally attached to the search device 2.
[0022] The output device 25 is a device that outputs information to the outside of the search device 2. For example, the output device 25 may output information as an image. That is, the output device 25 may include a display device (a so-called display) that can display an image showing the information to be output. Examples of display devices include a liquid crystal display and an OLED (organic light-emitting diode) display. For example, the output device 25 may output information as sound. That is, the output device 25 may include an audio device (a so-called speaker) that can output sound. For example, the output device 25 may output information on paper. That is, the output device 25 may include a printing device (a so-called printer) that can print desired information on paper. Furthermore, the input device 24 and the output device 25 may be integrally formed as a touch panel.
[0023] It should be noted that the hardware configuration shown in FIG. 2 is an example, and devices other than those shown in FIG. 2 may be added, or some devices may not be provided. Also, some devices may be replaced with other devices having similar functions. Also, some functions of the second embodiment may be provided by other devices via a network. The functions of the second embodiment may be distributed and realized among multiple devices. In this way, the hardware configuration shown in FIG. 2 can be modified as appropriate. [2-2: Model generation operation performed by the search device 2]
[0024] The flow of the model generation operation performed by the search device 2 in the second embodiment will be described with reference to Fig. 3. Fig. 3 is a flowchart showing the flow of the model generation operation performed by the search device 2 in the second embodiment.
[0025] As shown in FIG. 3 , the acquisition unit 211 acquires teacher data D including second situation information indicating the situation of a second predetermined range including the space-time where the second search target occurred and second location information indicating the location of the second search target located within the second predetermined range (step S20). The space-time where the second search target occurred may be, for example, the location of an incident such as an accident or crime. In this case, the second search target may be evidence related to the incident. Examples of evidence related to the incident include latent fingerprints, left behind articles, etc. The location of the second search target may include at least one of the location where the second search target is located and a location where the second search target is likely to be located. The second location information may be, for example, information indicating a location where evidence related to the incident may remain. The acquisition unit 211 may acquire the teacher data D stored in the storage device 22.
[0026] The output unit 212 inputs the second situation information included in the training data D acquired by the acquisition unit 211 into the search model (step S21).
[0027] The output unit 212 causes the search model to output first position information indicating the position of the second search target that exists within the second predetermined range (step S22).
[0028] The model generation unit 213 generates a search model by adjusting the parameters MP included in the search model based on the first location information output to the search model and the second location information included in the teacher data D acquired by the acquisition unit 211 (step S23). The model generation unit 213 may adjust the parameters MP included in the search model so that the second location information output to the search model is similar to the second location information included in the teacher data D acquired by the acquisition unit 211. In other words, the model generation unit 213 may generate the search model by adjusting the parameters MP included in the search model by performing machine learning using the teacher data D including the second situation information and the second location information. In other words, the search model is generated by machine learning using the teacher data D including the second situation information and the second location information. Note that a model generation device different from the search device 2 may perform the model generation operation. In this case, the search device 2 may acquire the search model generated by the model generation device from the model generation device. Furthermore, the search device 2 may not be equipped with the model generation unit 213. [2-3: Search Assistance Operation Performed by Search Device 2]
[0029] The flow of the search assist operation performed by the search device 2 in the second embodiment will be described with reference to Fig. 4. Fig. 4 is a flowchart showing the flow of the search assist operation performed by the search device 2 in the second embodiment.
[0030] 4, the acquisition unit 211 acquires first situation information indicating the situation of a first predetermined range including the time-space where the first search target occurred (step S24). The time-space where the first search target occurred may be, for example, the crime scene. In this case, the first search target may be evidence related to the crime.
[0031] The output unit 212 inputs the first situation information acquired by the acquisition unit 211 into the search model (step S25).
[0032] The output unit 212 causes the search model to output first position information indicating the position of the first search target present within the first predetermined range (sometimes referred to as the "first position") (step S26). That is, the output unit 212 outputs the first position information indicating the position of the first search target present within the first predetermined range. The first position may include at least one of the position where the first search target is present and the position where the first search target is likely to be present.
[0033] The likely location of the first search target changes depending on the situation in a first predetermined range that includes the space-time where the first search target occurred. For this reason, the search device 2 in the second embodiment uses the situation in the first predetermined range that includes the space-time where the first search target occurred to estimate the first position.
[0034] If the first location is a location where it is difficult for humans to collect evidence, a user using the first location information may, for example, direct a drone or a robot to the first location. In this case, the drone or robot executes movement control with the first location as its destination. [2-4: Examples of input and output by the search device 2]
[0035] FIG. 5 is a diagram illustrating an example of input and output by the search device 2 in the second embodiment. The acquisition unit 211 may acquire text information as the first situation information. The first predetermined range including the time and space where the first search target occurred may be a building where the incident occurred. In this case, the acquisition unit 211 may acquire text information indicating the layout of enclosed spaces such as rooms within the first predetermined range, as illustrated in FIG. 5( a). As illustrated in FIG. 5( a), the text information indicating the layout of the rooms may include text information indicating a floor plan such as a two-room, two-bedroom, and kitchen, and text information indicating the direction of each room. Furthermore, the first situation information acquired by the acquisition unit 211 may include text information indicating the intruder's entry route and text information indicating the date and time of the incident.
[0036] The first status information may be information input by a user operation. The acquisition unit 211 may be a mechanism that acquires the first status information received by the input device 24. The first status information may be input by the user performing a pull-down operation on a touch panel or the like mounted on a smartphone or the like carried by the user, for example.
[0037] The output unit 212 may also output text information as the first location information. As illustrated in FIG. 5B, the output unit 212 may output text information indicating the first location information alongside text information indicating the first situation information. FIG. 5B illustrates an example of an output indicating that in a "DK" located in the "north" of a "2DK" floor plan, the search target is likely to be found in the "corner of a table" and on the "television screen," and in a "hallway" located in the "center" of a "2DK" floor plan, the search target is likely to be found on the "floor." The output unit 212 may be a control mechanism that causes the output device 25 to output the first location information. For example, the output unit 212 may output text information as the first location information to a display or the like mounted on a smartphone or the like carried by the user. The output unit 212 may further output information indicating what the evidence is. [2-5: Technical Effects of the Search Device 2]
[0038] For example, when a forensic investigator collects evidence at a crime scene, the investigator must make inferences from various perspectives to find evidence, which often places a heavy workload on the investigator. Furthermore, inferences based on the circumstances of the crime scene often rely on past experience, so inexperienced forensic investigators often have difficulty finding sufficient evidence. For example, in the case of a home intrusion, an experienced forensic investigator knows from past experience where evidence is likely to be found, such as where intruders often place their hands on window frames, making it highly likely that fingerprints can be collected from window frames. Furthermore, in the case of a home intrusion, an experienced forensic investigator knows from past experience where evidence is likely to be found, such as where intruders often pass through gardens, making it highly likely that footprints can be collected from gardens.
[0039] The search device 2 in the second embodiment outputs first location information indicating the location of a first search target within a first predetermined range, so that the user, a forensic investigator, can be informed of locations where evidence is likely to be found, thereby reducing the workload of the forensic investigator. In particular, even inexperienced forensic investigators can be assisted in finding sufficient evidence. [3: Third Embodiment]
[0040] A third embodiment of a searching device, a searching method, and a recording medium will be described below. Hereinafter, the third embodiment of a searching device, a searching method, and a recording medium will be described using a searching device 3 to which the third embodiment of the searching device, the searching method, and the recording medium is applied.
[0041] The search device 3 in the third embodiment differs from the search device 2 in the second embodiment in the input and output targets. Other features of the search device 3 may be the same as other features of the search device 3. [3-1: Examples of input and output by the search device 3]
[0042] FIG. 6 is a diagram illustrating an example of input and output by the search device 3 in the third embodiment. In the third embodiment, the first situation information may be information related to an image of a first predetermined range including the space-time where the first search target occurred. The first predetermined range including the space-time where the first search target occurred may be a building where the incident occurred. In this case, the acquisition unit 211 may acquire, as the first situation information, information related to an image showing the layout of closed spaces such as rooms in the first predetermined range, as exemplified in FIG. 6( a). Furthermore, the first situation information acquired by the acquisition unit 211 may include information indicating the intruder's entry route and information indicating the date and time of the incident.
[0043] The acquisition unit 211 may acquire a floor plan as the first situation information generated by the autonomous mobile robot. The autonomous mobile robot may autonomously patrol a closed space, such as a building or a room within a building, and generate a floor plan of the building or a room within the building. The autonomous mobile robot may be equipped with a camera and may generate a floor plan based on captured images. The autonomous mobile robot may be equipped with at least one of a position detection device such as a GPS (Global Positioning System) sensor and an attitude detection device such as an acceleration sensor, and may generate a floor plan based on the detection results. The autonomous mobile robot may generate a two-dimensional floor plan. The autonomous mobile robot may be a flying object such as a drone and may generate a three-dimensional floor plan.
[0044] The output unit 212 may output an image as the first position information. As illustrated in FIG. 6B, the output unit 212 may output an image showing the first predetermined range with a circle indicating the first position superimposed thereon as the first position information. The output unit 212 may further output information indicating what the evidence is superimposed thereon. [3-2: Model Generation Operation Performed by Search Device 3]
[0045] The model generation unit 213 may perform machine learning using training data D including information about the image of the second predetermined range including the space-time in which the second search target occurred and the second location information, thereby adjusting parameters included in the search model and generating the search model. [3-3: Technical Effects of the Search Device 3]
[0046] The search device 3 in the third embodiment can provide information about an image to a user, thereby improving user convenience. [4: Fourth Embodiment]
[0047] A fourth embodiment of a search device, a search method, and a recording medium will be described below. Hereinafter, the fourth embodiment of the search device, the search method, and the recording medium will be described using a search device 4 to which the fourth embodiment of the search device, the search method, and the recording medium is applied. [4-1: Configuration of the search device 4]
[0048] The configuration of the search device 4 in the fourth embodiment will be described with reference to Fig. 7. Fig. 7 is a block diagram showing the configuration of the search device 4 in the fourth embodiment.
[0049] As shown in FIG. 7 , the search device 4 in the fourth embodiment includes a calculation device 21 and a storage device 22, similar to the search device 2 in the second embodiment and the search device 3 in the third embodiment. Furthermore, the search device 4 may include a communication device 23, an input device 24, and an output device 25, similar to the search device 2 in the second embodiment and the search device 3 in the third embodiment. However, the search device 4 does not have to include at least one of the communication device 23, the input device 24, and the output device 25. The search device 4 in the fourth embodiment differs from the search device 2 in the second embodiment and the search device 3 in the third embodiment in that the calculation device 21 includes a presentation unit 414. Other features of the search device 4 may be the same as other features of at least one of the search device 2 in the second embodiment and the search device 3 in the third embodiment. [4-2: Presentation Operation by the Search Device 4]
[0050] The flow of the presentation operation performed by the search device 4 in the fourth embodiment will be described with reference to Figures 8 and 9. Figure 8 is a flowchart showing the flow of the presentation operation performed by the search device 4 in the fourth embodiment.
[0051] As shown in FIG. 8 , the acquisition unit 211 acquires first situation information indicating the situation of a first predetermined range including the time-space where the first search target occurred (step S23). The output unit 212 inputs the first situation information acquired by the acquisition unit 211 to the search model (step S24), and the output unit 212 causes the search model to output first location information indicating the location of the first search target present within the first predetermined range (step S25). In step S25, the output unit 212 may output text information as the first location information, as exemplified in FIG. 5( b). Alternatively, the output unit 212 may output an image as the first location information, as exemplified in FIG. 6( b).
[0052] The presenting unit 414 determines whether there is a location indicated by the first location information that has not been confirmed by the user using the first location information (also referred to as an "overlooked location") (step S40). If there is an overlooked location (step S40: Yes), the presenting unit 414 presents the location indicated by the corresponding first location information as an overlooked location that may not have been confirmed by the user (step S41). If there is no overlooked location (step S40: No), the search assistance operation in the fourth embodiment ends.
[0053] 9A and 9B are diagrams showing examples of presentations made by the search device 4 in the fourth embodiment. Fig. 9A illustrates a case where the output unit 212 outputs text information as the first location information. As illustrated in Fig. 9A, if the "television screen" in the "dining kitchen" is the overlooked location, the presentation unit 414 may highlight the corresponding frame.
[0054] 9B illustrates an example in which the output unit 212 outputs an image as the first location information. As illustrated in FIG. 9B, if the "television screen" in the "dining kitchen" is the overlooked location, the presentation unit 414 may highlight a circle indicating the corresponding location.
[0055] The presentation unit 414 may also present the overlooked part by displaying an arrow or the like indicating the overlooked part on a head-mounted display or smart glasses worn by the user.
[0056] The operation of step S40 may be executed, for example, when the user inputs an input indicating the completion of the search operation. [4-3: Technical Effects of Search Device 4]
[0057] The search device 4 in the fourth embodiment presents the first location to the user as an overlooked location when the user has not confirmed the first location, thereby preventing the user from overlooking the first location. [5: Fifth Embodiment]
[0058] A fifth embodiment of a searching device, a searching method, and a recording medium will be described below. Hereinafter, the fifth embodiment of a searching device, a searching method, and a recording medium will be described using a searching device 5 to which the fifth embodiment of the searching device, the searching method, and the recording medium is applied.
[0059] The search device 5 of the fifth embodiment differs from the search device 4 of the fourth embodiment in the presentation operation by the presentation unit 414. Other features of the search device 5 may be the same as other features of the search device 4. [5-1: Presentation Operation by Search Device 5]
[0060] The flow of the presentation operation performed by the search device 5 in the fifth embodiment will be described with reference to Figures 10 and 11. Figure 10 is a flowchart showing the flow of the presentation operation performed by the search device 5 in the fifth embodiment.
[0061] As shown in FIG. 10 , the acquisition unit 211 acquires first situation information indicating the situation of a first predetermined range including the time-space where the first search target occurred (step S23). The output unit 212 inputs the first situation information acquired by the acquisition unit 211 to the search model (step S24), and the output unit 212 causes the search model to output first location information indicating the location of the first search target present within the first predetermined range (step S25). In step S25, the output unit 212 may output text information as the first location information, as exemplified in FIG. 5( b). Alternatively, the output unit 212 may output an image as the first location information, as exemplified in FIG. 6( b).
[0062] The presentation unit 414 determines whether confirmation status information has been acquired (step S50). The confirmation status information may include at least one of information regarding a checklist indicating the locations the user has confirmed within the first predetermined range and information regarding the user's movements within the first predetermined range. The presentation unit 414 may collect a log of the user's movements using a detection device installed in the terminal device carried by the user, the detection device including at least one of a position detection device such as a GPS (Global Positioning System) sensor and an orientation detection device such as an acceleration sensor. If the presentation unit 414 has acquired the confirmation status information (step S50: Yes), the presentation unit 414 presents the overlooked location based on the confirmation status information (step S51). Specifically, the presentation unit 414 determines whether there is an overlooked location based on the confirmation status information. For example, if information regarding a checklist, which is an example of confirmation status information, indicates that the user has not confirmed the location of the first search target indicated by the first location information, the presentation unit 414 may identify the location of the first search target indicated by the first location information as an overlooked location. For example, if information regarding the user's movement, which is an example of confirmation status information, indicates that the user has not passed the location of the first search target indicated by the first location information, the presentation unit 414 may identify the location of the first search target indicated by the first location information as an overlooked location. Thereafter, the presentation unit 414 may superimpose a check mark on the information indicating the first location information output by the output unit 212 to indicate the overlooked location. If confirmation status information has not been acquired (step S50: No), the process returns to step S50.
[0063] 11A and 11B are diagrams illustrating examples of presentations made by the search device 5 in the fifth embodiment. FIG. 11A illustrates a case in which the output unit 212 outputs text information as the first location information. As illustrated in FIG. 11A, if the "corner of the table" in the "dining room" and the "floor" in the "hallway" have been confirmed, the presentation unit 414 may superimpose a check mark in the corresponding frame. In other words, the presentation unit 414 may present locations without a check mark superimposed thereon as overlooked locations.
[0064] 11B illustrates an example in which the output unit 212 outputs an image as first position information. As illustrated in FIG. 11B, when the "corner of the table" in the "DK" and the "floor" in the "hallway" have been confirmed, the presentation unit 414 may superimpose a check mark on the circle indicating the corresponding position. In other words, the presentation unit 414 may present a circle without a check mark superimposed thereon as an overlooked location.
[0065] The presenting unit 414 determines whether or not there is an overlooked portion (step S52). The presenting unit 414 may determine whether or not there is an overlooked portion by determining whether or not there is a position where no check mark is superimposed. If there is an overlooked portion (step S52: Yes), the process returns to step S50. If there is no overlooked portion (step S52: No), the search assist operation ends.
[0066] The presenting unit 414 may also present the overlooked location by displaying an arrow or the like indicating the overlooked location on a head-mounted display or smart glasses worn by the user. [5-2: Technical Effects of the Search Device 5]
[0067] The search device 5 in the fifth embodiment presents the user at the first location with the location that has been overlooked based on at least one of information related to the checklist and information related to the user's movements, thereby more reliably preventing the user from overlooking the location. [6: Sixth Embodiment]
[0068] A sixth embodiment of a searching device, a searching method, and a recording medium will be described below. Hereinafter, the sixth embodiment of a searching device, a searching method, and a recording medium will be described using a searching device 6 to which the sixth embodiment of the searching device, the searching method, and the recording medium is applied.
[0069] The search device 6 of the sixth embodiment differs from the search devices 2 of the second embodiment to 5 of the fifth embodiment in the content output by the output unit 212. Other features of the search device 6 may be the same as at least one other feature of the search devices 2 of the second embodiment to 5 of the fifth embodiment. [6-1: Search assist operation and model generation operation by the search device 6]
[0070] The flow of the learning operation performed by the search device 6 in the sixth embodiment will be described with reference to Fig. 12. Fig. 12 is a flowchart showing the flow of the learning operation performed by the search device 6 in the sixth embodiment.
[0071] As shown in FIG. 12, the acquisition unit 211 acquires first situation information indicating the situation of a first predetermined range including the space-time in which the first search target occurred (step S23).
[0072] The output unit 212 inputs the first situation information acquired by the acquisition unit 211 into the search model, and causes the search model to output the first location information and first possibility information regarding the possibility that the user will overlook the location indicated by the first location information (step S60).
[0073] 13A and 13B are diagrams illustrating examples of outputs output by the search device 6 in the sixth embodiment. Fig. 13A illustrates a case where the output unit 212 outputs text information as the first location information. For example, if the first possibility information indicates that the possibility that the user will overlook the "television screen" in the "DK" is higher than a predetermined value, the output unit 212 may highlight the frame of the "television screen" in the "DK" as illustrated in Fig. 13A.
[0074] 13(b) illustrates an example in which the output unit 212 outputs an image as the first position information. For example, if the first possibility information indicates that the possibility that the user will miss the "television screen" in the "DK" is higher than a predetermined value, the output unit 212 may highlight a circle indicating the position of the "television screen" in the "DK" as illustrated in FIG.
[0075] The presenting unit 414 determines whether or not there is an overlooked portion (step S61). If there is an overlooked portion (step S61: Yes), the presenting unit 414 presents the overlooked portion to the user (step S62). The presenting unit 414 may present the overlooked portion to the user as in the presentation examples shown in FIGS. 9 and 11.
[0076] After the first position as the overlooked location is presented to the user, the model generation unit 213 determines whether the user has found the first search target at the first position (step S63).
[0077] If the user finds the first search target at the first location (step S63: Yes), the model generation unit 213 generates training data D (step S64). The training data D generated in step S64 includes the first situation information and the first location information as second situation information and second location information, respectively. Furthermore, the training data D generated in step S64 further includes second possibility information indicating that the possibility that the user will overlook the location indicated by the second location information is higher than a predetermined value.
[0078] The acquisition unit 211 acquires the training data D generated in step S64 (step S65).
[0079] The output unit 212 inputs the second situation information included in the training data D generated in step S64 and acquired by the acquisition unit 211 to the search model (step S66).
[0080] The output unit 212 causes the search model to output first location information indicating the location of a second search target located within a second specified range and first possibility information regarding the possibility that the user will overlook the location indicated by the first location information (step S67).
[0081] The model generation unit 213 regenerates the search model through machine learning using the teacher data D (step S68). The model generation unit 213 may regenerate the search model by adjusting the parameter MP included in the search model. The model generation unit 213 may adjust the parameter MP included in the search model based on the first location information output to the search model, the second location information included in the teacher data D acquired by the acquisition unit 211, the first possibility information output to the search model, and the second possibility information included in the teacher data D acquired by the acquisition unit 211. The model generation unit 213 may adjust the parameter MP included in the search model so that the first location information output to the search model is similar to the second location information included in the teacher data D acquired by the acquisition unit 211, and so that the first possibility information output to the search model is similar to the second possibility information included in the teacher data D acquired by the acquisition unit 211. That is, the model generation unit 213 may adjust the parameters MP included in the search model by performing machine learning using the teacher data D including the second situation information, the second location information, and the second possibility information, and regenerate the search model. In other words, the search model is regenerated by machine learning using the teacher data D including the second situation information, the second location information, and the second possibility information. [6-2: Technical Effects of the Search Device 6]
[0082] The search device 6 in the sixth embodiment outputs first possibility information in addition to the first location information, thereby enabling the user to recognize in advance locations that the user is likely to overlook. Furthermore, when the user finds the search target at a location that has been presented to the user as an overlooked location, the search device 6 regenerates a search model by machine learning using training data D that includes the first situation information, the first location information, and the first possibility information as second situation information, second location information, and second possibility information, respectively, thereby providing the user with information that will enable them to more easily find the search target.
[0083] In each of the above embodiments, the time-space where the search target occurred is the location of an incident such as an accident or crime, and the search target is evidence related to the incident, as an example. However, each of the embodiments may also be applied to search cases involving incidents other than accidents and crimes. For example, the search target may be a missing person, including a lost child, and the time-space where the search target occurred may be the location, time period, etc. where the missing person was lost. In this case, the search model may be generated by machine learning using training data D including second situation information, such as information indicating the surrounding circumstances, such as the location and time period where the missing person was lost, and second location information, such as information indicating the location where the missing person was found. The second situation information may include information indicating the age of the missing person. Furthermore, for example, the search target may be a lost item, and the time-space where the search target occurred may be the time period, etc., where the lost item was lost. In this case, the search model may be generated by machine learning using training data D including second situation information, such as information indicating the circumstances, such as the time period where the lost item was lost, and second location information, such as information indicating the location where the lost item was found. [7: Supplementary Note]
[0084] The following supplementary notes are further disclosed with respect to the above-described embodiment. [Supplementary Note 1] A search device comprising: an acquisition means for acquiring first situation information indicating a situation in a first predetermined range including a space-time in which a first search target has occurred; and an output means for outputting first position information indicating a position of the first search target present within the first predetermined range, wherein the output means inputs the first situation information acquired by the acquisition means to a search model and causes the search model to output the first position information, and the search model is generated by machine learning using training data including second situation information indicating a situation in a second predetermined range including a space-time in which a second search target has occurred and second position information indicating the position of the second search target present within the second predetermined range. [Supplementary Note 2] The search device according to Supplementary Note 1 further comprises a model generation means for generating the search model by performing machine learning using training data including the second situation information and the second position information, thereby adjusting parameters included in the search model. [Supplementary Note 3] The search device according to Supplementary Note 1 or 2, wherein the first situation information is information relating to an image of the first predetermined range including the space-time where the first search target occurred, and the second situation information is information relating to an image of the second predetermined range including the space-time where the second search target occurred. [Supplementary Note 4] The search device according to Supplementary Note 2, further comprising a presentation means for, when a user using the first location information has not confirmed the location indicated by the first location information, presenting the location indicated by the first location information as an overlooked location that may not have been confirmed by the user. [Supplementary Note 5] The search device according to Supplementary Note 4, wherein the presentation means presents the location indicated by the first location information as an overlooked location that may not have been confirmed by the user, based on at least one of information relating to a checklist indicating locations confirmed by the user within the first predetermined range and information relating to the user's movements within the first predetermined range.[Supplementary Note 6] The output means inputs the first situation information acquired by the acquisition means to the search model, and causes the search model to output the first location information and first possibility information regarding the possibility that the user will overlook the location indicated by the first location information, and the model generation means, when the user finds the first search target at the location indicated by the first location information after the location indicated by the first location information has been presented to the user as the overlooked location, regenerates the search model by machine learning using training data that includes the first situation information and the first location information as the second situation information and the second location information, respectively, and further includes second possibility information indicating that the possibility that the user will overlook the location indicated by the second location information is higher than a predetermined value. The search device described in Supplementary Note 4 or 5. [Supplementary Note 7] A search method that acquires first situation information indicating the situation of a first predetermined range including the space-time where a first search target occurred, and outputs first location information indicating the location of the first search target that exists within the first predetermined range, wherein the output is performed by inputting the acquired first situation information into a search model and causing the search model to output the first location information, and the search model is generated by machine learning using training data that includes second situation information indicating the situation of a second predetermined range including the space-time where a second search target occurred, and second location information indicating the location of the second search target that exists within the second predetermined range. [Supplementary Note 8] A recording medium having recorded thereon a computer program for causing a computer to execute a search method, the search method comprising: acquiring first situation information indicating the situation of a first predetermined range including the space-time in which a first search target occurred; and outputting first position information indicating the position of the first search target that exists within the first predetermined range; wherein the output is performed by inputting the acquired first situation information into a search model and causing the search model to output the first position information; and the search model is generated by machine learning using training data including second situation information indicating the situation of a second predetermined range including the space-time in which a second search target occurred and second position information indicating the position of the second search target that exists within the second predetermined range.
[0085] Although the present disclosure has been described above with reference to the embodiments, the present disclosure is not limited to the above-described embodiments. Various modifications that would be understood by a person skilled in the art can be made to the configuration and details of the present disclosure within the scope of the present disclosure. Furthermore, each embodiment can be combined with other embodiments as appropriate. Furthermore, to the extent permitted by law, the disclosures of all documents (e.g., published patent applications) cited in the above disclosure are incorporated by reference as part of the description of this disclosure.
[0086] To the extent permitted by law, this application claims priority to Japanese Patent Application No. 2022-104556, filed June 29, 2022, the disclosure of which is incorporated herein in its entirety by reference.
[0087] 1, 2, 3, 4, 5, 6 Search device 11, 211 Acquisition unit 12, 212 Output unit 213 Model generation unit 414 Presentation unit
Claims
1. An acquisition means for acquiring first situation information indicating a situation of a first predetermined range including a time-space in which the first search target occurs; an output means for outputting first position information indicating a position of the first search target existing within the first predetermined range; Equipped with The output means inputs the first situation information acquired by the acquisition means to a search model, and causes the search model to output the first position information; The search model is generated by machine learning using training data including second situation information indicating a situation of a second predetermined range including a time and space in which a second search target occurs and second location information indicating a location of the second search target existing within the second predetermined range. Search equipment.
2. The vehicle further includes a model generating unit that generates the search model by performing machine learning using teacher data including the second situation information and the second position information to adjust parameters included in the search model. The search device according to claim 1.
3. The first situation information is information regarding an image of the first predetermined range including a time-space in which the first search target occurred, The second situation information is information regarding an image of the second predetermined range including a time and space in which the second search target occurred. The search device according to claim 1 or 2.
4. The information processing device further includes a presentation unit that, when a user who uses the first location information has not confirmed a location indicated by the first location information, presents the location indicated by the first location information as an overlooked location that may not have been confirmed by the user. The search device according to claim 2.
5. The presenting means presents the position indicated by the first position information as an overlooked location that may not have been confirmed by the user, based on at least one of information on a checklist indicating the positions confirmed by the user within the first specified range and information on the movement of the user within the first specified range. The search device according to claim 4.
6. The output means inputs the first situation information acquired by the acquisition means to the search model, and causes the search model to output the first location information and first possibility information regarding a possibility that the user will overlook a location indicated by the first location information; When the user finds the first search target at the location indicated by the first location information after the location indicated by the first location information is presented to the user as the overlooked location, the model generation means regenerates the search model by machine learning using teacher data including the first situation information and the first location information as the second situation information and the second location information, respectively, and further including second possibility information indicating that the possibility that the user will overlook the location indicated by the second location information is higher than a predetermined value. A search device according to claim 4 or 5.
7. Acquire first situation information indicating a situation of a first predetermined range including a time-space in which the first search target occurs; A search method for outputting first position information indicating a position of the first search target existing within the first predetermined range, The output is performed by inputting the acquired first situation information to a search model and causing the search model to output the first position information; The search model is generated by machine learning using training data including second situation information indicating a situation of a second predetermined range including a time and space in which a second search target occurs and second location information indicating a location of the second search target existing within the second predetermined range. A computer implemented search method.
8. On the computer, Acquire first situation information indicating a situation of a first predetermined range including a time-space in which the first search target occurs; A search method for outputting first position information indicating a position of the first search target existing within the first predetermined range, The output is performed by inputting the acquired first situation information to a search model and causing the search model to output the first position information; The search model is generated by machine learning using training data including second situation information indicating a situation of a second predetermined range including a time and space in which a second search target occurs and second location information indicating a location of the second search target existing within the second predetermined range. A computer program for carrying out a search method.