Stereocognition ability evaluation system, stereocognition ability evaluation device, stereocognition ability evaluation program, and stereocognition ability evaluation method

Abstract

The present invention realizes objective evaluation of stereognostic capabilities. This stereognostic capability evaluation system (100) evaluates the stereognostic capabilities of a person being measured on the basis of a change in the position of a moving object (Cs1) that moves upon being manipulated by the person being measured. A position acquisition unit (101b) acquires information relating to the position of the moving object (Cs1). A stereognostic capability assessment unit (101e) assesses the stereognostic capabilities of the person being measured from at least one of a first feature value derived from a value for which a first distance between the moving object (Cs1) and a reference object preceding the moving object (Cs1) in the advancement direction of the moving object (Cs1) is temporally differentiated at least once, and a second feature value derived from a second distance between the moving object (Cs1) and a reference line that follows the advancement direction of the moving object (Cs1).

Description

[Technical field]

[0001] The present disclosure relates to a stereoscopic cognitive ability assessment system, a stereoscopic cognitive ability assessment device, a stereoscopic cognitive ability assessment program, and a stereoscopic cognitive ability assessment method. [Background technology]

[0002] Conventionally, systems for evaluating the cognitive function of a subject based on the subject's response have been known. For example, Japanese Patent Publication No. 6000968 (Patent Document 1) discloses a system using a portable touch screen personal computing device as a system for evaluating cognitive function using a visual measurement. In this system, an individual's cognitive evaluation is performed based on the reaction speed to a displayed cognitive evaluation stimulus. In the cognitive evaluation test, the reaction time from when a character is displayed on the display of the personal computing device to when the user responds by pressing a button is measured.

[0003] JP2015-502238A (Patent Document 2) discloses a system that provides a video game for mapping a subject's peripheral vision, including a test in which the subject finds a visual stimulus presented for a short period of time. In this system, targets are displayed on a display, and the visual field defect due to glaucoma is measured based on the user's response to the targets. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Patent No. 6000968 [Patent Document 2] Special Publication No. 2015-502238 Summary of the Invention [Problem to be solved by the invention]

[0005] Stereocognition is the ability to recognize the distance between objects and respond appropriately, and can be expressed as a sense of perspective. In recent years, it has been known that stereocognition tends to decline when cognitive function declines in the elderly and other people. Therefore, evaluating stereocognition may have an effect similar to evaluating cognitive function.

[0006] However, in the system disclosed in Patent Document 1, the displayed object is moved, but the user's perspective of the object is not taken into consideration. In addition, in this system, the measurement result of the cognitive function is influenced by various attributes such as the subject's concentration level or the degree of proficiency in the operation method, so it is difficult to objectively evaluate the user's three-dimensional cognitive ability. Furthermore, in the system disclosed in Patent Document 2, the distance between the subject and the monitor is measured, but the subject's perspective of the displayed target is not taken into consideration.

[0007] The present disclosure has been made to solve the above-mentioned problems, and has an object to realize an objective evaluation of stereoscopic perception ability. [Means for solving the problem]

[0008] A stereoscopic cognitive ability assessment system according to one aspect of the present disclosure assesses the stereoscopic cognitive ability of a subject based on a change in position of a moving object that is moved by being operated by the subject. The stereoscopic cognitive ability assessment system includes a position acquisition unit and a stereoscopic cognitive ability determination unit. The position acquisition unit acquires information regarding the position of the moving object. The stereoscopic cognitive ability determination unit determines the stereoscopic cognitive ability of the subject from at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in the moving direction of the moving object, and a second feature amount derived from a second distance between the moving object and a reference line along the moving direction of the moving object.

[0009] A stereoscopic cognitive ability assessment system according to another aspect of the present disclosure assesses the stereoscopic cognitive ability of a subject based on a change in behavior of a moving object that is moved by being operated by the subject. The stereoscopic cognitive ability assessment system includes a behavior acquisition unit and a stereoscopic cognitive ability determination unit. The behavior acquisition unit acquires information regarding the behavior of the moving object. The stereoscopic cognitive ability determination unit determines the stereoscopic cognitive ability from at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in the moving direction of the moving object, and a second feature amount relating to a second distance between the moving object and a reference line along the moving direction of the moving object.

[0010] A stereoscopic cognitive ability assessment device according to another aspect of the present disclosure evaluates the stereoscopic cognitive ability of a subject based on a change in the position of a moving object that is moved by being operated by the subject. The stereoscopic cognitive ability assessment device includes a position acquisition unit, a stereoscopic cognitive ability determination unit, and a housing. The position acquisition unit acquires information on the position of the moving object. The stereoscopic cognitive ability determination unit determines the stereoscopic cognitive ability of the subject from at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in the moving direction of the moving object, and a second feature amount derived from a second distance between the moving object and a reference line along the moving direction of the moving object. The housing accommodates the position acquisition unit and the stereoscopic cognitive ability determination unit.

[0011] A stereoscopic cognitive ability assessment device according to another aspect of the present disclosure assesses the stereoscopic cognitive ability of a subject based on a change in behavior of a moving object that is moved by being operated by the subject. The stereoscopic cognitive ability assessment device includes a behavior acquisition unit, a stereoscopic cognitive ability determination unit, and a housing. The behavior acquisition unit acquires information regarding the behavior of the moving object. The stereoscopic cognitive ability determination unit determines the stereoscopic cognitive ability from at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in the moving direction of the moving object, and a second feature amount relating to a second distance between the moving object and a reference line along the moving direction of the moving object. The housing accommodates the behavior acquisition unit and the stereoscopic cognitive ability determination unit.

[0012] A stereoscopic cognitive ability evaluation program according to another aspect of the present disclosure is executed by a computer to configure a stereoscopic cognitive ability evaluation device for evaluating the stereoscopic cognitive ability of a subject based on a change in the position of a moving object that is moved by being operated by the subject. The stereoscopic cognitive ability evaluation device includes a position acquisition unit and a stereoscopic cognitive ability determination unit. The position acquisition unit acquires information on the position of the moving object. The stereoscopic cognitive ability determination unit evaluates the stereoscopic cognitive ability of the subject from at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in the moving direction of the moving object, and a second feature amount derived from a second distance between the moving object and a reference line along the moving direction of the moving object.

[0013] A stereoscopic cognition ability evaluation program according to another aspect of the present disclosure is executed by a computer to cause the computer to configure a stereoscopic cognition ability evaluation device for evaluating the stereoscopic cognition ability of a subject based on a change in behavior of a moving object that is moved by being operated by the subject. The stereoscopic cognition ability evaluation device includes a behavior acquisition unit and a stereoscopic cognition ability determination unit. The behavior acquisition unit acquires information regarding the behavior of the moving object. The stereoscopic cognition ability determination unit evaluates the stereoscopic cognition ability from at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in the moving direction of the moving object, and a second feature amount relating to a second distance between the moving object and a reference line along the moving direction of the moving object.

[0014] A stereoscopic cognition ability evaluation method according to another aspect of the present disclosure evaluates the stereoscopic cognition ability of a subject based on a change in the position of a moving object that is moved by being operated by the subject. The stereoscopic cognition ability evaluation method includes a step of acquiring information regarding the position of the moving object. The stereoscopic cognition ability evaluation method further includes a step of determining the stereoscopic cognition ability of the subject from at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in the moving direction of the moving object, and a second feature amount derived from a second distance between the moving object and a reference line along the moving direction of the moving object.

[0015] A stereoscopic cognition ability evaluation method according to another aspect of the present disclosure evaluates the stereoscopic cognition ability of a subject based on a change in behavior of a moving object that is moved by being operated by the subject. The stereoscopic cognition ability evaluation method includes a step of acquiring information about the behavior of the moving object. The stereoscopic cognition ability evaluation method further includes a step of determining the stereoscopic cognition ability from at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in the moving direction of the moving object, and a second feature amount relating to a second distance between the moving object and a reference line along the moving direction of the moving object. Effect of the Invention

[0016] According to the stereoscopic perception ability assessment system, stereoscopic perception ability assessment device, stereoscopic perception ability assessment program, and stereoscopic perception ability assessment method disclosed herein, an objective assessment of the stereoscopic perception ability of the subject can be realized by determining the stereoscopic perception ability of the subject from at least one of a first feature derived from a value obtained by time-differentiating at least once a first distance between a reference object and a moving object, and a second feature related to a second distance between a reference line and the moving object. [Brief description of the drawings]

[0017] [Figure 1] 1 is a block diagram showing a configuration of a stereoscopic perception ability evaluation system according to a first embodiment. [Diagram 2] FIG. 2 is a diagram illustrating a hardware configuration of the stereocognition ability evaluation system of FIG. 1. [Diagram 3] 3 is a functional block diagram showing a configuration of a function of evaluating stereoscopic perception ability of the processor in FIG. 2. FIG. [Figure 4] FIG. 3 is a functional block diagram showing the configuration of a machine learning function of the processor in FIG. 2. [Diagram 5] 3 is a diagram for explaining distances related to an automobile required for deriving feature amounts to be input to the evaluation model of FIG. 2. FIG. [Figure 6]6 is a time chart of the speeds of the preceding vehicle and the automobile in FIG. 5 . [Figure 7] 7 is a time chart showing a result of identifying a section in which the preceding vehicle is traveling at a constant speed in the time chart of FIG. 6. [Figure 8] FIG. 4 is a diagram showing a plurality of feature amounts derived from travel data of a plurality of sections. [Figure 9] FIG. 6 shows the trajectory of the preceding vehicle in FIG. 5 together with the trajectory of the vehicle itself when the vehicle is driven by a driver labeled with a high level of stereoscopic perception ability. [Figure 10] FIG. 6 shows the trajectory of the preceding vehicle in FIG. 5 together with the trajectory of the vehicle itself when the vehicle is driven by a driver labeled as having poor stereoscopic perception ability. [Figure 11] FIG. 3 is a diagram showing an example of the decision tree in FIG. 2, the decision tree including, as explanatory variables, feature quantities derived from interval acceleration. [Figure 12] FIG. 12 is a diagram showing a classification result when a driver with low stereoscopic perception ability is added to the classification targets of the decision tree in FIG. [Figure 13] FIG. 3 is a diagram showing another example of the decision tree in FIG. 2, which is a decision tree DT2 including, as explanatory variables, feature amounts derived from the absolute value of the slide amount and the absolute value of the slide speed. [Figure 14] FIG. 3 is another example of the decision tree in FIG. 2, showing a decision tree including, as explanatory variables, feature quantities derived from the absolute value of the slide amount, the absolute value of the slide speed, and the absolute value of the slide acceleration. [Figure 15] FIG. 3 is another example of the decision tree in FIG. 2, showing a decision tree including, as explanatory variables, feature quantities derived from the absolute value of the slide amount, the absolute value of the slide speed, and the absolute value of the slide acceleration. [Figure 16] FIG. 3 is a diagram showing another example of the decision tree in FIG. 2, which includes a feature amount derived from the slide amount as an explanatory variable. [Figure 17] FIG. 3 is a diagram showing another example of the decision tree in FIG. 2, which includes a feature amount derived from the slide amount as an explanatory variable. [Figure 18]3 is a flowchart showing a process flow in a stereocognition ability evaluation method performed by a processor that executes the stereocognition ability evaluation program of FIG. 2. [Figure 19] FIG. 13 is a block diagram showing a configuration of a stereoscopic perception ability evaluation system according to a modified example of the first embodiment. [Figure 20] FIG. 20 is a diagram illustrating a hardware configuration of the stereocognition ability evaluation system in FIG. 19. [Figure 21] FIG. 13 is a diagram showing how the stereoscopic perception ability of a subject is measured through a measurement test using a driving simulation in the stereoscopic perception ability assessment system according to the second embodiment. [Figure 22] 22 is a diagram showing an example of a screen displayed on a subject via a stereocognition ability evaluation device in a driving simulation performed by the stereocognition ability evaluation system of FIG. 21. FIG. [Diagram 23] FIG. 22 is an external perspective view of the stereocognition ability evaluation device of FIG. 21. [Figure 24] FIG. 24 is a diagram illustrating a hardware configuration of the stereocognition ability evaluation device in FIG. 23. [Diagram 25] FIG. 1 is a diagram showing the concept of an optotype used to measure eye function. [Figure 26] FIG. 2 is a diagram showing the relationship between the target distance (the distance between the subject's eye and the target) and pupil diameter due to the pupillary near-distance reflex. [Figure 27] FIG. 1 is a diagram showing the relationship between target distance and pupil position (convergence and divergence movement). [Figure 28] 25 is a functional block diagram showing the configuration of a function of evaluating the stereoscopic perception ability of the processor in FIG. 24. FIG. [Figure 29] 25 is a flowchart showing a process flow in a stereocognition ability evaluation method performed by a processor that executes the stereocognition ability evaluation program of FIG. 24. [Diagram 30] FIG. 11 is a diagram showing a hardware configuration of a stereoscopic perception ability evaluation system according to a third embodiment. [Diagram 31] FIG. 31 is a functional block diagram showing the configuration of a function of evaluating stereoscopic perception ability of the processor in FIG. 30. [Diagram 32]FIG. 31 is a functional block diagram showing the configuration of a machine learning function of the processor of FIG. 30. [Diagram 33] FIG. 13 is a diagram illustrating a hardware configuration of a stereoscopic perception ability evaluation system according to a modified example of the third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0018] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. In the drawings, the same or corresponding parts are designated by the same reference characters, and their description will not be repeated in principle.

[0019] [Embodiment 1] Fig. 1 is a block diagram showing a configuration of a stereoscopic cognitive ability evaluation system 100 according to the first embodiment. As shown in Fig. 1, the stereoscopic cognitive ability evaluation system 100 includes an information processing device 110 (stereoscopic cognitive ability evaluation device), a terminal device 800 of a subject Sb1, and a drive recorder 900 (movement recording device) mounted on an automobile Cs1 (moving object). The information processing device 110, the terminal device 800, the automobile Cs1, and the drive recorder 900 are connected to each other via a network NW. The network NW includes, for example, the Internet, a LAN (Local Area Network), or a cloud system.

[0020] The subject Sb1 is a subject whose stereoscopic cognitive ability is measured by the stereoscopic cognitive ability evaluation system 100. The subject Sb1 gets into an automobile Cs1 and drives the automobile Cs1. Travel data (video data) of the automobile Cs1 driven by the subject Sb1 is recorded by a drive recorder 900.

[0021] The information processing device 110 acquires travel data of the subject Sb1 from the drive recorder 900. The information processing device 110 derives an index (feature amount) that characterizes a change in the position of the automobile Cs1 that is operated and moved by the subject Sb1 from the travel data. The information processing device 110 evaluates the stereoscopic cognitive ability of the subject Sb1 based on the feature amount. The information processing device 110 transmits the evaluation result of the stereoscopic cognitive ability to the terminal device 800 of the subject Sb1. The information processing device 110 includes, for example, a personal computer or a workstation. The information processing device 110 may be mounted on the automobile Cs1.

[0022] The subject Sb1 can check the evaluation result by referring to the terminal device 800 wherever the subject Sb1 can establish a connection between the terminal device 800 and the network NW. The terminal device 800 includes, for example, a smartphone capable of wireless communication.

[0023] Fig. 2 is a diagram showing a hardware configuration of the stereoscopic perception ability evaluation system 100 of Fig. 1. As shown in Fig. 2, the information processing device 110 includes a processor 101, a RAM (Random Access Memory) 102, a storage 103, a communication unit 104, a memory interface 105, and a housing Hs1. The housing Hs1 contains the processor 101, the RAM (Random Access Memory) 102, the storage 103, the communication unit 104, and the memory interface 105.

[0024] The processor 101 is a processing circuit for executing various functions that control the operation of the information processing device 110. The processor 101 includes a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit) that operates an information device such as a computer.

[0025] The RAM 102 includes a volatile memory and is used as a work area when the processor 101 operates, a temporary data storage area, and the like.

[0026] The storage 103 includes a non-volatile memory such as a flash ROM (Read Only Memory) or an SSD (Solid State Drive). The storage 103 stores computer programs and data referenced when the programs are executed. For example, the storage 103 stores an OS (Operating System) program (not shown), a stereoscopic perception ability evaluation program 103a, and a machine learning program 103b as the computer programs. The storage 103 also stores an evaluation model 103c (specific model) and learning data 103d as data referenced when the computer programs are executed.

[0027] The evaluation model 103c evaluates the stereoscopic cognitive ability of the subject Sb1 from the feature amount derived from the driving data of the subject Sb1. The evaluation model 103c may be a classification model that classifies the stereoscopic cognitive ability of the subject Sb1 into one of a plurality of predetermined levels, or may be a regression model that outputs a numerical value corresponding to the stereoscopic cognitive ability. The evaluation model 103c is referenced by the stereoscopic cognitive ability evaluation program 103a. The evaluation model 103c includes a decision tree DT (Decision Tree). The evaluation model 103c may include, for example, a neural network or a support vector machine. The learning data 103d is referenced by the machine learning program 103b. The machine learning program 103b executes, for example, CART (Classification and Regression Trees) as a machine learning algorithm. The learning data 103d includes the driving data (teacher data) of the driver that is labeled in advance according to the driver's cognitive ability. Examples of labels that can be attached to driving data include the level of stereoscopic perception ability determined in advance from existing visual-spatial cognition assessment batteries such as mental rotation tasks, visual ability assessment tests, "seeing ability" comprehension tests, K-ABC (Kaufman Assessment Battery for Children) tests, visual perception skills tests, and motor-exclusion visual cognition tests, as well as the driver's driving skills, occupation, age, or driving history (information about the attributes of the person being measured).

[0028] The communication unit 104 communicates with each of the terminal device 800, the automobile Cs1, and the drive recorder 900 via the network NW. The communication unit 104 and the network NW may be connected by wired communication (e.g., Ethernet (registered trademark)) or wireless communication (e.g., Wi-Fi (registered trademark)).

[0029] The memory interface 105 enables access to an external storage medium of the information processing device 110. The memory interface 105 supports a Secure Digital (SD) card and a Universal Serial Bus (USB).

[0030] An input / output unit (not shown) that receives input from a user and displays to the user the processing results of the processor 101 may be connected to the information processing device 110. The input / output unit includes, for example, a display, a touch panel, a keyboard, a speaker, a lamp, and a mouse.

[0031] The drive recorder 900 includes a processor 901, a camera 902, a RAM 903, a communication unit 904, a storage 905, a memory interface 906, and a position measurement unit 907. The processor 901, the RAM 903, the communication unit 904, and the memory interface 906 have substantially the same functions as the processor 101, the RAM 102, the communication unit 104, and the memory interface 105 of the information processing device 110, respectively, and therefore descriptions of the functions of the processor 901, the RAM 903, the communication unit 904, and the memory interface 906 will not be repeated.

[0032] The camera 902 generates driving data 905a when the driving direction of the automobile Cs1 is viewed from the driver's seat of the automobile Cs1. The driving data 905a is stored in a storage 905. The driving data 905a is transmitted to the information processing device 110 by a communication unit 904. A position measurement unit 907 measures the position of the automobile Cs1 by using a global navigation satellite system (GNSS) such as a global positioning system (GPS).

[0033] The automobile Cs1 transmits data on the reaction of the subject Sb1 who is operating the automobile Cs1 to the information processing device 110. Specifically, the automobile Cs1 transmits data on the reaction of the subject Sb1 to the information processing device 110 based on input information (for example, the rotation angle of the steering wheel, the depression of the accelerator pedal, and the depression of the brake pedal) of the subject to the steering wheel, accelerator pedal, brake pedal, and the like. Note that the reaction means recognizing the proximity of a moving object and responding accordingly.

[0034] The stereoscopic cognition ability assessment function of the stereoscopic cognition ability assessment system 100 is realized by the processor 101 reading out a stereoscopic cognition ability assessment program 103a stored in the storage 103 and executing it by utilizing the work area of ​​the RAM 102. By executing the stereoscopic cognition ability assessment program 103a, modules that realize various functions related to stereoscopic cognition ability assessment are formed, and the modules execute operations that realize the functions.

[0035] Fig. 3 is a functional block diagram showing the configuration of the stereoscopic cognitive ability evaluation function of the processor 101 in Fig. 2. In the stereoscopic cognitive ability evaluation system 100, a module forming functional blocks, namely, a position acquisition unit 101b, a reaction input unit 101d, and a stereoscopic cognitive ability determination unit 101e, is configured by executing a stereoscopic cognitive ability evaluation program 103a stored in a storage 103 by the processor 101. Therefore, in Fig. 3, instead of the processor 101 and the stereoscopic cognitive ability evaluation program 103a in Fig. 2, functional blocks realized by them are shown. Below, these functional blocks will be described.

[0036] The position acquisition unit 101b acquires travel data of the subject Sb1 and a measurement result (positioning result) of the position of the automobile Cs1 from the drive recorder 900. The position acquisition unit 101b acquires information about the position of the automobile Cs1 from the travel data and the positioning result. The position acquisition unit 101b outputs information about the position of the automobile Cs1 to the stereoscopic perception ability assessment unit 101e. The information about the position of the automobile Cs1 includes information from which at least one of a value (interval speed) obtained by differentiating the distance between the automobile Cs1 and another automobile (preceding automobile) once with respect to time, a value (interval acceleration) obtained by differentiating the distance twice with respect to time, a distance (slide amount) between the automobile Cs1 and a reference line along the traveling direction of the automobile Cs1, a value (slide speed) obtained by differentiating the amount of slide once with respect to time, and a value (slide acceleration) obtained by differentiating the amount of slide twice with respect to time. For example, information regarding the position of the automobile Cs1 includes a correspondence between time and three-dimensional space coordinates of the automobile Cs1, the three-dimensional space coordinates of each of at least one reference position, and information capable of identifying a reference line in the three-dimensional space.

[0037] The reaction input unit 101d receives input of the subject Sb1's active reaction corresponding to the three-dimensional position of the car Cs1 recognized by the subject Sb1. The reaction input unit 101d identifies the reaction of the subject Sb1 who is driving the car Cs1 based on input information of the subject Sb1 to the steering wheel, accelerator pedal, brake pedal, etc. of the car Cs1. The reaction input unit 101d outputs information related to the reaction to the stereoscopic perception ability assessment unit 101e.

[0038] An active response means a proactive operation for achieving a predetermined purpose, and specifically means an operation performed on a target whose position is dynamically determined by the position of the automobile Cs1. An active response includes, for example, an operation to move the automobile Cs1 closer to a predetermined location. In this case, the difference between the position of the automobile Cs1 and the position of the predetermined location is dynamically determined by the position of the automobile Cs1, and the goal is to reduce the difference. An active response also includes, for example, an operation to keep the distance between the automobile Cs1 and another automobile (preceding automobile) traveling ahead constant. In this case, the difference between the preceding automobile and the automobile Cs1 is dynamically determined by the position of the preceding automobile, and the goal is to keep the difference constant.

[0039] The stereoscopic cognitive ability assessment unit 101e derives a feature amount derived from at least one of the interval speed, interval acceleration, absolute value of the sliding amount, absolute value of the sliding speed, absolute value of the sliding acceleration, sliding amount, sliding speed, and sliding acceleration of the automobile Cs1 based on information about the position of the automobile Cs1. The stereoscopic cognitive ability assessment unit 101e uses the evaluation model 103c to assess the stereoscopic cognitive ability of the subject Sb1 from the feature amount. Information about the reaction from the reaction input unit 101d may be used to assess the stereoscopic cognitive ability of the subject Sb1. The stereoscopic cognitive ability assessment unit 101e transmits the evaluation result of the stereoscopic cognitive ability to the terminal device 800 of the subject Sb1 via the communication unit 104.

[0040] In the stereoscopic perception ability assessment system 100, the processor 101 executes the machine learning program 103b, so that the information processing device 110 functions as a learning device that generates a trained assessment model 103c.

[0041] Fig. 4 is a functional block diagram showing the configuration of the machine learning function of the processor 101 in Fig. 2. As shown in Fig. 4, in the stereoscopic perception ability assessment system 100, the machine learning program 103b is executed by the processor 101 to configure modules forming functional blocks of a position acquisition unit 111b and a learning unit 101f. Therefore, in Fig. 4, instead of the processor 101 and the machine learning program 103b in Fig. 2, functional blocks realized by them are shown. Below, these functional blocks will be described.

[0042] The position acquisition unit 111b acquires driving data from the learning data 103d. The position acquisition unit 111b acquires information about the position of the vehicle corresponding to the driving data, similar to the position acquisition unit 101b in Fig. 3. The position acquisition unit 111b outputs information about the position of the vehicle and a label attached to the driving data to the learning unit 101f.

[0043] The learning unit 101f derives a feature amount derived from at least one of the interval speed, interval acceleration, absolute value of the slide amount, absolute value of the slide speed, absolute value of the slide acceleration, slide amount, slide speed, and slide acceleration of the vehicle based on the information on the vehicle's position acquired from the position acquisition unit 111b. The learning unit 101f performs machine learning on the evaluation model 103c using the feature amount and the label attached to the information on the vehicle's position, and sets the evaluation model 103c as a trained model. For example, the learning unit 101f executes CART to generate a decision tree DT with the driver's stereoscopic perception ability corresponding to the feature amount as a target variable and the feature amount as an explanatory variable.

[0044] If the distance between the reference object and the moving object is defined as a first distance, and the distance (slide amount) between the moving object and a reference line along the moving object's direction of travel is defined as a second distance, the present disclosure includes embodiments in the following range.

[0045] That is, the stereocognitive ability assessment system of the present disclosure assesses the stereocognitive ability of a subject based on a change in the position of a moving object that is moved by being operated by the subject. The stereocognitive ability assessment system includes a position acquisition unit and a stereocognitive ability determination unit. The position acquisition unit acquires information regarding the position of the moving object.

[0046] The stereoscopic perception ability assessment unit assesses the subject's stereoscopic perception ability from at least one of a first feature amount relating to a first distance between the moving object and a reference object that precedes the moving object in the moving direction of the moving object, and a second feature amount relating to a second distance between the moving object and a reference line along the moving direction of the moving object.

[0047] Furthermore, the position acquisition unit acquires information on the position of the moving object, and the first feature amount and / or the second feature amount are derived from the information on the position of the moving object. The first feature amount may be derived from any one of a value of a first distance, a value obtained by differentiating the first distance once with respect to time, and a value obtained by differentiating the first distance twice with respect to time. The second feature amount may be derived from any one of a value of a second distance, a value obtained by differentiating the second distance once with respect to time, and a value obtained by differentiating the second distance twice with respect to time.

[0048] FIG. 5 is a diagram for explaining the distance related to the automobile Cs1 required to derive the feature amount input to the evaluation model 103c of FIG. 2. As shown in FIG. 5, the preceding vehicle Ca1 is traveling ahead of the automobile Cs1 driven by the subject Sb1. The preceding vehicle Ca1 and the automobile Cs1 are traveling along a traveling direction Dr1. The sliding direction Dr2 is perpendicular to the traveling direction Dr1. The preceding vehicle Ca1 includes a reference position Pr1 defined as the position of the preceding vehicle Ca1. The automobile Cs1 includes a reference position Pr0 defined as the position of the automobile Cs1. The distance (first distance) between the automobile Cs1 and the preceding vehicle Ca1 is defined as the magnitude of a vector Vc1 directed from the reference position Pr0 toward the reference position Pr1.

[0049] The reference line VL1 is parallel to the traveling direction Dr1 and passes through the reference position Pr1. The reference line VL1 includes the reference position Pr2. The vector Vc2 from the reference position Pr2 to Pr0 is perpendicular to the traveling direction Dr1. The distance (second distance) between the automobile Cs1 and the reference line VL1 is defined as a value obtained by adding a sign according to the direction of the vector Vc2 in the sliding direction Dr2 to the magnitude of the vector Vc2. That is, when the vector Vc2 is directed in the positive direction of the sliding direction Dr2, the sliding amount is a positive value. When the vector Vc2 is directed in the negative direction of the sliding direction Dr2, the sliding amount is a negative value. The sliding amount can be defined even when there is no preceding automobile Ca1. In this case, the reference line VL1 is, for example, the center line of the lane in which the automobile Cs1 is traveling. Therefore, even when there is no preceding automobile Ca1, the stereoscopic perception ability of the subject Sb1 can be evaluated using the feature amount derived from the sliding amount.

[0050] 6 to 8, the following will specifically explain what features are derived in the stereoscopic cognitive ability assessment system 100 to assess the stereoscopic cognitive ability of the subject Sb1. FIG. 6 is a time chart of the speeds of the preceding vehicle Ca1 and the vehicle Cs1 in FIG. 5. The graph Cva shows the relationship between the speed of the preceding vehicle Ca1 measured at each sampling time and the sampling time. The graph Cvs shows the relationship between the speed of the vehicle Cs1 measured at each sampling time and the sampling time.

[0051] FIG. 7 is a time chart showing the result of identifying the section where the preceding vehicle Ca1 is traveling at a constant speed in the time chart of FIG. 6. The section where the preceding vehicle Ca1 is traveling at a constant speed is a section where the change in the speed of the preceding vehicle Ca1 is included in a predetermined range. As shown in FIG. 7, a plurality of sections Sc1, Sc2, Sc3, Sc4, Sc5, and Sc6 are identified as the section where the preceding vehicle Ca1 is traveling at a constant speed. Five or fewer sections may be extracted as the section where the preceding vehicle Ca1 is traveling at a constant speed, or seven or more sections may be extracted. By deriving the feature amount using the travel data of the section where the preceding vehicle Ca1 is traveling at a constant speed, it is possible to reduce the inclusion of noise (information that is less related to the stereoscopic perception ability) in the feature amount. As a result, it is possible to improve the accuracy of the evaluation of the stereoscopic perception ability of the subject Sb1. It is also possible to evaluate the stereoscopic perception ability of the subject Sb1 by using the feature amount derived from the travel data of the section where the preceding vehicle Ca1 is not traveling at a constant speed.

[0052] FIG. 8 is a diagram showing a number of feature quantities derived from driving data of a number of sections. In the following, "X_m" represents the average value (section average value) of value X in one section, and "X_s" represents the standard deviation (section standard deviation) of value X in one section. "X_mm" represents the average value (section representative average value) of section average values ​​(section representative values) of value X, and "X_ms" represents the standard deviation (section representative standard deviation) of section average values ​​of value X. "X_sm" represents the average value (section representative average value) of section standard deviations (section representative values) of value X, and "X_ss" represents the standard deviation (section representative standard deviation) of section standard deviations of value X. Value X1 represents a value obtained by time-differentiating value X0 once, and value X2 represents a value obtained by time-differentiating value X0 twice.

[0053] As shown in FIG. 8, from the driving data for each sampling time, the vehicle distance D0, interval speed D1, interval acceleration D2, slide amount S0, slide speed S1, slide acceleration S2, absolute value AS0 of the slide amount S0, absolute value AS1 of the slide speed, and absolute value AS2 of the slide acceleration are calculated.

[0054] When focusing on an interval ScN (N is a natural number), an interval average value D0_m and an interval standard deviation D0_s are calculated as feature quantities for the vehicle distance D0. An interval average value D1_m and an interval standard deviation D1_s are calculated as feature quantities for the interval speed D1. An interval average value D2_m and an interval standard deviation D2_s are calculated as feature quantities for the interval acceleration D2.

[0055] For the slide amount S0, an interval average value S0_m and an interval standard deviation S0_s are calculated as feature quantities. For the slide velocity S1, an interval average value S1_m and an interval standard deviation S1_s are calculated as feature quantities. For the slide acceleration S2, an interval average value S2_m and an interval standard deviation S2_s are calculated as feature quantities.

[0056] For the absolute value AS0 of the slide amount S0, an interval average value AS0_m and an interval standard deviation AS0_s are calculated as feature quantities. For the absolute value AS1 of the slide velocity, an interval average value AS1_m and an interval standard deviation AS1_s are calculated as feature quantities. For the absolute value AS2 of the slide acceleration, an interval average value AS2_m and an interval standard deviation AS2_s are calculated as feature quantities.

[0057] From the interval average values ​​D0_m, D1_m, D2_m and the interval standard deviations D0_s, D1_s, D2_s of each interval, interval representative average values ​​D0_mm, D1_mm, D2_mm, D0_sm, D1_sm, D2_sm are calculated as features, respectively, and interval representative standard deviations D0_ms, D1_ms, D2_ms, D0_ss, D1_ss, D2_ss are calculated as features, respectively.

[0058] From the interval average values ​​S0_m, S1_m, S2_m and the interval standard deviations S0_s, S1_s, S2_s of each interval, interval representative average values ​​S0_mm, S1_mm, S2_mm, S0_sm, S1_sm, S2_sm are calculated as feature quantities, respectively, and interval representative standard deviations S0_ms, S1_ms, S2_ms, S0_ss, S1_ss, S2_ss are calculated as feature quantities, respectively.

[0059] From the interval average values ​​AS0_m, AS1_m, AS2_m and the interval standard deviations AS0_s, AS1_s, AS2_s of each interval, interval representative average values ​​AS0_mm, AS1_mm, AS2_mm, AS0_sm, AS1_sm, AS2_sm are calculated as features, respectively, and interval representative standard deviations AS0_ms, AS1_ms, AS2_ms, AS0_ss, AS1_ss, AS2_ss are calculated as features, respectively.

[0060] In the stereoscopic cognitive ability assessment system 100, the stereoscopic cognitive ability of the subject Sb1 is assessed using at least one of a feature amount (first feature amount) derived from a value obtained by time-differentiating the inter-vehicle distance at least once and a feature amount (second feature amount) derived from the amount of sliding. That is, in this assessment, at least one of the feature amounts included in the area Rg surrounded by the dotted line in FIG. 8 is used.

[0061] In the following, the relationship between the amount of sliding and the stereoscopic perception ability will be described with reference to FIG. 9 and FIG. 10. FIG. 9 shows the trajectory Tra1 of the preceding vehicle Ca1 and the trajectory Trs1 of the vehicle Cs1 in FIG. 5 when the vehicle Cs1 is driven by a driver labeled with a label indicating high stereoscopic perception ability. FIG. 10 shows the trajectory Tra2 of the preceding vehicle Ca1 and the trajectory Trs2 of the vehicle Cs1 in FIG. 5 when the vehicle Cs1 is driven by a driver labeled with a label indicating low stereoscopic perception ability. As shown in FIG. 9, the trajectory Trs1 changes so as to follow the trajectory Tra1. On the other hand, as shown in FIG. 10, the trajectory Trs2 often departs from the trajectory Tra2 and meanders. A driver with relatively low stereoscopic perception ability tends to have a lower ability to maintain the traveling direction of the vehicle than a driver with relatively high stereoscopic perception ability. Therefore, the feature derived from the slide amount is effective for evaluating the stereoscopic perception ability of the driver of the vehicle (subject) as a feature reflecting the ability to maintain the vehicle's heading in that direction.

[0062] A specific example of the decision tree DT in FIG. 2 will be described below with reference to FIGS. 11 to 17. FIG. 11 is an example of the decision tree DT in FIG. 2, and is a diagram showing a decision tree DT1 including a feature amount derived from the interval acceleration D2 as an explanatory variable. As shown in FIG. 11, the decision tree DT1 includes a root node Rn10, a branch node Bn11, and leaf nodes Ln11, Ln12, and Ln13. Of the 19 drivers (traveling data as learning data) to be classified in the root node Rn10, 14 drivers are given a label ("high") indicating that they have high stereoscopic perception ability (objective variable), and 5 drivers are given a label ("low") indicating that they have low stereoscopic perception ability.

[0063] The root node Rn10 includes a feature D2_ms (a section representative standard deviation of the section average value of interval acceleration D2) as an explanatory variable. If the feature D2_ms of a driver is equal to or less than 0.417, the driver is classified into the leaf node Ln11. If the feature D2_ms of a driver is greater than 0.417, the driver is classified into the branch node Bn11. Of the 19 drivers to be classified in the root node Rn10, 13 drivers are classified into the leaf node Ln11, and 6 drivers are classified into the branch node Bn11. All drivers classified into the leaf node Ln11 are labeled as "high." Of the 6 drivers classified into the branch node Bn11, 1 driver is labeled as "high" and 5 drivers are labeled as "low."

[0064] Branch node Bn11 includes feature D2_ss (section representative standard deviation of section standard deviation of interval acceleration D2) as an explanatory variable. If a driver's feature D2_ss is less than or equal to 1.594, the driver is classified into leaf node Ln12. If a driver's feature D2_ss is greater than 1.594, the driver is classified into leaf node Ln13. Of the six drivers to be classified in branch node Bn11, five drivers are classified into leaf node Ln12 and one driver is classified into leaf node Ln13. All drivers classified into leaf node Ln12 are labeled "low." All drivers classified into leaf node Ln13 are labeled "high."

[0065] As described above, the decision tree DT1 includes the features D2_ms and D2_ss as explanatory variables. The decision tree DT1 classifies subjects with high stereoscopic ability into leaf nodes Ln11 and Ln13, and subjects with low stereoscopic ability into leaf node Ln12.

[0066] The Gini coefficient shown in each node of the decision tree DT1 is an index value indicating the degree to which multiple types are mixed in the labels attached to the driver corresponding to the node (impurity). The larger the Gini coefficient, the higher the degree to which multiple types are mixed in the labels attached to the driver. For example, the type of label attached to the drivers classified in each of the leaf nodes Ln11 to Ln13 is one type, and no other labels are mixed in, so the Gini coefficient of the leaf node is the minimum value of 0. On the other hand, for example, the label attached to the driver classified in the branch node Bn11 is a mixture of "high" and "low" labels, so the Gini coefficient is greater than 0.

[0067] Fig. 12 is a diagram showing the classification result when a driver with low stereoscopic perception ability is added to the classification target of the decision tree DT1 in Fig. 11. As shown in Fig. 12, two drivers labeled with "low" are added to the classification target of the root node Rn10 in Fig. 11. The driver is classified into the leaf node Ln12 as in Fig. 11, via the branch node Bn11 from the root node Rn10. Since the decision tree DT1 has versatility for driving data other than the learning data, it is possible to determine the stereoscopic perception ability of the subject with high accuracy.

[0068] Fig. 13 is another example of the decision tree DT in Fig. 2, and shows a decision tree DT2 including, as explanatory variables, feature quantities derived from the absolute value AS0 of the slide amount S0 and the absolute value AS1 of the slide speed S1. As shown in Fig. 13, the decision tree DT2 includes a root node Rn20, branch nodes Bn21, Bn22, and Bn23, and leaf nodes Ln21, Ln22, Ln23, Ln24, and Ln25. Of the 21 drivers to be classified in the root node Rn20, 14 drivers are labeled as "high" and 7 drivers are labeled as "low."

[0069] The root node Rn20 includes a feature AS0_ss (an interval representative standard deviation of an interval standard deviation of the absolute value AS0 of the slide amount S0) as an explanatory variable. If the feature AS0_ss of a driver is equal to or less than 0.093, the driver is classified into the branch node Bn21. If the feature AS0_ss of a driver is greater than 0.093, the driver is classified into the leaf node Ln21. Of the 21 drivers to be classified in the root node Rn20, 16 drivers are classified into the branch node Bn21, and 5 drivers are classified into the leaf node Ln21. All drivers classified into the leaf node Ln21 are labeled "low". Of the 16 drivers classified into the branch node Bn21, 14 drivers are labeled "high" and 2 drivers are labeled "low".

[0070] Branch node Bn21 includes age as an explanatory variable. If the driver's age is 60 or less, the driver is classified into branch node Bn22. If the driver's age is more than 60, the driver is classified into leaf node Ln22. Of the 16 drivers to be classified in branch node Bn21, 15 drivers are classified into branch node Bn22 and one driver is classified into leaf node Ln22. All drivers classified into leaf node Ln22 are labeled "low." Of the 15 drivers classified into branch node Bn22, 14 drivers are labeled "high" and one driver is labeled "low."

[0071] Branch node Bn22 includes feature AS0_ms (interval representative standard deviation of interval average value of absolute value AS0 of slide amount S0) as an explanatory variable. If the feature AS0_ms of a driver is 0.206 or less, the driver is classified into leaf node Ln23. If the feature AS0_ms of a driver is greater than 0.206, the driver is classified into branch node Bn23. Of the 15 drivers to be classified in branch node Bn22, 13 drivers are classified into leaf node Ln23, and two drivers are classified into branch node Bn23. All drivers classified into leaf node Ln23 are labeled "high". Of the two drivers classified into branch node Bn23, one driver is labeled "high" and one driver is labeled "low".

[0072] Branch node Bn23 includes feature AS1_sm (a representative average value of the section standard deviation of the absolute value AS1 of the slide speed S1) as an explanatory variable. If the feature AS1_sm of a driver is equal to or less than 0.001, the driver is classified into leaf node Ln24. If the feature AS1_sm of a driver is greater than 0.001, the driver is classified into leaf node Ln25. Of the two drivers to be classified in branch node Bn23, one driver is classified into leaf node Ln24 and one driver is classified into branch node Bn25. All drivers classified into leaf node Ln24 are labeled as "low". All drivers classified into leaf node Ln25 are labeled as "low".

[0073] As described above, the decision tree DT2 includes the features AS0_ss, AS0_ms, and AS1_sm, and age as explanatory variables. The decision tree DT2 classifies subjects with high stereocognition ability into leaf nodes Ln23 and Ln25, and subjects with low stereocognition ability into leaf nodes Ln21, Ln22, and Ln24.

[0074] Fig. 14 is another example of the decision tree DT in Fig. 2, and shows a decision tree DT3 including, as explanatory variables, feature quantities derived from the absolute value AS0 of the slide amount S0, the absolute value AS1 of the slide speed S1, and the absolute value AS2 of the slide acceleration S2. As shown in Fig. 14, the decision tree DT3 includes a root node Rn30, branch nodes Bn31 and Bn32, and leaf nodes Ln31, Ln32, Ln33, and Ln34. Of the 21 drivers to be classified in the root node Rn30, 14 drivers are labeled as "high" and 7 drivers are labeled as "low."

[0075] The root node Rn30 includes a feature AS0_ss (a section representative standard deviation of a section standard deviation of the absolute value AS0 of the slide amount S0) as an explanatory variable. If the feature AS0_ss of a driver is equal to or less than 0.093, the driver is classified into the branch node Bn31. If the feature AS0_ss of a driver is greater than 0.093, the driver is classified into the leaf node Ln31. Of the 21 drivers to be classified in the root node Rn30, 16 drivers are classified into the branch node Bn31, and 5 drivers are classified into the leaf node Ln31. All drivers classified into the leaf node Ln31 are labeled "low". Of the 16 drivers classified into the branch node Bn31, 14 drivers are labeled "high" and 2 drivers are labeled "low".

[0076] Branch node Bn31 includes feature AS2_mm (a section representative average value of the section average value of the absolute value AS2 of the slide acceleration S2) as an explanatory variable. If the feature AS2_mm of a driver is equal to or less than 0.0, the driver is classified into leaf node Ln32. If the feature AS2_mm of a driver is greater than 0.0, the driver is classified into branch node Bn32. Of the 16 drivers to be classified in branch node Bn31, 12 drivers are classified into leaf node Ln32 and 4 drivers are classified into branch node Bn32. All drivers classified into leaf node Ln32 are labeled "high". Of the 4 drivers classified into branch node Bn32, 2 drivers are labeled "high" and 2 drivers are labeled "low".

[0077] Branch node Bn33 includes feature AS1_mm (a section representative average value of the section average value of the absolute value AS1 of the slide speed S1) as an explanatory variable. If the feature AS1_mm of a driver is equal to or less than 0.0, the driver is classified into leaf node Ln33. If the feature AS1_mm of a driver is greater than 0.0, the driver is classified into leaf node Ln34. Of the four drivers to be classified in branch node Bn32, two drivers are classified into leaf node Ln33 and two drivers are classified into leaf node Ln34. All drivers classified into leaf node Ln33 are labeled as "high." All drivers classified into leaf node Ln34 are labeled as "low."

[0078] As described above, the decision tree DT3 includes the features AS0_ss, AS2_mm, and AS1_mm as explanatory variables. The decision tree DT3 classifies subjects with high stereocognition ability into leaf nodes Ln32 and Ln33, and subjects with low stereocognition ability into leaf nodes Ln31 and Ln34.

[0079] Fig. 15 is another example of the decision tree DT in Fig. 2, and shows a decision tree DT4 including, as explanatory variables, feature quantities derived from the absolute value AS0 of the slide amount S0, the absolute value AS1 of the slide speed S1, and the absolute value AS2 of the slide acceleration S2. As shown in Fig. 15, the decision tree DT4 includes a root node Rn40, branch nodes Bn41 and Bn42, and leaf nodes Ln41, Ln42, Ln43, and Ln44. Of the 21 drivers to be classified in the root node Rn40, 14 drivers are labeled as "high" and 7 drivers are labeled as "low."

[0080] The route node Rn40 includes a feature AS2_ms (a section representative standard deviation of a section average value of the absolute value AS2 of the slide acceleration S2) as an explanatory variable. When the feature AS2_ms of a driver is equal to or less than 0.0, the driver is classified into a branch node Bn41. When the feature AS2_ms of a driver is greater than 0.0, the driver is classified into a branch node Bn42. Of the 21 drivers to be classified in the route node Rn40, 14 drivers are classified into the branch node Bn41, and 7 drivers are classified into the branch node Bn42. Of the 14 drivers classified into the branch node Bn41, 13 drivers are labeled as "high" and one driver is labeled as "low". Of the 7 drivers classified into the branch node Bn42, one driver is labeled as "high" and six drivers are labeled as "low".

[0081] Branch node Bn41 includes feature AS1_ms (interval representative standard deviation of the interval average value of absolute value AS1 of slide speed S1) as an explanatory variable. If a driver's feature AS1_ms is less than or equal to 0.001, the driver is classified into leaf node Ln41. If a driver's feature AS1_ms is greater than 0.001, the driver is classified into leaf node Ln42. All drivers classified into leaf node Ln41 are labeled as "high." All drivers classified into leaf node Ln42 are labeled as "low."

[0082] Branch node Bn42 includes feature AS0_mm (interval representative average value of interval average values ​​of absolute value AS0 of slide amount S0) as an explanatory variable. If a driver's feature AS0_mm is 0.467 or less, the driver is classified into leaf node Ln43. If a driver's feature AS0_mm is greater than 0.467, the driver is classified into leaf node Ln44. All drivers classified into leaf node Ln43 are labeled as "high." All drivers classified into leaf node Ln44 are labeled as "low."

[0083] As described above, the decision tree DT4 includes the features AS2_ms, AS1_ms, and AS0_mm as explanatory variables. The decision tree DT4 classifies subjects with high stereoscopic ability into leaf nodes Ln41 and Ln43, and subjects with low stereoscopic ability into leaf nodes Ln42 and Ln44.

[0084] Fig. 16 is another example of the decision tree DT in Fig. 2, and shows a decision tree DT5 including a feature amount derived from the slide amount S0 as an explanatory variable. As shown in Fig. 16, the decision tree DT5 includes a root node Rn50, branch nodes Bn51, Bn52, Bn53, and Bn54, and leaf nodes Ln51, Ln52, Ln53, Ln54, Ln55, and Ln56. Of the 21 drivers to be classified in the root node Rn50, 14 drivers are labeled as "high" and 7 drivers are labeled as "low."

[0085] The root node Rn50 includes a feature S0_s (interval standard deviation of the slide amount S0) as an explanatory variable. If the feature S0_s of a driver is equal to or less than 0.314, the driver is classified into the branch node Bn51. If the feature S0_s of a driver is greater than 0.314, the driver is classified into the leaf node Ln51. Of the 21 drivers to be classified in the root node Rn50, 17 drivers are classified into the branch node Bn51 and 4 drivers are classified into the leaf node Ln51. All drivers classified into the leaf node Ln51 are labeled "low". Of the 17 drivers classified into the branch node Bn51, 14 drivers are labeled "high" and 3 drivers are labeled "low".

[0086] The branch node Bn51 includes the feature S0_s as an explanatory variable. If the feature S0_s of a driver is equal to or less than 0.156, the driver is classified into the leaf node Ln52. If the feature S0_s of a driver is greater than 0.156, the driver is classified into the branch node Bn52. Of the 17 drivers to be classified in the branch node Bn51, 10 drivers are classified into the leaf node Ln52 and 7 drivers are classified into the branch node Bn52. All the drivers classified into the leaf node Ln52 are labeled as "high". Of the 7 drivers classified into the branch node Bn52, 4 drivers are labeled as "high" and 3 drivers are labeled as "low".

[0087] The branch node Bn52 includes the feature S0_s as an explanatory variable. If the feature S0_s of a driver is equal to or less than 0.173, the driver is classified into the leaf node Ln53. If the feature S0_s of a driver is greater than 0.173, the driver is classified into the branch node Bn53. Of the seven drivers to be classified in the branch node Bn52, two drivers are classified into the leaf node Ln53, and five drivers are classified into the branch node Bn53. All drivers classified into the leaf node Ln53 are labeled "low". Of the five drivers classified into the branch node Bn53, four drivers are labeled "high" and one driver is labeled "low".

[0088] Branch node Bn53 includes feature S0_m as an explanatory variable. If the feature S0_m of a driver is equal to or less than 0.613, the driver is classified into leaf node Ln54. If the feature S0_m of a driver is greater than 0.613, the driver is classified into branch node Bn54. Of the five drivers to be classified in branch node Bn53, three drivers are classified into leaf node Ln54 and two drivers are classified into branch node Bn54. All drivers classified into leaf node Ln54 are labeled "high". Of the two drivers classified into branch node Bn54, one driver is labeled "high" and one driver is labeled "low".

[0089] Branch node Bn54 includes feature S0_m as an explanatory variable. If the feature S0_m of a driver is equal to or less than 0.853, the driver is classified into leaf node Ln55. If the feature S0_m of a driver is greater than 0.853, the driver is classified into leaf node Ln56. Of two drivers to be classified in branch node Bn54, one driver is classified into leaf node Ln55 and one driver is classified into leaf node Ln56. All drivers classified into leaf node Ln55 are labeled as "low." All drivers classified into leaf node Ln56 are labeled as "high."

[0090] As described above, the decision tree DT5 includes the features S0_s and S0_m as explanatory variables. The decision tree DT5 classifies subjects with high stereocognition ability into leaf nodes Ln52, Ln54, and Ln56, and subjects with low stereocognition ability into leaf nodes Ln51, Ln53, and Ln55.

[0091] Fig. 17 is another example of the decision tree DT in Fig. 2, and shows a decision tree DT6 including a feature amount derived from the slide amount S0 as an explanatory variable. As shown in Fig. 17, the decision tree DT6 includes a root node Rn60, branch nodes Bn61 and Bn62, and leaf nodes Ln61, Ln62, Ln63, and Ln64. Of the 21 drivers to be classified in the root node Rn60, 14 drivers are labeled as "high" and 7 drivers are labeled as "low."

[0092] The root node Rn60 includes the feature S0_s as an explanatory variable. If the feature S0_s of a driver is equal to or less than 0.314, the driver is classified into the branch node Bn61. If the feature S0_s of a driver is greater than 0.314, the driver is classified into the leaf node Ln61. Of the 21 drivers to be classified in the root node Rn60, 17 drivers are classified into the branch node Bn61, and 4 drivers are classified into the leaf node Ln61. All drivers classified into the leaf node Ln61 are labeled "low". Of the 17 drivers classified into the branch node Bn61, 14 drivers are labeled "high" and 3 drivers are labeled "low".

[0093] Branch node Bn61 includes age as an explanatory variable. If the driver's age is 47 or less, the driver is classified into leaf node Ln62. If the driver's age is more than 47, the driver is classified into branch node Bn62. Of the 17 drivers to be classified in branch node Bn61, 12 drivers are classified into leaf node Ln62 and 5 drivers are classified into branch node Bn62. All drivers classified into leaf node Ln62 are labeled "high". Of the 5 drivers classified into branch node Bn62, 2 drivers are labeled "high" and 3 drivers are labeled "low".

[0094] Branch node Bn62 includes feature S0_s as an explanatory variable. If the feature S0_s of a driver is less than or equal to 0.134, the driver is classified into leaf node Ln63. If the feature S0_s of a driver is greater than 0.134, the driver is classified into leaf node Ln64. Of the five drivers to be classified in branch node Bn62, two drivers are classified into leaf node Ln63 and three drivers are classified into leaf node Ln64. All drivers classified into leaf node Ln63 are labeled "high". All drivers classified into leaf node Ln64 are labeled "low".

[0095] As described above, the decision tree DT6 includes the feature S0_s and age as explanatory variables. The decision tree DT6 classifies subjects with high stereocognition ability into leaf nodes Ln62 and Ln63, and subjects with low stereocognition ability into leaf nodes Ln61 and Ln64.

[0096] Comparing the decision tree DT5 in FIG. 16 with the decision tree DT6 in FIG. 17, the conditions regarding the explanatory variables included in the root node and the classification targets are the same. Meanwhile, regarding the explanatory variables, the explanatory variables of the decision tree DT5 do not include information regarding the attributes of the subjects, whereas the explanatory variables of the decision tree DT6 include information regarding the attributes of the subjects. As a result, the number of branchings (number of root nodes and branching nodes) is five in the decision tree DT5, whereas it is three in the decision tree DT6. In this way, by including information regarding the attributes of the subjects in the explanatory variables, it becomes possible to generate a decision tree with an efficient branching structure.

[0097] Fig. 18 is a flowchart showing the flow of processing in the stereocognition ability evaluation method performed by the processor 101 that executes the stereocognition ability evaluation program 103a in Fig. 2. In the following, each step will be simply abbreviated as S.

[0098] 18, in S101, the processor 101 selects the subject Sb1 as the driver of the vehicle Cs1, and proceeds to S102. In S102, the processor 101 reads calculation data (e.g., feature amounts derived from previous driving data) related to the subject Sb1 from the storage 103, and proceeds to S103. In S103, the processor 101 acquires the current driving data from the drive recorder 900, and proceeds to S105. In S105, the processor 101 calculates feature amounts related to the amount of sliding, and proceeds to S106. In S106, the processor 101 determines whether or not a preceding vehicle is present.

[0099] If there is no preceding vehicle (NO in S106), the processor 101 determines whether the driver has been changed in S107. If the driver has not been changed (NO in S107), the processor 101 returns the process to S102 and writes the calculation data including the feature amount calculated in S105 to the storage 103. If the driver has been changed (YES in S107), the processor 101 writes the calculation data including the feature amount calculated in S105 to the storage 103 in S108 and proceeds to S109. In S109, the processor 101 evaluates the stereoscopic perception ability of the subject Sb1 based on the feature amount related to the slide amount using the evaluation model 103c, and proceeds to S114.

[0100] If a preceding vehicle exists (YES in S106), the processor 101 calculates a feature quantity related to the following distance in S110, and proceeds to processing in S111. The processor 101 determines whether the driver has been changed in S111. If the driver has not been changed (NO in S111), the processor 101 returns the processing to S102, and writes the calculation data including the feature quantities calculated in each of S105 and S110 to the storage 103. If the driver has been changed (YES in S111), the processor 101 writes the calculation data including the feature quantities calculated in each of S105 and S110 to the storage 103 in S112, and proceeds to processing in S113. The processor 101 evaluates the stereoscopic perception ability of the subject Sb1 based on the feature quantities related to the sliding amount and the following distance using the evaluation model 103c in S113, and proceeds to processing in S114. In S114, the processor 101 transmits the evaluation result of the stereoscopic perception ability to the terminal device of the subject Sb1, and ends the process.

[0101] According to the stereoscopic cognitive ability assessment system 100, the trained assessment model 103c generated by machine learning can be used to objectively assess the stereoscopic cognitive ability of the subject based on the quantified feature values ​​(see FIG. 8).

[0102] [Modification of the first embodiment] In the first embodiment, a configuration has been described in which the information processing device 110 has both a function of evaluating stereoscopic perception ability and a learning function of learning an evaluation model. In a modification of the first embodiment, a configuration will be described in which a device having a function of evaluating stereoscopic perception ability and a device having a learning function are separate devices.

[0103] FIG. 19 is a block diagram showing a configuration of a stereoscopic cognitive ability evaluation system 100A according to a modified example of the first embodiment. In the configuration of the stereoscopic cognitive ability evaluation system 100A, the information processing device 110 in FIG. 1 is replaced with 110A, and a learning device 600 is added. The configuration of the stereoscopic cognitive ability evaluation system 100A other than these is the same as the configuration of the stereoscopic cognitive ability evaluation system 100 in FIG. 1, so the description of the configuration will not be repeated. As shown in FIG. 19, the information processing device 110A, the learning device 600, the terminal device 800, the automobile Cs1, and the drive recorder 900 are connected to each other via a network NW.

[0104] Fig. 20 is a diagram showing a hardware configuration of the stereoscopic perception ability evaluation system 100A of Fig. 19. The hardware configuration of the information processing device 110A is a configuration in which the machine learning program 103b and the learning data 103d are removed from the storage 103 of Fig. 2. The hardware configuration of the information processing device 110A other than these is similar to the hardware configuration of the information processing device 110 of Fig. 2, and therefore description of the hardware configuration will not be repeated.

[0105] 20, the learning device 600 includes a processor 601, a RAM 602, a storage 603, a communication unit 604, a memory interface 605, and a housing Hs11. The housing Hs11 houses the processor 601, the RAM 602, the storage 603, the communication unit 604, and the memory interface 605. The processor 601, the RAM 602, the storage 603, the communication unit 604, and the memory interface 605 have similar functions to the processor 101, the RAM 102, the storage 103, the communication unit 104, and the memory interface 105 in FIG. 2, respectively, and therefore the description of the similar functions of the processor 601, the RAM 602, the storage 603, the communication unit 604, and the memory interface 605 will not be repeated.

[0106] The storage 603 stores a machine learning program 103b, an evaluation model 103c, and training data 103d. The processor 601 executes the machine learning program 103b to generate a trained evaluation model 103c. The processor 601 provides the trained evaluation model 103c to the information processing device 110A via the communication unit 604.

[0107] As described above, the stereoscopic cognitive ability assessment system, stereoscopic cognitive ability assessment device, stereoscopic cognitive ability assessment program, and stereoscopic cognitive ability assessment method according to embodiment 1 and its modified examples make it possible to realize an objective assessment of stereoscopic cognitive ability.

[0108] [Embodiment 2] In the first embodiment and the modified example, a configuration is described in which the stereoscopic cognitive ability of a subject is evaluated based on a change in the position of a moving object operated by the subject in the real world. In the second embodiment, a configuration is described in which the stereoscopic cognitive ability of a subject is evaluated based on a change in the position of a moving object operated by the subject in virtual reality (virtual space). Note that in the second embodiment, as in the first embodiment, a case is described in which the moving object is a car, but the moving object is not limited to a car. As long as the position of the moving object changes in response to the subject's operation, the moving object may be any object, such as a motorcycle, bicycle, airplane, or ship.

[0109] FIG. 21 is a diagram showing a state in which the stereoscopic cognitive ability of the subject Sb2 is measured by a measurement test using a driving simulation in the stereoscopic cognitive ability evaluation system 200 according to the second embodiment. As shown in FIG. 21, the stereoscopic cognitive ability evaluation system 200 includes a stereoscopic cognitive ability evaluation device 210 and a control device Cd. The stereoscopic cognitive ability evaluation device 210 and the control device Cd are connected to each other wirelessly or by wire. The control device Cd includes a steering wheel Hd, an accelerator pedal Ap, and a brake pedal Bp. The stereoscopic cognitive ability evaluation system 200 evaluates the stereoscopic cognitive ability of the subject Sb2 using at least one of a feature amount derived from a value obtained by time-differentiating the inter-vehicle distance at least once and a feature amount derived from the slide amount, as in the stereoscopic cognitive ability evaluation system according to the first embodiment.

[0110] FIG. 22 is a diagram showing an example of a screen displayed to the subject Sb2 via the stereoscopic cognitive ability evaluation device 210 in a driving simulation by the stereoscopic cognitive ability evaluation system 200 in FIG. 21. As shown in FIG. 22, a view from the driver's seat of the automobile Cs2 (moving object) driven by the subject Sb2 in a virtual space is displayed to the subject Sb2. A preceding automobile Ca2 (reference object) traveling in front of the automobile Cs2 is displayed in the view. The automobile Cs2 moves in the virtual space according to the input of the subject Sb2 to the control device in FIG. 21. Note that, when the stereoscopic cognitive ability of the subject Sb2 is evaluated using a feature amount related to the sliding amount of the automobile Cs2, the preceding automobile Ca2 does not need to be displayed.

[0111] Fig. 23 is an external perspective view of the stereocognitive ability assessment device 210 of Fig. 21. In Fig. 23, the configuration shown by the dashed lines is housed inside the housing Hs2 of the stereocognitive ability assessment device 210 and cannot be seen from the outside. Details of the configuration will be described later with reference to Fig. 24. The stereocognitive ability assessment device 210 evaluates the stereocognitive ability of the subject by having the subject visually recognize a moving object and evaluating the subject's reaction thereto.

[0112] The stereocognitive ability assessment device 210 is typically a head-mounted display (goggles) equipped with an electronic display that displays moving images representing a three-dimensional virtual reality, and takes the form of a virtual reality headset. A wearing band Rb, typically a rubber band, is attached to the stereocognitive ability assessment device 210. The user (subject) places the stereocognitive ability assessment device 210 so as to cover the area around the eyes and wraps the wearing band Rb around the head, thereby wearing the stereocognitive ability assessment device 210 around the eyes.

[0113] Fig. 24 is a diagram showing a hardware configuration of the stereoscopic perception ability assessment device 210 of Fig. 23. As shown in Fig. 24, the stereoscopic perception ability assessment device 210 includes a processor 201, a RAM 202, a storage 203, a communication unit 204, a memory interface 205, an electronic display 211, and a gaze / pupil sensor 212. The processor 201, the RAM 202, the storage 203, the communication unit 204, and the memory interface 205 have similar functions to the processor 101, the RAM 102, the storage 103, the communication unit 104, and the memory interface 105 of Fig. 2, respectively, and therefore the description of the similar functions of the processor 201, the RAM 202, the storage 203, the communication unit 204, and the memory interface 205 will not be repeated.

[0114] The storage 203 stores an OS (Operating System) program (not shown), a stereoscopic perception ability evaluation program 203a, a machine learning program 203b, an evaluation model 203c (specific model), and learning data 203d.

[0115] The electronic display 211 includes a flat panel display such as an LCD (Liquid Crystal Display) or an organic EL (Electro Luminescence) display. The electronic display 211 displays a moving image of a moving object moving in a virtual space to a subject wearing the stereocognition ability assessment device 210 around the eyes through an eyepiece arranged on the subject side. When data of the moving image is transferred from the processor 201 to a data buffer area of ​​the electronic display 211, the electronic display 211 reads the image data from the data buffer area and displays the moving image represented by the data. The electronic display 211 is independent for the right eye and the left eye, and each is visually recognized by the user through an eyepiece. When an object displayed on the electronic display 211 is located at infinity, it is displayed at the same position on the electronic displays 211 for the right eye and the left eye. As a result, no parallax occurs between the left and right eyes, and the left and right eyes are in a divergent state, giving the subject a sense of being at infinity. As the object moves closer to the user, it is displayed inward on the right-eye electronic display 211 and the left-eye electronic display 211. As a result, parallax occurs between the left and right eyes, and the left and right eyes enter a convergence state, giving the subject the sensation that the object is nearby.

[0116] The gaze / pupil sensor 212 detects the gaze direction and pupil size of each of the left and right eyes, which are arranged on the upper side of the electronic display 211 or the like toward the subject. The gaze / pupil sensor 212 acquires images of each of the left and right eyes using an image acquisition means such as a camera, and determines the gaze direction and pupil size by identifying the position of the pupil and the size of the pupil in the image, and outputs them to the processor 201 as gaze information. A visible light camera or an infrared camera can be used as the camera. In order to determine whether the subject is viewing a moving object, the gaze direction is important data. The visibility of the moving object can be confirmed by confirming that the gazes of the left and right eyes (normal lines to the center of the pupil) are accurately passing through the moving object. At this time, if a nearby moving object is viewed, the gazes of the left and right eyes will move inward due to parallax, resulting in a convergence state. In addition, the pupil diameter can be additionally used to determine whether the subject is continuously viewing an approaching moving object. When the subject is continuously viewing an approaching moving object, the pupil diameter gradually decreases due to the pupillary short-distance reflex, and by detecting this, it is possible to confirm whether or not the object has been viewed successfully.

[0117] Stereocognitive ability is based on the assumption that the eyes have normal functions (e.g., pupil accommodation or eye movement) to obtain visual information. Therefore, stereocognitive ability is defined as the ability to accurately grasp the positional relationship of a moving object from the visual information of the moving object when the subject sees it, and to appropriately and accurately perform an action corresponding to the moving object based on the grasped positional relationship.

[0118] As a method for checking the normal function of the eye, for example, pupil measurement and near point distance measurement are known. These measurements can be performed using a device called TriIris. Pupil measurement includes at least measuring the pupil reaction (light reaction) to a visible light stimulus from a visual target, and measuring the pupil change when looking at a moving visual target. Specifically, the pupil change occurs due to the pupil near reflex, and the pupil contracts when the visual target moves closer. In addition, in the near point distance measurement, the subject presses a hand switch when the visual target approaching at a constant refraction speed becomes blurred during observation, and the position of the visual target at that time is recorded as the near point distance. These measurements are the measurement of the near point distance based on the state of the eyeball (pupil) and the pressing of the switch at the time when it becomes blurred, but the main purpose is to measure the near point distance.

[0119] The stereocognitive ability includes the ability to accurately track a moving object with the eyes, correctly recognize its position, and respond appropriately to it. In order to measure the ability to track a moving object with the eyes, a method of measuring pupil accommodation and convergence reaction is considered to be particularly useful. These measurement methods are described below. FIG. 25 is a diagram showing the concept of the visual target used to measure eye function. The visual target (the preceding vehicle Ca2) moves between a direction moving away and a direction moving closer so that it is visually recognized in front of the subject's eyes. It is desirable to repeat the indicator and check the state of the subject's eyes each time.

[0120] FIG. 26 is a diagram showing the relationship between the target distance (the distance between the subject's eye and the target) and pupil diameter due to the pupil near reflex. FIG. 26(A) shows that the pupil diameter becomes smaller when the target distance is close. FIG. 26(B) shows that the pupil diameter becomes larger when the target distance is far. FIG. 26(C) shows a graph with the target distance on the horizontal axis and the pupil diameter on the vertical axis. The solid line shows the graph for the left eye, and the dashed line shows the graph for the right eye. These graphs show that the pupil diameter becomes smaller when the target distance is close, and the pupil diameter becomes larger when the target distance is far. It also shows that the pupil diameter of the right eye and the left eye is almost the same regardless of the target distance.

[0121] FIG. 27 is a diagram showing the relationship between the target distance and the pupil position (convergence and divergence movement). FIG. 27(A) shows that when the target distance is close, the left and right eyes are in a convergent state with the eyes moving inwards. FIG. 27(B) shows that when the target distance is far, the left and right eyes are in a divergent state with the eyes parallel to each other. FIG. 27(C) shows a graph with the target distance on the horizontal axis and the pupil position on the vertical axis. The solid line shows the graph for the left eye, and the dashed line shows the graph for the right eye. These graphs show that when the target distance is close, the distance between the pupils of the left and right eyes becomes small, resulting in a convergent state, and when the target distance is far, the distance between the pupils of the left and right eyes becomes large, resulting in a divergent state.

[0122] When the indicator is moved so as to change the distance of the visual target that the subject is looking at, the subject's reaction such as the change in pupil diameter and convergence movement occurs accordingly. If the subject's visual function is impaired, these reactions are reduced. Therefore, the visual function can be measured by measuring the change in pupil diameter and convergence movement of the subject when the visual target is changed. According to the stereoscopic cognition ability evaluation system 200, since a large-scale device for changing the distance of the visual target is not required, the stereoscopic cognition ability evaluation system 200 can be made compact.

[0123] Fig. 28 is a functional block diagram showing the configuration of the stereoscopic cognition ability evaluation function of the processor 201 in Fig. 24. In the stereoscopic cognition ability evaluation system 200, the processor 201 executes the stereoscopic cognition ability evaluation program 203a stored in the storage 203 to configure modules forming functional blocks of a moving object display unit 201a, a position acquisition unit 201b, a visual recognition determination unit 201c, a reaction input unit 201d, and a stereoscopic cognition ability determination unit 201e. Therefore, in Fig. 28, instead of the processor 201 and the stereoscopic cognition ability evaluation program 203a in Fig. 24, functional blocks realized by them are shown. Below, these functional blocks will be described.

[0124] The moving object display unit 201a continuously generates images of the scenery ahead seen by the subject Sb2 from the driver's seat of the car Cs2 by three-dimensional rendering in accordance with the movement of the car Cs2, and transfers them as video data to the data buffer area of ​​the electronic display 211. The video data is data for each of the right eye and the left eye on the electronic display 211. The video data generates parallax between the video data for the right eye and the video data for the left eye in the position of the scenery seen by the subject Sb2 from the car Cs2 according to the position of the moving object. Therefore, the subject who views the video data on the electronic display 211 can see the ball with a realistic perspective. The moving object display unit 201a transmits driving data of the car Cs2 based on the video data to the position acquisition unit 201b.

[0125] The position acquisition unit 201b acquires running data of the subject Sb2 from the moving object display unit 201a. The position acquisition unit 201b acquires information about the position of the car Cs2 from the running data. The position acquisition unit 201b outputs the information about the position of the car Cs2 to the visual recognition determination unit 201c and the stereoscopic perception ability determination unit 201e.

[0126] The visibility determination unit 201c determines whether the subject Sb2 visually recognizes the vehicle Cs2 spatially by determining whether the gaze direction correctly corresponds to the position of the vehicle Cs2. The visibility determination unit 201c receives data on the gaze directions of the left and right eyes of the subject Sb2 detected by the gaze / pupil sensor 212, and determines whether the gaze directions of the left and right eyes correspond to the position of the vehicle Cs2 transmitted from the moving object display unit 201a, thereby determining whether the subject Sb2 spatially recognizes the vehicle Cs2. The visibility determination unit 201c may further receive data on the pupil diameters of the left and right eyes of the subject Sb2 detected by the gaze / pupil sensor 212. The visual recognition determination unit 201c may operate to determine that the subject spatially recognizes an object when it is further determined that the pupil diameter of both eyes is gradually decreasing while the position of the preceding vehicle Ca2 approaches a predetermined viewpoint and the distance between the subject and the predetermined viewpoint is decreasing (when the pupillary short-distance reflex occurs in response to the distance between the subject and the predetermined viewpoint is decreasing). The visual recognition determination unit 201c outputs the determination result to the stereoscopic perception ability determination unit 201e as visual recognition information.

[0127] The reaction input unit 201d receives input of the active reaction of the subject Sb2 corresponding to the three-dimensional position in the virtual space of the car Cs2 recognized by the subject Sb2. The reaction input unit 201d identifies the reaction of the subject Sb2 who is driving the car Cs2 based on the input information of the subject Sb2 to the control device Cd. The reaction input unit 201d outputs information related to the reaction to the stereoscopic perception ability assessment unit 201e.

[0128] The stereoscopic cognitive ability determination unit 201e derives a feature amount derived from at least one of the interval speed, interval acceleration, absolute value of the sliding amount, absolute value of the sliding speed, absolute value of the sliding acceleration, sliding amount, sliding speed, and sliding acceleration of the automobile Cs2 based on information about the position of the automobile Cs2. The stereoscopic cognitive ability determination unit 201e uses the evaluation model 203c to determine the stereoscopic cognitive ability of the subject Sb2 from the feature amount and the visual recognition information from the visual recognition determination unit 201c. The visual recognition information may not be used to determine the stereoscopic cognitive ability of the subject Sb2. Information about the reaction from the reaction input unit 201d may also be used to determine the stereoscopic cognitive ability of the subject Sb2. The stereoscopic cognitive ability determination unit 201e transmits the evaluation result of the stereoscopic cognitive ability to the terminal device 820 of the subject Sb2 via the communication unit 204.

[0129] In the stereoscopic cognitive ability assessment system 200, the processor 201 executes the machine learning program 203b, and the stereoscopic cognitive ability assessment device 210 functions as a learning device that generates a trained assessment model 203c. As in the modification of the first embodiment, machine learning for the assessment model 203c may be performed by a learning device separate from the stereoscopic cognitive ability assessment device 210.

[0130] Fig. 29 is a flowchart showing the flow of processing in the stereoscopic perception ability evaluation method performed by the processor 201 that executes the stereoscopic perception ability evaluation program 203a in Fig. 24. The processing shown in Fig. 29 is a flowchart in which S204 is added between S103 and S105 in the flowchart shown in Fig. 18, and S109 and S113 are replaced with S209 and S213, respectively.

[0131] 29, the processor 201 executes S101 to S103 in the same manner as in the first embodiment, and then acquires gaze information from the gaze / pupil sensor 212 in S204, and proceeds to processing at S105. The processor 201 executes S105 to S108 or S105, S106, and S110 to S112 in the same manner as in the first embodiment, and then evaluates the stereoscopic perception ability of the subject Sb2 based on the feature amount related to the sliding amount and the visual information using the evaluation model 203c in each of S209 and S213, and proceeds to processing at S114. The processor 201 executes S114 in the same manner as in the first embodiment, and then ends the processing.

[0132] As described above, the stereoscopic cognitive ability assessment system, the stereoscopic cognitive ability assessment device, the stereoscopic cognitive ability assessment program, and the stereoscopic cognitive ability assessment method according to the second embodiment can realize an objective assessment of stereoscopic cognitive ability.

[0133] [Embodiment 3] In the first embodiment, a configuration is described in which a feature amount is calculated from the position of a moving object measured by a position measurement unit of a drive recorder. In the third embodiment, a configuration is described in which a feature amount is directly measured by an acceleration sensor of a drive recorder.

[0134] FIG. 30 is a diagram showing a hardware configuration of a stereoscopic perception ability evaluation system 300 according to the third embodiment. The configuration of the information processing device 310 in FIG. 30 is a configuration in which the stereoscopic perception ability evaluation program 103a, the machine learning program 103b, the evaluation model 103c, and the learning data 103d in FIG. 2 are replaced with a stereoscopic perception ability evaluation program 303a, a machine learning program 303b, an evaluation model 303c, and learning data 303d, respectively. In addition, the drive recorder 930 in FIG. 30 is a configuration in which the position measurement unit 907 in FIG. 2 is replaced with an acceleration sensor 937. The acceleration sensor 937 measures, for example, the acceleration of the automobile Cs1 in a direction (lateral direction) perpendicular to a reference line along the traveling direction of the automobile Cs1. The acceleration sensor 937 may measure the acceleration of the automobile Cs1 in the direction of the reference line. The configuration of the stereoscopic perception ability evaluation system 300 other than these is similar to the configuration of the stereoscopic perception ability evaluation system 100, and therefore the description of the similar configuration will not be repeated.

[0135] Fig. 31 is a functional block diagram showing the configuration of the stereoscopic perception ability evaluation function of the processor 101 in Fig. 30. In the stereoscopic perception ability evaluation system 300, a module forming functional blocks, a behavior acquisition unit 301b, a reaction input unit 301d, and a stereoscopic perception ability determination unit 301e, is configured by executing a stereoscopic perception ability evaluation program 303a stored in the storage 103 by the processor 101. Therefore, in Fig. 31, instead of the processor 101 and the stereoscopic perception ability evaluation program 303a in Fig. 30, functional blocks realized by them are shown. Below, these functional blocks will be described.

[0136] 31, the behavior acquisition unit 301b acquires the travel data of the subject from the drive recorder 930 and the measurement result of the acceleration sensor 937 as information on the behavior of the vehicle Cs1. The behavior acquisition unit 301b acquires the acceleration (e.g., slide acceleration) of the vehicle Cs1 from the travel data and the measurement result. The behavior acquisition unit 301b outputs the acceleration of the vehicle Cs1 to the stereoscopic perception ability assessment unit 301e. The behavior acquisition unit 301b may acquire information on the position of the vehicle Cs1 as information on the behavior of the vehicle Cs1 from the travel data of the subject from the drive recorder 930 and the positioning result of the vehicle Cs1.

[0137] The reaction input unit 301d receives input of the subject's active reaction corresponding to the three-dimensional position of the car Cs1 recognized by the subject. The reaction input unit 301d identifies the reaction of the subject who is driving the car Cs1 based on input information of the subject to the steering wheel, accelerator pedal, brake pedal, etc. of the car Cs1. The reaction input unit 301d outputs information related to the reaction to the stereoscopic perception ability assessment unit 301e.

[0138] The stereoscopic cognitive ability assessment unit 301e assesses the subject's stereoscopic cognitive ability from the acceleration of the automobile Cs1 using the evaluation model 303c. Information on the reaction from the reaction input unit 301d may be used to assess the subject's stereoscopic cognitive ability. The stereoscopic cognitive ability assessment unit 301e transmits the evaluation result of the stereoscopic cognitive ability to the subject's terminal device 800 via the communication unit 104, similar to the stereoscopic cognitive ability assessment unit 101e of the first embodiment.

[0139] Fig. 32 is a functional block diagram showing the configuration of the machine learning function of the processor 101 in Fig. 30. In the stereoscopic perception ability evaluation system 300, a module forming functional blocks, that is, a behavior acquisition unit 311b and a learning unit 301f, is configured by executing the machine learning program 303b by the processor 101. Therefore, in Fig. 32, instead of the processor 101 and the machine learning program 303b in Fig. 30, functional blocks realized by them are shown. The functional blocks will be described below.

[0140] As shown in Fig. 32, the behavior acquisition unit 311b acquires driving data from the learning data 303d. The behavior acquisition unit 311b acquires the acceleration of the automobile corresponding to the driving data, similar to the behavior acquisition unit 301b in Fig. 31. The behavior acquisition unit 311b outputs the acceleration and the label attached to the driving data to the learning unit 301f.

[0141] The learning unit 301f performs machine learning on the evaluation model 303c using the acceleration and the label from the behavior acquisition unit 311b, and sets the evaluation model 303c as a trained model.

[0142] Table 1 below shows an example of correlation coefficients between each of the feature amounts S0_s, S1_s, and S2_s derived from the sliding acceleration acquired by the acceleration sensor 937 and each of the feature amounts S0_s and AS1_s derived from the sliding amount in the virtual space acquired by the three-dimensional perception ability evaluation system according to embodiment 2, for a plurality of subjects. Each feature amount shown in Table 1 corresponds to the feature amount shown in FIG. 8 for the same character string as the feature amount.

[0143] [Table 1]

[0144] Regarding the feature amounts derived from the slide acceleration acquired by the acceleration sensor 937, the feature amount S2_s is the interval standard deviation of the slide acceleration acquired by the acceleration sensor 937. The feature amount S1_s is the interval standard deviation of the slide speed, and is derived by integrating the feature amount S2_s once over time. The feature amount S0_s is the interval standard deviation of the slide amount, and is derived by integrating the feature amount S1_s once over time (integrating the feature amount S2_s twice over time). The slide acceleration acquired by the acceleration sensor 937 can be acquired, for example, from a drive recorder of a driving school vehicle traveling on a course at a driving school that includes S-shaped curves, cranks, straight lines, and the like.

[0145] Regarding the feature amounts derived from the sliding amounts in the virtual space, the feature amount S0_s is the section standard deviation of the sliding amounts, and the feature amount AS1_s is the section standard deviation of the absolute value of the sliding speed.

[0146] As shown in Table 1, a positive correlation is observed between each of the feature amounts S0_s, S1_s, and S2_s and each of the feature amounts S0_s and AS1_s derived from the amount of sliding acquired in the virtual space of embodiment 2. Therefore, like the feature amounts S0_s and AS1_s derived from the amount of sliding in the virtual space, the feature amounts S0_s, S1_s, and S2_s relating to the amount of sliding acquired by the acceleration sensor 937 can also be used to realize an objective evaluation of the ability to perceive three-dimensional reality.

[0147] If the distance between the reference object and the moving object is defined as a first distance, and the distance (slide amount) between the moving object and a reference line along the moving object's direction of travel is defined as a second distance, the present disclosure includes embodiments in the following range.

[0148] The present disclosure provides a system for evaluating the stereoscopic cognitive ability of a subject based on a change in behavior of a moving object that is moved by being operated by the subject. The system for evaluating the stereoscopic cognitive ability includes a behavior acquisition unit and a stereoscopic cognitive ability determination unit.

[0149] The stereoscopic cognitive ability determination unit determines the stereoscopic cognitive ability of the subject from at least one of a first feature amount related to a first distance between the moving object and a reference object preceding the moving object in the moving direction of the moving object and a second feature amount related to a second distance between the moving object and a reference line along the moving direction of the moving object. In addition, the behavior acquisition unit acquires information regarding the acceleration of the moving object.

[0150] The first feature amount and / or the second feature amount are feature amounts derived from information on the acceleration of the moving object. The first feature amount related to the first distance may be derived from any one of a value of acceleration in a first distance direction, a value obtained by integrating the acceleration in the first distance direction once over time, and a value obtained by integrating the acceleration in the first distance direction twice over time. The second feature amount related to the second distance may be derived from any one of a value of acceleration in a second distance direction, a value obtained by integrating the acceleration in the second distance direction once over time, and a value obtained by integrating the acceleration in the second distance direction twice over time.

[0151] In the third embodiment, a configuration using an acceleration sensor has been described, but a gyro sensor (angular velocity sensor) may be used instead of the acceleration sensor, or both the acceleration sensor and the gyro sensor may be used together. When a gyro sensor is used, the feature amount for determining the stereoscopic perception ability may be derived from any one of the value of the angular velocity, the value obtained by integrating the angular velocity once with respect to time, and the value obtained by integrating the angular velocity twice with respect to time.

[0152] [Modification of the third embodiment] Fig. 33 is a diagram showing a hardware configuration of a stereoscopic perception ability evaluation system 300A according to a modification of embodiment 3. A drive recorder 930 in Fig. 33 has a configuration in which the acceleration sensor 937 in Fig. 30 is added to the drive recorder 900 in Fig. 2. Other than these, the configuration of the stereoscopic perception ability evaluation system 300A is similar to the configuration of the stereoscopic perception ability evaluation system 100, and therefore description of the similar configuration will not be repeated.

[0153] As described above, the stereoscopic cognitive ability assessment system, stereoscopic cognitive ability assessment device, stereoscopic cognitive ability assessment program, and stereoscopic cognitive ability assessment method according to embodiment 3 and its modified examples can achieve an objective assessment of stereoscopic cognitive ability.

[0154] [Embodiment 4] In the modified example of the third embodiment, a configuration using data measured by an acceleration sensor or a position measuring unit of a drive recorder installed in an automobile is described, whereas in the fourth embodiment, a configuration using data measured by a camera, an acceleration sensor, or a position measuring unit included in a portable mobile terminal such as a smartphone is described. Note that the information processing device that stores the various programs may be included in the mobile terminal, or may be a device separate from the mobile terminal that communicates with the mobile terminal via a communication unit.

[0155] In the fourth embodiment, since the mobile terminal is equipped with a camera, an acceleration sensor, or a position measurement unit, the moving object is not limited to a car. The moving object may be anything that moves with the subject and whose position changes in response to the subject's operation, such as a bicycle, a motorcycle, an airplane, or a ship.

[0156] As described above, the stereoscopic perception ability assessment system, the stereoscopic perception ability assessment device, the stereoscopic perception ability assessment program, and the stereoscopic perception ability assessment method according to the fourth embodiment can realize an objective assessment of stereoscopic perception ability.

[0157] The objective evaluation of stereocognitive ability described in the first to fourth embodiments, which are examples of the present disclosure, can be utilized in various situations. For example, the evaluation can be utilized as an index of the development of higher brain functions in the process of human growth. In particular, stereocognitive ability has been conventionally evaluated by solving problems in the form of a quiz or based on empirical rules in sports and other competitions. However, by quantitatively evaluating stereocognitive ability using the stereocognitive ability evaluation system, stereocognitive ability evaluation device, stereocognitive ability evaluation program, and stereocognitive ability evaluation method described in this specification, it becomes possible to evaluate stereocognitive ability with higher accuracy.

[0158] The stereocognitive ability evaluation system, stereocognitive ability evaluation device, stereocognitive ability evaluation program, and stereocognitive ability evaluation method described herein enable continuous evaluation of stereocognitive ability during activities of daily life such as driving a car, etc. As a result, it becomes possible to objectively evaluate stereocognitive ability, which cannot be obtained by conventional evaluation methods such as quiz-style evaluation methods conducted within a certain period of time.

[0159] Another example of the application of the objective evaluation of stereocognitive ability described in this specification can be the evaluation of cognitive impairment in brain damage, mental illness, dementia, etc. In this example, the stereocognitive ability can be safely evaluated objectively by using virtual reality (virtual space) according to the subject's situation. It is also possible to confirm the recovery of stereocognitive ability by taking medication, or changes such as adverse effects.

[0160] As another example, the use of the ability required for driving a moving object such as a car can be mentioned in the objective definition. Specifically, stereoscopic cognition ability can be defined as one of the abilities. In particular, the decline in the stereoscopic cognition ability of elderly people due to aging may lead to accidents, and an objective evaluation of stereoscopic cognition ability is important in determining the aptitude of a person to drive. In addition, by objectively evaluating the stereoscopic cognition ability of subjects such as inexperienced drivers, drivers with high fatigue levels, or drivers with low attention, in addition to elderly people, the decline in the stereoscopic cognition ability of the subjects can be detected early. As a result, it is possible to prevent accidents by warning the subjects of the decline in their stereoscopic cognition ability. Note that subjects whose stereoscopic cognition ability can be objectively evaluated are not limited to drivers of cars, but include operators of moving objects that can run in straight and curved lines. Moving objects that can run in straight and curved lines include, for example, bicycles, motorcycles, airplanes, or ships. In addition, the objective evaluation of stereoscopic cognition ability can also be an indicator for the design of moving objects equipped with auxiliary functions that compensate for the decline in stereoscopic cognition ability.

[0161] The embodiments disclosed herein are intended to be combined appropriately within a range that is not inconsistent. The embodiments disclosed herein should be considered to be illustrative and not restrictive in all respects. The scope of the present disclosure is indicated by the claims, not the above description, and is intended to include all modifications within the meaning and scope of the claims. [Explanation of symbols]

[0162] 100,100A,200,300,300A Stereocognitive Ability Evaluation System, 101,201,601,901 Processor, 101b,111b,201b Position Acquisition Unit, 101d,201d,301d Response Input Unit, 101e,201e,301e Stereocognitive Ability Judgment Unit, 101f,301f Learning Unit, 102,202,602,903 RAM, 103,203,603,905 Storage, 103a,203a,303a Stereocognitive Ability Evaluation Program, 103b,203b,303b Machine Learning Program, 103c,203c,303c Evaluation Model, 103d,203d,303d Learning Data, 104,204,604,904 Communication unit, 105, 205, 605, 906 memory interface, 110, 110A, 310 information processing device, 201a moving object display unit, 201c visual recognition determination unit, 210 stereoscopic perception ability evaluation device, 211 electronic display, 212 pupil sensor, 301b, 311b behavior acquisition unit, 600 learning device, 800, 820 terminal device, 900 drive recorder, 902 camera, 905a driving data, Ap accelerator pedal, Bp brake pedal, Ca1, Ca2 preceding vehicle, Cd control device, Cs1, Cs2 automobile, DT, DT1 to DT6 decision tree, Hd steering wheel, Hs1, Hs2, Hs11 housing, NW network, Rb wearing band, Sb1, Sb2 subject.

Claims

1. A stereocognitive ability assessment system for assessing the stereocognitive ability of a subject based on a change in position of a moving object that is moved by the subject, comprising: a position acquisition unit for acquiring information regarding a position of the moving object; a stereoscopic perception ability assessment unit that assesses the stereoscopic perception ability using a trained specific model; the explanatory variables of the specific model include at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in a moving direction of the moving object, and a second feature amount derived from a second distance between the moving object and a reference line along the moving direction of the moving object, The specific model is a trained model through machine learning using training data that is information regarding the position of the moving object operated and moved by the operator, the model being pre-labeled based on the operator's stereoscopic perception ability.

2. The stereoscopic perception ability assessment system according to claim 1 , wherein the second feature amount includes a feature amount derived from a value obtained by time-differentiating the second distance at least once.

3. The stereoscopic perception ability assessment system according to claim 1 , wherein the second feature amount includes a feature amount derived from a value obtained by time-differentiating the second distance at least twice.

4. The stereoscopic perception ability assessment system according to claim 1 , wherein the first feature amount includes a feature amount derived from a value obtained by time-differentiating the first distance at least twice.

5. The stereocognitive ability assessment system according to claim 4 , wherein the specific model includes a decision tree having the stereocognitive ability as a response variable.

6. The stereocognitive ability assessment system according to claim 5 , wherein the explanatory variables further include information regarding attributes of the subject.

7. a movement recording device mounted on the moving object; The movement recording device includes a camera for generating video data when the moving object is viewed in a moving direction of the moving object; a position measurement unit for measuring a position of the moving object, The stereoscopic perception ability assessment system according to claim 1 , wherein the position acquisition unit acquires information regarding the position of the moving object from the video data including at least one of the reference object and the reference line and from the measurement result of the position measurement unit.

8. the moving object is provided by a virtual reality; The stereocognition ability evaluation system comprises: a virtual reality headset including an electronic display for displaying said virtual reality motion imagery; a moving object display unit that displays a moving image of the moving object moving in the virtual reality on the electronic display, The stereoscopic perception ability assessment system according to claim 1 , wherein the position acquisition unit acquires a position of the moving object in the virtual reality displayed by the moving object display unit.

9. A stereocognitive ability assessment system for assessing the stereocognitive ability of a subject based on a change in behavior of a moving object that is moved by being operated by the subject, comprising: A behavior acquisition unit that acquires information regarding a behavior of the moving object; a stereoscopic perception ability assessment unit that assesses the stereoscopic perception ability using a trained specific model; the explanatory variables of the specific model include at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in a moving direction of the moving object, and a second feature amount relating to a second distance between the moving object and a reference line along the moving direction of the moving object, The specific model is a trained model through machine learning using training data that is information regarding the position of the moving object operated and moved by the operator, the model being pre-labeled based on the operator's stereoscopic perception ability.

10. The behavior acquisition unit acquires information regarding a position of the moving object, The stereoscopic perception ability assessment system according to claim 9 , wherein the first feature amount and / or the second feature amount is derived from information relating to a position of the moving object.

11. The behavior acquisition unit acquires information regarding the acceleration of the moving object, The stereoscopic cognitive ability assessment system according to claim 9 , wherein the stereoscopic cognitive ability assessment unit assesses the stereoscopic cognitive ability from the second feature amount.

12. The stereoscopic perception ability assessment system according to claim 11 , wherein the second feature amount is derived from an acceleration of the moving object.

13. a movement recording device mounted on the moving object; The movement recording device includes a camera for generating video data when the moving object is viewed in a moving direction of the moving object; an acceleration sensor for measuring the acceleration of the moving object; The stereoscopic perception ability assessment system according to claim 11 , wherein the behavior acquisition unit acquires information about the acceleration of the moving object from the video data and the measurement result of the acceleration sensor.

14. A terminal device of the subject; The stereoscopic cognitive ability assessment system according to any one of claims 1 to 13, further comprising a communication unit that transmits the determination result of the stereoscopic cognitive ability determination unit to the terminal device.

15. A stereocognitive ability assessment device for assessing the stereocognitive ability of a subject based on a change in position of a moving object that is moved by the subject, the device comprising: a position acquisition unit for acquiring information regarding a position of the moving object; A stereoscopic cognitive ability assessment unit that assesses the stereoscopic cognitive ability using a trained specific model; a housing that houses the position acquisition unit and the stereoscopic perception ability assessment unit, the explanatory variables of the specific model include at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in a moving direction of the moving object, and a second feature amount derived from a second distance between the moving object and a reference line along the moving direction of the moving object, The specific model is a trained model through machine learning using training data that is information regarding the position of the moving object that is operated and moved by the operator, the training data being labeled in advance based on the operator's stereoscopic perception ability.

16. A stereocognitive ability assessment device for assessing the stereocognitive ability of a subject based on a change in behavior of a moving object that is moved by being operated by the subject, comprising: A behavior acquisition unit that acquires information regarding a behavior of the moving object; A stereoscopic cognitive ability assessment unit that assesses the stereoscopic cognitive ability using a trained specific model; a housing that houses the behavior acquisition unit and the stereoscopic perception ability assessment unit, the explanatory variables of the specific model include at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in a moving direction of the moving object, and a second feature amount relating to a second distance between the moving object and a reference line along the moving direction of the moving object, The specific model is a trained model through machine learning using training data that is information regarding the position of the moving object that is operated and moved by the operator, the training data being labeled in advance based on the operator's stereoscopic perception ability.

17. A stereocognitive ability evaluation program for causing a computer to configure a stereocognitive ability evaluation device for evaluating a subject's stereocognitive ability based on a change in position of a moving object that is moved by the subject, the stereocognitive ability evaluation device comprising: a position acquisition unit for acquiring information regarding a position of the moving object; a stereoscopic perception ability assessment unit that assesses the stereoscopic perception ability using a trained specific model; the explanatory variables of the specific model include at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in a moving direction of the moving object, and a second feature amount derived from a second distance between the moving object and a reference line along the moving direction of the moving object, The specific model is a trained model through machine learning using training data that is information about the position of the moving object operated and moved by the operator, the training data being pre-labeled based on the operator's stereoscopic perception ability.

18. A stereocognitive ability evaluation program for causing a computer to configure a stereocognitive ability evaluation device for evaluating a subject's stereocognitive ability based on a change in behavior of a moving object that is moved by being operated by the subject, the stereocognitive ability evaluation device comprising: A behavior acquisition unit that acquires information regarding a behavior of the moving object; a stereoscopic perception ability assessment unit that assesses the stereoscopic perception ability using a trained specific model; the explanatory variables of the specific model include at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in a moving direction of the moving object, and a second feature amount relating to a second distance between the moving object and a reference line along the moving direction of the moving object, The specific model is a trained model through machine learning using training data that is information about the position of the moving object operated and moved by the operator, the training data being pre-labeled based on the operator's stereoscopic perception ability.

19. A method for evaluating stereoscopic cognition ability, the method being executed by a computer to cause the computer to evaluate the stereoscopic cognition ability of a subject based on a change in position of a moving object that is moved by being operated by the subject, comprising: obtaining information regarding a position of the moving object; and determining said stereoscopic ability using a trained specific model; the explanatory variables of the specific model include at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in a moving direction of the moving object, and a second feature amount derived from a second distance between the moving object and a reference line along the moving direction of the moving object, A method for evaluating stereoscopic perception ability, in which the specific model is a trained model through machine learning using training data that is information regarding the position of the moving object operated and moved by the operator, the model being pre-labeled based on the operator's stereoscopic perception ability.

20. A method for evaluating stereoscopic cognition ability, the method being executed by a computer to cause the computer to evaluate the stereoscopic cognition ability of a subject based on a change in behavior of a moving object that is moved by being operated by the subject, comprising: obtaining information about a behavior of the moving object; and determining said stereoscopic ability using a trained specific model; the explanatory variables of the specific model include at least one of a first feature amount derived from a value obtained by time-differentiating at least once a first distance between the moving object and a reference object preceding the moving object in a moving direction of the moving object, and a second feature amount relating to a second distance between the moving object and a reference line along the moving direction of the moving object, A method for evaluating stereoscopic perception ability, in which the specific model is a trained model through machine learning using training data that is information regarding the position of the moving object operated and moved by the operator, the model being pre-labeled based on the operator's stereoscopic perception ability.

Images