Dementia Prediction Device

Abstract

To provide a dementia estimation device that can easily estimate dementia with minimal burden on the subject. [Solution] This disclosure relates to a dementia estimation device comprising a smartphone having a GPS receiver, a camera, a microphone, and an acceleration sensor; a learning data creation unit that creates learning data by machine learning the subject's route calculated from signals received by the GPS receiver, the degree of change in the subject's facial expression obtained from image data of the subject's face captured by the camera, the content of the subject's responses obtained from the subject's voice collected by the microphone, and the position of the subject's center of gravity calculated from the three-dimensional acceleration measured by the acceleration sensor; and an estimation unit that estimates whether the subject has dementia using artificial intelligence based on the learning data.

Description

Technical Field

[0001] The present disclosure relates to a dementia estimation apparatus.

Background Art

[0002] In Japan, along with the progress of aging, the population with dementia is increasing, and it is expected that about 1 in 5 people aged 65 or older will have dementia by 2025. Dementia causes numerous social issues, including family breakdown due to communication failure, violence, missing persons caused by wandering, various crimes, and traffic accidents caused by wrong-way driving and reckless driving, etc. In recent years, therapeutic drugs for dementia have been developed, but they are expensive, and even when covered by insurance, they put pressure on the national budget. Therefore, early detection and early treatment are required for dementia. Under such circumstances, there have traditionally been various tests used for the diagnosis of dementia, but in Japan, cognitive function tests such as the Revised Hasegawa Dementia Scale (HDS-R) and MMSE (Mini-Mental State Examination) are mainly used. However, the Revised Hasegawa Dementia Scale is a static means implemented in a question format asking for daily information (e.g., "What is the date today?") and carrying out situation confirmation and calculation, etc. Also, the MMSE is also a method of confirming time, place, short-term memory, etc. in a question format, and can be said to be a static means similarly to the Hasegawa scale. Incidentally, in recent years, it has been confirmed that the progression of dementia appears in dynamic factors such as behavior, facial expression, eye movement, speech and behavior, and balance. For example, it is also clear from the fact that as dementia progresses, activity itself becomes less vigorous. This is also clear from experiments on small animals such as mice. Therefore, by converting these dynamic factors into training data on a daily basis, and artificial intelligence (AI) extracting while comparing and examining their changes, it is possible to screen whether there is a suspicion of dementia at an early stage. Therefore, in the present invention, collection of these daily data is performed using basic functions installed in a portable terminal (e.g., a smartphone, etc.), and the present invention is developed for the purpose of enabling elderly people to carry out the screening easily and on a daily basis.

[0003] Patent Document 1 discloses a cognitive impairment diagnostic device that diagnoses cognitive impairment based on a distribution map created by detecting the subject's gaze over time while they are viewing a diagnostic video. Patent Document 2 discloses an information providing device that receives a first input from a first user for the start of a first service, and provides a first service recommendation to a first output unit based on the subject information, including the dementia symptoms of the person being cared for, and time information at the time the first input was received. Patent Document 3 discloses a standing posture evaluation device that evaluates the standing posture balance of a subject based on the position of the head's center of gravity detected by an overhead 3D camera and the position of the body's center of gravity detected by a body pressure sensor. Patent Document 4 discloses a dementia testing method that includes a first test, which is a practice test, and a second test, which determines whether or not a person has dementia. In the second test, if the verbal fluency value calculated based on one word extracted in response to the second theme is smaller than a standard value, the person is determined to be a candidate for a subsequent test. Patent Document 5 discloses a diagnostic support information provision device that determines the possibility of brain dysfunction based on the percentage of no change in facial orientation calculated from the subject's facial orientation. [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Patent No. 6867715 [Patent Document 2] Patent No. 7547250 [Patent Document 3] Patent No. 6631250 [Patent Document 4] Patent No. 7509878 [Patent Document 5] Patent No. 7123539 [Overview of the project] [Problems that the invention aims to solve]

[0005] In Patent Document 1, the subject must continuously stare at a diagnostic image displayed on a screen or other display unit for a certain period of time, which places a heavy burden on the subject. In Patent Document 2, the method for determining dementia symptoms, which is one of the subject information items, is not disclosed. In Patent Document 3, the method for determining dementia symptoms is not disclosed, and a large device is required to evaluate standing posture balance. In Patent Document 4, the comparison with standard data for each age, gender, and language type is not verified, nor is the change in personal data collected over time verified. In Patent Document 5, the subject is filmed using an imaging device such as a home video camera, but home video cameras cannot capture the subject's natural facial expressions. This disclosure addresses these challenges and provides a dementia estimation device that can easily estimate dementia with minimal burden on the subject. [Means for solving the problem]

[0006] This disclosure relates to a dementia estimation device that estimates whether a subject has dementia based on the route the subject takes from leaving their home to returning home, and comprises: a terminal device having a GPS receiver; a route calculation unit that calculates the route the subject takes from leaving their home to returning home based on signals received by the GPS receiver; a learning data creation unit that creates learning data by machine learning the route of the subject calculated by the route calculation unit; and an estimation unit that estimates whether the subject has dementia using artificial intelligence based on the route of the subject calculated by the route calculation unit and the learning data created by the learning data creation unit. This configuration allows for easy estimation of dementia with minimal burden on the subject.

[0007] This disclosure relates to a dementia estimation device that estimates whether a subject has dementia based on the degree of change in the subject's facial expression, comprising: an imaging device that images the subject's face; an expression acquisition unit that acquires the degree of change in the subject's facial expression based on the image data of the subject's face captured by the imaging device; a learning data creation unit that creates learning data by machine learning the degree of change in the subject's facial expression acquired by the expression acquisition unit; and an estimation unit that estimates whether the subject has dementia using artificial intelligence based on the degree of change in the subject's facial expression acquired by the expression acquisition unit and the learning data created by the learning data creation unit. This configuration allows for easy estimation of dementia with minimal burden on the subject.

[0008] This disclosure relates to a dementia estimation device that estimates whether a subject has dementia based on the subject's answers to questions from an interviewer, comprising: a question acquisition unit that acquires the content of the interviewer's questions; a sound collection device that collects the subject's voice; an answer acquisition unit that acquires the subject's answers to the interviewer's questions from the voice of the subject collected by the sound collection device; a learning data creation unit that creates learning data by machine learning the content of the interviewer's questions acquired by the question acquisition unit and the content of the subject's answers acquired by the answer acquisition unit; and an estimation unit that estimates whether the subject has dementia using artificial intelligence based on the content of the interviewer's questions acquired by the question acquisition unit, the content of the subject's answers acquired by the answer acquisition unit, and the learning data created by the learning data creation unit. This configuration allows for easy estimation of dementia with minimal burden on the subject.

[0009] This disclosure relates to a dementia estimation device that estimates whether a subject has dementia based on the position of the subject's center of gravity, comprising: a terminal device having an acceleration sensor; a center of gravity calculation unit that calculates the position of the subject's center of gravity when the terminal device is attached; a learning data creation unit that creates learning data by machine learning the position of the subject's center of gravity calculated by the center of gravity calculation unit; and an estimation unit that estimates whether the subject has dementia using artificial intelligence based on the position of the subject's center of gravity calculated by the center of gravity calculation unit and the learning data created by the learning data creation unit. This configuration allows for easy estimation of dementia with minimal burden on the subject. [Effects of the Invention]

[0010] This disclosure provides a dementia estimation device that can easily estimate dementia with minimal burden on the subject. [Modes for carrying out the invention]

[0011] Embodiment 1 The dementia estimation device 10 according to Embodiment 1 will be explained using Figures 1 to 3. The dementia estimation device 10 is a device that estimates whether a subject has dementia based on the route the subject takes from leaving their home to returning home.

[0012] Figure 1 shows the system configuration of the dementia estimation device 10. The dementia estimation device 10 is installed in a smartphone 11. The smartphone 11 is a terminal device that the subject carries with them when going out. Note that the terminal device does not have to be a smartphone; it may be a wearable device, a tablet, or other terminal device. The smartphone 11 has a GPS (Global Positioning System) receiver 12, a first control unit 13, a memory 14, an input unit 15, a display unit 16, and a communication unit 17.

[0013] The GPS receiver 12 receives signals transmitted from an artificial satellite (not shown). The GPS receiver 12 transmits the received signal to the first control unit 13. The first control unit 13 includes a CPU (Central Processing Unit) and the like, and estimates dementia of the subject. The memory 14 includes a RAM (Random Access Memory), a ROM (Read Only Memory) and the like, and stores a control program executed by the first control unit 13, learning data created by the first control unit 13 and the like. The input unit 15 includes a touch panel and the like, and inputs data relating to the subject. The display unit 16 includes a display and the like, and displays a result of estimating dementia of the subject executed by the first control unit 13 and the like. The communication unit 17 is communicably connected to a communication network N, and is connected to a server 24 described later via a mobile phone line (for example, 4G or 5G), WIFI (Wireless Fidelity) or the like.

[0014] The first control unit 13 includes a route calculation unit 21, a learning data creation unit 22, and an estimation unit 23. The route calculation unit 21 calculates a route (behavior pattern) from when the subject leaves home to when the subject returns home based on the signal received by the GPS receiver 12.

[0015] The learning data creation unit 22 performs machine learning on the subject's route calculated by the route calculation unit 21 to create learning data. The learning data creation unit 22 creates learning data using artificial intelligence. The created learning data is stored in the memory 14, and is used by the estimation unit 23 for estimating dementia of the subject.

[0016] The estimation unit 23 estimates whether the subject has dementia using artificial intelligence based on the subject's route calculated by the route calculation unit 21 and the learning data created by the learning data creation unit 22. The result of the subject's dementia estimated by the estimation unit 23 is displayed on the display unit 16.

[0017] Note that the route calculation unit 21, learning data creation unit 22, and estimation unit 23 included in the first control unit 13 may be provided not in the smartphone 11 carried by the subject, but for example in a smartphone owned by the subject's family, or in a personal computer owned by the subject's attending physician or attending care manager. In this case, the signal received by the GPS receiver 12 from the smartphone 11 carried by the subject is transmitted to the smartphone of the subject's family or the like via wireless communication or the like.

[0018] Next, a dementia estimation processing program executed by the first control unit 13 will be described with reference to FIG. 2. FIG. 2 is an example of a flowchart of the dementia estimation processing program executed by the first control unit 13.

[0019] In step S1, when the subject carries the smartphone 11, goes out from home and returns home, the route calculation unit 21 calculates the route (behavior pattern) from when the subject goes out from home to when he / she returns home based on the signal received by the GPS receiver 12. In the initial state, the subject is a healthy person and is assumed to take habitual actions, so the subject's usual route (behavior pattern) is calculated. The route calculation unit 21 transmits the calculated route of the subject to the learning data creation unit 22 and the estimation unit 23.

[0020] In step S2, the learning data creation unit 22 performs machine learning on the subject's route calculated by the route calculation unit 21 to create learning data. The learning data creation unit 22 creates (updates) learning data by performing machine learning on a new route (learning data) against learning data created in the past using artificial intelligence. The learning data creation unit 22 performs machine learning using the route when the subject is a healthy person as teacher data. On the other hand, the subject may have dementia in the route calculated by the route calculation unit 21. Therefore, the learning data creation unit 22 may create learning data after, for example, the subject's family, attending physician, attending care manager or the like check the movement information. Alternatively, artificial intelligence may be used to determine whether learning is required. The created learning data is stored in the memory 14 and used by the estimation unit 23 for estimating dementia.

[0021] In step S3, the estimation unit 23 uses artificial intelligence to determine whether the subject's route is the same as usual, based on the subject's route calculated by the route calculation unit 21 and the learning data created by the learning data creation unit 22. For example, as shown in Figure 3, the route (behavior pattern) when a healthy subject goes out to XX supermarket and XX park is shown by a solid arrow. This route is built into the learning data. If the subject's route calculated by the route calculation unit 21 is a solid arrow, it is determined to be the same as usual. On the other hand, if the subject's route calculated by the route calculation unit 21 is a route that has not been seen before, as shown by a dashed line, it is determined to be different from the usual route. If the subject's route calculated by the route calculation unit 21 is the same as the usual route, the process proceeds to step S1; otherwise, the process proceeds to step S4.

[0022] In step S4, the estimation unit 23 estimates whether the subject has dementia. Even if the route calculated by the route calculation unit 21 in step S3 is determined to be different from the subject's usual route, the subject may intentionally take a different route than usual. Therefore, the number of times the route was determined to be different from the usual route in step S3 is counted, and if it is less than a predetermined number, the subject is estimated not to have dementia and the process proceeds to step S1. If the number of times the route was determined to be different from the usual route in step S3 exceeds a predetermined number, the subject is estimated to have dementia and the process proceeds to step S5. Note that the determination may be made based on consecutive occurrences rather than the total number of occurrences. Also, the value of the predetermined number may be set by artificial intelligence.

[0023] In step S5, the result estimated by the estimation unit 23 is displayed on the display unit 16. For example, the display unit 16 displays "Suspected dementia." If the estimation unit 23 estimates that the patient has dementia, the display unit 16 may also display the reason for this.

[0024] The dementia estimation processing program performed by the first control unit 13 may, for example, be performed on a server 24 owned by the patient's attending physician or care manager. In this case, the signal received by the GPS receiver 12 is transmitted from the smartphone 11 to the server 24 via the communication network N. The server 24 includes a sixth control unit 25, memory 26, and communication unit 27. The sixth control unit 25 includes a CPU, etc., and performs dementia estimation of the patient, etc., in the same way as the first control unit 13. The memory 26 includes RAM, ROM, etc., and stores control programs executed by the sixth control unit 25 and learning data created by the sixth control unit 25. The communication unit 27 is connected to the communication network N in a communication-enabled manner and connects to the smartphone 11 via a mobile phone line, Wi-Fi, etc. The dementia estimation processing program performed by the sixth control unit 25 is the same as the dementia estimation processing program performed by the first control unit 13, so its explanation is omitted.

[0025] In Embodiment 1, the subject can have their dementia status estimated simply by carrying a smartphone 11 equipped with a GPS receiver 12 when leaving their home. Therefore, dementia can be easily estimated with minimal burden on the subject.

[0026] Embodiment 2 The dementia estimation device 30 according to Embodiment 2 will be explained using Figures 4 and 5. The dementia estimation device 30 is a device that estimates whether a subject has dementia based on the degree of change in the subject's facial expression.

[0027] Figure 4 shows the system configuration of the dementia estimation device 30. The dementia estimation device 30 is installed in a smartphone 11. The smartphone 11 is a terminal device owned by the subject. Note that the terminal device may be other terminal devices such as wearable devices, tablet devices, or personal computers, instead of a smartphone. The smartphone 11 has a camera 32, a speaker 70, a second control unit 33, a memory 14, an input unit 15, a display unit 16, and a communication unit 17.

[0028] Camera 32 is an imaging device that captures images of the subject's face. It is preferable to fix the smartphone 11 with a holder or other fixing means so that the camera 32 can capture images of the subject's face. Camera 32 transmits the captured image data of the subject's face to the second control unit 33. Speaker 70 is a device that generates sound. Speaker 70 generates sound based on control signals from the second control unit 33.

[0029] The second control unit 33 includes a CPU and performs tasks such as estimating the subject's dementia. The memory 14 includes RAM, ROM, etc., and stores control programs executed by the second control unit 33. The input unit 15 includes a touch panel, etc., and inputs data about the subject. The display unit 16 includes a display, etc., and displays conversation text output by a chatbot program on the server 24 (described later), and the results of the subject's dementia estimation performed by the second control unit 33. The communication unit 17 is connected to the communication network N in a communication-enabled manner and connects to the server 24 (described later) via a mobile phone line (for example, 4G or 5G), WIFI (Wireless Fidelity), etc.

[0030] The second control unit 33 includes an expression acquisition unit 37, a learning data creation unit 38, and an estimation unit 39. The expression acquisition unit 37 acquires the degree of change in the subject's facial expression from image data of the subject's face captured by the camera 32. The learning data creation unit 38 creates learning data by machine learning the degree of change in the subject's facial expression acquired by the expression acquisition unit 37. The learning data creation unit 38 creates the learning data using artificial intelligence. The created learning data is stored in the memory 14 and the memory 26 on the server 24 side and is used by the estimation unit 39 to estimate the subject's dementia. Based on the degree of change in the subject's facial expression acquired by the expression acquisition unit 37 and the learning data created by the learning data creation unit 38, the estimation unit 39 uses artificial intelligence to estimate whether the subject has dementia. The result of the subject's dementia estimated by the estimation unit 39 is displayed on the display unit 16.

[0031] Furthermore, the facial expression acquisition unit 37, learning data creation unit 38, and estimation unit 39 of the second control unit 33 may be located not on the subject's smartphone 11, but on, for example, a smartphone owned by a family member of the subject, or a personal computer owned by the subject's doctor or care support specialist. In this case, the image data of the subject's face captured by the camera 32 from the subject's smartphone 11 is transmitted wirelessly or otherwise to the subject's family member's smartphone or the like.

[0032] The smartphone 11 is connected to the server 24 via a communication network N. The server 24 comprises a sixth control unit 25, memory 26, and a communication unit 27. The sixth control unit 25 includes a CPU and other components and executes a chatbot program that interacts with the subject. The memory 26 includes RAM, ROM, etc., and stores the chatbot program executed by the sixth control unit 25. The communication unit 27 is connected to the communication network N in a communicable manner and connects to the subject's smartphone 11 via a mobile phone line, Wi-Fi, etc. The communication unit 27 transmits the conversation text output by the chatbot program to the sixth control unit 25. In this embodiment 2, the chatbot program is executed on a server 24 separate from the subject's smartphone 11, but it may also be executed on the subject's smartphone 11.

[0033] Next, the dementia estimation processing program performed by the second control unit 33 will be explained using Figure 5. Figure 5 is an example of a flowchart of the dementia estimation processing program performed by the second control unit 33.

[0034] In step S11, the facial expression acquisition unit 37 acquires the degree of change in the subject's facial expression from the image data of the subject's face captured by the camera 32. For example, it acquires the degree of change in the subject's facial expression when the speaker 70 gives voice instructions such as "Please smile" or "Please make a sad face." Alternatively, it acquires the degree of change in the subject's facial expression when the chatbot program on the server 24 is activated and the subject is conversing with the chatbot. In the initial state, the subject is a healthy person, and it is expected that their facial expression will change, so the degree of change in the subject's usual facial expression is acquired. The facial expression acquisition unit 37 transmits the acquired degree of change in the subject's facial expression to the learning data creation unit 38 and the estimation unit 39.

[0035] In step S12, the learning data creation unit 38 creates learning data by machine learning the degree of facial expression change of the subject acquired by the facial expression acquisition unit 37. The learning data creation unit 38 uses artificial intelligence to machine learning new facial expression change degrees (learning data) to create (update) learning data to previously created learning data. The learning data creation unit 38 uses the degree of facial expression change when the subject is healthy as training data for machine learning. On the other hand, the degree of facial expression change acquired by the facial expression acquisition unit 37 may also be in the case of a subject with dementia. Therefore, the learning data creation unit 38 may create learning data after, for example, the subject's family, attending physician, or care support specialist has confirmed the facial expression information. Alternatively, it may use artificial intelligence to determine whether or not learning is necessary. The created learning data is stored in memory 14 and the server 24's memory 26 and used by the estimation unit 39 to estimate dementia.

[0036] In step S13, the estimation unit 39 uses artificial intelligence to determine whether the degree of change in the subject's facial expression is the same as usual, based on the degree of change in the subject's facial expression acquired by the facial expression acquisition unit 37 and the learning data created by the learning data creation unit 38. Specifically, the estimation unit 39 determines that the degree of change in the subject's facial expression acquired by the facial expression acquisition unit 37 is the same as the usual degree of change in the subject's facial expression if it is the same as the degree of change in the subject's facial expression built into the learning data, and determines that it is not the same as the usual degree of change in the subject's facial expression if it is not the same as the usual degree of change in the subject's facial expression. If the degree of change in the subject's facial expression acquired by the facial expression acquisition unit 37 is the same as the usual degree of change in the subject's facial expression, the process proceeds to step S11; otherwise, the process proceeds to step S14.

[0037] In step S14, the estimation unit 39 estimates whether the subject has dementia. Even if, in step S13, the degree of change in the subject's facial expression acquired by the facial expression acquisition unit 37 is determined to be different from the subject's usual degree of change, the subject may intentionally make a different facial expression than usual. Therefore, the number of times in step S13 that the degree of change in the subject's facial expression is determined to be different from the usual degree is counted, and if it is less than a predetermined number, it is estimated that the subject does not have dementia and the process proceeds to step S11. If the number of times in step S13 that the degree of change in the subject's facial expression is determined to be different from the usual number exceeds a predetermined number, it is estimated that the subject has dementia and the process proceeds to step S15. Note that the determination may be made based on consecutive counts rather than the total number of counts. Also, the value of the predetermined count may be set by artificial intelligence.

[0038] In step S15, the result estimated by the estimation unit 39 is displayed on the display unit 16. If the estimation unit 39 estimates that the patient has dementia, the display unit 16 may also display the reason for this estimation.

[0039] In Embodiment 2, it is possible to estimate whether a subject has dementia simply by capturing an image of the subject's face with the camera 32. Therefore, dementia can be easily estimated with minimal burden on the subject. Furthermore, by using the camera 32 of the subject's smartphone 11, it is possible to capture images while visually observing their usual facial expressions, and dementia can be estimated based on natural facial images.

[0040] Embodiment 3 The dementia estimation device 40 according to Embodiment 3 will be explained using Figures 6 and 7. The dementia estimation device 40 is a device that estimates whether a subject has dementia based on the subject's answers to the questioner's questions.

[0041] Figure 6 shows the system configuration of the dementia estimation device 40. The dementia estimation device 40 is installed in a smartphone 11. The smartphone 11 is a terminal device owned by the subject. Note that the terminal device does not have to be a smartphone; it may be a wearable device, a tablet device, a personal computer, or other terminal device. The smartphone 11 has a microphone 42, a GPS receiver 12, a third control unit 43, a memory 14, an input unit 15, a display unit 16, and a communication unit 17.

[0042] The microphone 42 is a sound collection device that collects the subject's voice. The microphone 42 may be separate from the smartphone 11 in order to collect the subject's voice clearly. The microphone 42 transmits the collected voice of the subject to the third control unit 43. The GPS receiver 12 receives signals transmitted from an artificial satellite (not shown). The GPS receiver 12 transmits the received signals to the third control unit 43.

[0043] The third control unit 43 includes a CPU and performs tasks such as estimating the subject's dementia. The memory 14 includes RAM, ROM, etc., and stores control programs executed by the third control unit 43. The input unit 15 includes a touch panel and is used to input data about the subject. The display unit 16 includes a display and shows questions output by a chatbot program on the server 24 (described later), and the results of the subject's dementia estimation performed by the third control unit 43. The communication unit 17 is connected to the communication network N in a communication-enabled manner and connects to the server 24 (described later) via a mobile phone line (for example, 4G or 5G), Wi-Fi (Wireless Fidelity), etc.

[0044] The third control unit 43 includes a question acquisition unit 47, an answer acquisition unit 48, a learning data creation unit 49, and an estimation unit 50. The question acquisition unit 47 acquires the question content of the questioner as text data from the question text output by the chatbot program on the server 24, which will be described later. The answer acquisition unit 48 acquires the subject's answer to the questioner's question as text data from the audio collected by the microphone 42.

[0045] The learning data creation unit 49 creates learning data by machine learning the question content of the questioner obtained by the question acquisition unit 47 and the answer content of the subject obtained by the answer acquisition unit 48. The learning data creation unit 49 creates the learning data using artificial intelligence. The created learning data is stored in memory 14 and the memory 26 on the server 24 side and is used by the estimation unit 50 to estimate the subject's dementia. The estimation unit 50 uses artificial intelligence to estimate whether the subject has dementia based on the question content of the questioner obtained by the question acquisition unit 47, the answer content of the subject obtained by the answer acquisition unit 48, and the learning data created by the learning data creation unit 49. The result of the subject's dementia estimation by the estimation unit 50 is displayed on the display unit 16.

[0046] Furthermore, the question acquisition unit 47, answer acquisition unit 48, learning data creation unit 49, and estimation unit 50 of the third control unit 43 may be located not on the subject's smartphone 11, but on, for example, a smartphone owned by a family member of the subject, or a personal computer owned by the subject's doctor or care support specialist. In this case, the question text output from the chatbot program of the server 24 (described later), and the subject's voice collected by the microphone 42 are transmitted wirelessly from the subject's smartphone 11 to the subject's family member's smartphone or the like.

[0047] The smartphone 11 is connected to the server 24 via a communication network N. The server 24 comprises a sixth control unit 25, memory 26, and a communication unit 27. The sixth control unit 25 includes a CPU and other components and executes a chatbot program that interacts with the subject. The memory 26 includes RAM, ROM, etc., and stores the chatbot program executed by the sixth control unit 25. The communication unit 27 is connected to the communication network N in a communicable manner and connects to the subject's smartphone 11 via a mobile phone line, Wi-Fi, etc. The communication unit 27 transmits the question text output by the chatbot program to the third control unit 43. In this embodiment 3, the chatbot program is executed on a server 24 separate from the subject's smartphone 11, but it may also be executed on the subject's smartphone 11.

[0048] Next, the dementia estimation processing program performed by the third control unit 43 will be explained using Figure 7. Figure 7 is an example of a flowchart of the dementia estimation processing program performed by the third control unit 43.

[0049] In step S21, the question acquisition unit 47 acquires the question from the questioner as text data from the question output by the chatbot program on the server 24. When the dementia estimation processing program is started in the third control unit 43, the chatbot program on the server 24 operates and outputs a question, and the outputted question is sent to the third control unit 43 of the smartphone 11 and displayed on the display unit 16. In Embodiment 3, the questioner is the server 24, but it may also be a robot or a personal computer. The questioner may also be a natural person such as a family member of the subject, the attending physician, or the assigned care support specialist. In Embodiment 3, the questioner asks the subject via the communication network N, but the questioner may also ask the subject face-to-face. Examples of questions from the questioner include, "What was your behavioral pattern today?", "What is today's date and day of the week?", and "What was your behavioral pattern yesterday?". The question acquisition unit 47 transmits the acquired question from the questioner to the learning data creation unit 49 and the estimation unit 50.

[0050] In step S22, the response acquisition unit 48 acquires the subject's response to the questioner's question from the subject's voice collected by the microphone 42. When the subject responds verbally to the question displayed on the display unit 16, the response acquisition unit 48 acquires the subject's response to the questioner's question as text data from the subject's voice collected by the microphone 42. In the initial state, the subject is assumed to be healthy and to respond correctly to the questioner's question, so the subject's usual response content is acquired. The response acquisition unit 48 transmits the acquired subject's response content to the learning data creation unit 49 and the estimation unit 50.

[0051] In step S23, the learning data creation unit 49 creates learning data by machine learning the question content of the questioner obtained by the question acquisition unit 47 and the answer content of the subject obtained by the answer acquisition unit 48. The learning data creation unit 49 uses artificial intelligence to create (update) learning data by machine learning new question content and answer content (learning data) to previously created learning data. The learning data creation unit 49 uses the answer content to the question content when the subject is healthy as training data for machine learning. On the other hand, the answer content of the subject obtained by the answer acquisition unit 48 may indicate that the subject has dementia. Therefore, the learning data creation unit 49 may create the learning data after, for example, the subject's family, attending physician, or care support specialist has confirmed the subject's answer content. The created learning data is stored in memory 14 and the memory 26 on the server 24 side and is used by the estimation unit 50 to estimate dementia.

[0052] In step S24, the estimation unit 50 uses artificial intelligence to determine whether the subject's answer to the question is correct, based on the question content obtained by the question acquisition unit 47, the subject's answer content obtained by the answer acquisition unit 48, and the learning data creation unit 49. Specifically, the estimation unit 50 determines that the subject's answer to the question is correct if it matches the subject's answer content built in the learning data, and incorrect if it does not. Furthermore, if the questioner's question is "What was your behavior pattern yesterday?", the estimation unit 50 compares the subject's actual behavior pattern calculated based on the signal received by the GPS receiver 12, along with the answer content built in the learning data, to make a judgment. If the subject's answer to the question is correct, the process proceeds to step S21; otherwise, it proceeds to step S25.

[0053] In step S25, the estimation unit 50 estimates whether the subject has dementia. Even if the subject's answer to the questioner's question is judged to be incorrect in step S24, the subject may intentionally give an answer that is different from their usual one. Therefore, the number of times the answer was judged to be incorrect in step S24 is counted, and if it is less than a predetermined number, the subject is estimated not to have dementia and the process proceeds to step S21. If the number of times the answer was judged to be incorrect in step S24 exceeds a predetermined number, the subject is estimated to have dementia and the process proceeds to step S26. Note that the judgment may be made based on consecutive counts rather than the total number. Also, the value of the predetermined count may be set by artificial intelligence.

[0054] In step S26, the result estimated by the estimation unit 50 is displayed on the display unit 16. If the estimation unit 50 estimates that the patient has dementia, the display unit 16 may also display the reason for this estimation.

[0055] In Embodiment 3, it is possible to estimate whether a subject has dementia simply by collecting the subject's voice with the microphone 42. Therefore, dementia can be estimated easily with minimal burden on the subject. Furthermore, dementia is estimated using a learning data creation unit 49 that has acquired conversation information between the questioner and the subject through machine learning. By comparing this data with standard data for each age, gender, and language type, and by verifying changes in personal data collected over time, dementia can be estimated with high accuracy.

[0056] Embodiment 4 The dementia estimation device 60 according to Embodiment 4 will be explained using Figures 8 to 12. The dementia estimation device 60 is a device that estimates whether a subject has dementia based on the position of the subject's center of gravity.

[0057] Figure 8 shows the system configuration of the dementia estimation device 60. The dementia estimation device 60 is installed in a smartphone 11. The smartphone 11 is a terminal device worn by the subject when measuring the center of gravity. Note that the terminal device does not have to be a smartphone; it may be a wearable device, a tablet, or other terminal device. The smartphone 11 has an acceleration sensor 62, a fourth control unit 63, a memory 14, an input unit 15, a display unit 16, and a communication unit 17.

[0058] As shown in Figure 9, the acceleration sensor 62 is a three-axis acceleration sensor that measures acceleration in three dimensions, indicated by the X, Y, and Z axes, relative to the smartphone 11. The acceleration sensor 62 transmits the measured acceleration in three dimensions to the fourth control unit 63. The fourth control unit 63 includes a CPU and performs tasks such as estimating the subject's dementia. The memory 14 includes RAM, ROM, etc., and stores control programs executed by the fourth control unit 63. The input unit 15 includes a touch panel and is used to input data related to the subject. The display unit 16 includes a display and shows the results of the subject's dementia estimation performed by the fourth control unit 63. The communication unit 17 is connected to the communication network N and connects to the server 24, which will be described later, via a mobile phone line (for example, 4G or 5G), WIFI (Wireless Fidelity), etc.

[0059] The fourth control unit 63 includes a center of gravity calculation unit 67, a learning data creation unit 68, and an estimation unit 69. The center of gravity calculation unit 67 calculates the position of the subject's center of gravity from the three-dimensional acceleration measured by the acceleration sensor 62. The learning data creation unit 68 creates learning data by machine learning the position of the subject's center of gravity calculated by the center of gravity calculation unit 67. The learning data creation unit 68 creates the learning data using artificial intelligence. The created learning data is stored in the memory 14 and used by the estimation unit 69 to estimate the subject's dementia. The estimation unit 69 uses artificial intelligence to estimate whether the subject has dementia based on the position of the subject's center of gravity calculated by the center of gravity calculation unit 67 and the learning data created by the learning data creation unit 68. The result of the subject's dementia estimation by the estimation unit 69 is displayed on the display unit 16.

[0060] Furthermore, the center of gravity calculation unit 67, learning data creation unit 68, and estimation unit 69 of the fourth control unit 63 may be located not on the smartphone 11 worn by the subject, but on, for example, a smartphone owned by a family member of the subject, or a personal computer owned by the subject's doctor or care support specialist. In this case, the orientation and acceleration measured by the acceleration sensor 62 are transmitted wirelessly from the smartphone 11 worn by the subject to the family member's smartphone or the like.

[0061] Next, the dementia estimation processing program performed by the fourth control unit 63 will be explained using Figure 10. Figure 10 is an example of a flowchart of the dementia estimation processing program performed by the fourth control unit 63.

[0062] In step S31, with the subject wearing the smartphone 11, the center of gravity calculation unit 67 calculates the subject's center of gravity based on the three-dimensional acceleration measured by the acceleration sensor 62. To measure the subject's center of gravity more accurately, it is preferable to attach the smartphone 11 around the subject's waist. Also, to transmit the subject's movements more accurately to the smartphone 11, it is preferable to wrap the belt containing the smartphone 11 around the subject's waist for measurement. The subject, wearing the smartphone 11, goes barefoot and maintains an upright posture with both feet together for several tens of seconds. During this time, the center of gravity calculation unit 67 calculates the subject's center of gravity. In the initial state, it is assumed that the subject is healthy and in a balanced posture, and the subject's usual center of gravity is calculated.

[0063] As shown in Figure 11, the center of gravity calculation unit 67 calculates the center of gravity of the subject using two-dimensional coordinate values ​​in the left-right and front-back directions. The center of gravity of the subject calculated by the center of gravity calculation unit 67 is expressed in terms of the direction of the center of gravity and the magnitude of the deviation. The direction of the center of gravity and the magnitude of the deviation are calculated by integrating twice the three-dimensional acceleration measured by the acceleration sensor 62. The center of gravity calculation unit 67 measures the center of gravity at predetermined time intervals and displays the average value of the center of gravity in a radar chart 67a as shown in Figure 12. The radar chart 67a consists of three concentric regular octagons and eight axes connecting the center to each vertex of the regular octagon. The eight axes point in eight directions: front-back, front-left, left-right, front-right, rear-right, rear-left, and front-left, representing the direction of the center of gravity. Each axis also represents the magnitude of the deviation of the center of gravity, with the deviation increasing as you move radially outward from the center. The center of gravity is displayed on the radar chart 67a as black circles, corresponding to the direction and magnitude of the deviation. The center of gravity calculation unit 67 transmits the calculated center of gravity of the subject to the learning data creation unit 68 and the estimation unit 69.

[0064] In step S32, the learning data creation unit 68 creates learning data by machine learning the center of gravity position of the subject calculated by the center of gravity calculation unit 67. The learning data creation unit 68 uses artificial intelligence to machine learning new center of gravity positions (learning data) to previously created learning data and creates (updates) the learning data. The learning data creation unit 68 uses the center of gravity position when the subject is healthy as training data for machine learning. On the other hand, the center of gravity position calculated by the center of gravity calculation unit 67 may also be the case when the subject has dementia. Therefore, the learning data creation unit 68 may, for example, have the subject's family, doctor, or care manager confirm the movement information before creating the learning data. Alternatively, it may use artificial intelligence to determine whether or not learning is necessary. The created learning data is stored in memory 14 and used by the estimation unit 69 to estimate dementia.

[0065] In step S33, the estimation unit 69 uses artificial intelligence to determine whether the subject's center of gravity is the same as usual, based on the subject's center of gravity calculated by the center of gravity calculation unit 67 and the learning data created by the learning data creation unit 68. Specifically, if the subject's center of gravity calculated by the center of gravity calculation unit 67 is the same as the subject's center of gravity constructed in the learning data, it is determined to be the same as usual; otherwise, it is determined to be different from usual. If the subject's center of gravity calculated by the center of gravity calculation unit 67 is the same as usual, the process proceeds to step S31; otherwise, the process proceeds to step S34.

[0066] In step S34, the estimation unit 69 estimates whether the subject has dementia. Even if the center of gravity calculation unit 67 determines in step S33 that the subject's center of gravity is not the same as the usual center of gravity, the subject's physical condition may cause the center of gravity to be different from usual. Therefore, the number of times the center of gravity was determined to be different from the usual in step S33 is counted, and if it is less than a predetermined number, the subject is estimated not to have dementia and the process proceeds to step S31. If the number of times the center of gravity was determined to be different from the usual in step S33 exceeds a predetermined number, the subject is estimated to have dementia and the process proceeds to step S35. Note that the determination may be based on consecutive occurrences rather than the total number of occurrences. Also, the value of the predetermined number may be set by artificial intelligence.

[0067] In step S35, the results estimated by the estimation unit 69 are displayed on the display unit 16. The display unit 16 may also display the centroid position of the subject shown in Figure 12 along with the estimation results.

[0068] The dementia estimation processing program performed by the fourth control unit 63 may, for example, be performed on a server 24 owned by the patient's attending physician or care support specialist. In this case, the three-dimensional acceleration measured by the acceleration sensor 62 is transmitted from the smartphone 11 to the server 24 via the communication network N. The server 24 includes a sixth control unit 25, a memory 26, and a communication unit 27. The sixth control unit 25 includes a CPU, etc., and performs dementia estimation of the patient, etc., similar to the fourth control unit 63. The memory 26 includes RAM, ROM, etc., and stores the control program executed by the sixth control unit 25 and the learning data created by the sixth control unit 25. The communication unit 27 is connected to the communication network N in a communicative manner and connects to the smartphone 11 via a mobile phone line, Wi-Fi, etc. The dementia estimation processing program performed by the sixth control unit 25 is the same as the dementia estimation processing program performed by the fourth control unit 63, so its explanation is omitted.

[0069] In Embodiment 4, the subject can have dementia estimated simply by wearing a smartphone 11 equipped with an acceleration sensor 62. Therefore, dementia can be estimated easily with minimal burden on the subject. Since the center of gravity can be calculated using the smartphone 11, dementia can be estimated without the need for large-sized devices.

[0070] Embodiment 5 The dementia estimation device 100 according to Embodiment 5 will be explained using Figures 13 to 14. The dementia estimation device 100 is a device that estimates whether a subject has dementia based on the results of Embodiments 1 to 4.

[0071] Figure 13 shows the system configuration of the dementia estimation device 100. The dementia estimation device 100 is installed in a smartphone 11. The smartphone 11 is a terminal device owned by the subject. The smartphone 11 has a GPS receiver 12, a camera 32, a speaker 70, a microphone 42, an acceleration sensor 62, a fifth control unit 107, an input unit 15, a memory 14, a display unit 16, and a communication unit 17. The configuration and functions of the GPS receiver 12, camera 32, speaker 70, microphone 42, and acceleration sensor 62 are the same as those of the GPS receiver 12, etc. in Embodiments 1 to 4, so their explanation is omitted.

[0072] The fifth control unit 107 includes a CPU and performs tasks such as estimating the subject's dementia. The memory 14 includes RAM, ROM, etc., and stores control programs executed by the fifth control unit 107. The input unit 15 includes a touch panel, etc., and inputs data about the subject. The display unit 16 includes a display, etc., and displays the results of the subject's dementia estimation performed by the fifth control unit 107. The communication unit 17 is connected to the communication network N in a communication-enabled manner and connects to the server 24, which will be described later, via a mobile phone line (for example, 4G or 5G), WIFI (Wireless Fidelity), etc.

[0073] The fifth control unit 107 includes a first control unit 13, a second control unit 33, a third control unit 43, and a fourth control unit 63. The first control unit 13 is the first control unit 13 of the dementia estimation device 10 according to Embodiment 1, and estimates whether the subject has dementia based on the route the subject takes from leaving their home to returning home. The second control unit 33 is the second control unit 33 of the dementia estimation device 30 according to Embodiment 2, and estimates whether the subject has dementia based on the degree of change in the subject's facial expression. The third control unit 43 is the third control unit 43 of the dementia estimation device 40 according to Embodiment 3, and estimates whether the subject has dementia based on the subject's answers to the questioner's questions. The fourth control unit 63 is the fourth control unit 63 of the dementia estimation device 60 according to Embodiment 4, and estimates whether the subject has dementia based on the position of the subject's center of gravity. The dementia results of the subject estimated by the first control unit 13, second control unit 33, third control unit 43, and fourth control unit 63 are displayed on the display unit 16.

[0074] The smartphone 11 is connected to the server 24 via the communication network N. The server 24 comprises a sixth control unit 25, memory 26, and a communication unit 27. The sixth control unit 25 includes a CPU and other components and executes a chatbot program that interacts with the subject. The memory 26 includes RAM, ROM, etc., and stores the chatbot program executed by the sixth control unit 25. The communication unit 27 is connected to the communication network N in a communication-enabled manner and connects to the subject's smartphone 11 via a mobile phone line, Wi-Fi, etc. The communication unit 27 transmits the question text and other information output by the chatbot program to the fifth control unit 107.

[0075] Next, we will explain the dementia estimation processing program performed by the fifth control unit 107 using Figure 14. Figure 14 is an example of a flowchart of the dementia estimation processing program performed by the fifth control unit 107.

[0076] In step S101, the first control unit 13 estimates whether the subject has dementia. Specifically, it estimates whether the subject has dementia based on the route the subject takes from leaving their home to returning home. The estimation method is the same as in Embodiment 1, so a detailed explanation is omitted. If it is estimated that the subject has dementia, the process proceeds to step S102; if it is estimated that the subject does not have dementia, the process ends.

[0077] In step S102, the second control unit 33 estimates whether the subject has dementia. Specifically, it estimates whether the subject has dementia based on the degree of change in the subject's facial expression. The estimation method is the same as in Embodiment 2, so a detailed explanation is omitted. If it is estimated that the subject has dementia, the process proceeds to step S103; if it is estimated that the subject does not have dementia, the process ends.

[0078] In step S103, the third control unit 43 estimates whether the subject has dementia. Specifically, it estimates whether the subject has dementia based on the subject's answers to the questioner's questions. The estimation method is the same as in Embodiment 3, so a detailed explanation is omitted. If it is estimated that the subject has dementia, the process proceeds to step S104; if it is estimated that the subject does not have dementia, the flow ends.

[0079] In step S104, the fourth control unit 63 estimates whether the subject has dementia. Specifically, it estimates whether the subject has dementia based on the position of the subject's center of gravity. The estimation method is the same as in Embodiment 4, so a detailed explanation is omitted. If it is estimated that the subject has dementia, the process proceeds to step S105; if it is estimated that the subject does not have dementia, the process ends.

[0080] In step S105, the results estimated by the first control unit 13, the second control unit 33, the third control unit 43, and the fourth control unit 63 are displayed on the display unit 16.

[0081] In Embodiment 5, multiple means are combined to estimate whether the subject has dementia, thus enabling a highly accurate estimation of the subject's dementia. [Brief explanation of the drawing]

[0082] [Figure 1] This is a system configuration diagram of the dementia estimation device according to Embodiment 1. [Figure 2] This is an example of a flowchart for a dementia estimation processing program using the dementia estimation device according to Embodiment 1. [Figure 3] This is an explanatory diagram of the dementia estimation device according to Embodiment 1. [Figure 4] This is a system configuration diagram of the dementia estimation device according to Embodiment 2. [Figure 5] This is an example of a flowchart for a dementia estimation processing program using the dementia estimation device according to Embodiment 2. [Figure 6] This is a system configuration diagram of the dementia estimation device according to Embodiment 3. [Figure 7] This is an example of a flowchart for a dementia estimation processing program using the dementia estimation device according to Embodiment 3. [Figure 8] This is a system configuration diagram of the dementia estimation device according to Embodiment 4. [Figure 9] This is an explanatory diagram of the acceleration sensor of the dementia estimation device according to Embodiment 4. [Figure 10] This is an example of a flowchart for a dementia estimation processing program using the dementia estimation device according to Embodiment 4. [Figure 11] This figure shows the change in the center of gravity of the subject over time. [Figure 12] This is a radar chart showing the center of gravity of the subject. [Figure 13] This is a system configuration diagram of the dementia estimation device according to Embodiment 5. [Figure 14] This is an example of a flowchart for a dementia estimation processing program using the dementia estimation device according to Embodiment 5.

[0083] This disclosure is not limited to the embodiments described above, and may be modified as appropriate without departing from its spirit. [Explanation of Symbols]

[0084] 10, 30, 40, 60, 100...Dementia estimation device 11. Smartphone 12. GPS receiver 32...Camera 42... Microphone 62. Accelerometer

Claims

1. A dementia estimation device that estimates whether a subject has dementia based on the route the subject takes from leaving their home to returning home, A terminal device having a GPS receiver, and a route calculation unit that calculates the route taken by the subject from leaving their home to returning home based on the signals received by the GPS receiver. A learning data creation unit creates learning data by machine learning the subject's route calculated by the route calculation unit, The system includes an estimation unit that uses artificial intelligence to estimate whether the subject has dementia, based on the subject's route calculated by the route calculation unit and the learning data created by the learning data creation unit. Dementia estimation device.

2. A dementia estimation device that estimates whether a subject has dementia based on the degree of change in the subject's facial expression, An imaging device for capturing images of the subject's face, A facial expression acquisition unit acquires the degree of change in the facial expression of the subject based on the image data of the subject's face captured by the imaging device, A learning data creation unit creates learning data by machine learning the degree of change in the subject's facial expression acquired by the facial expression acquisition unit, The system includes an estimation unit that uses artificial intelligence to estimate whether the subject has dementia, based on the degree of change in the subject's facial expression acquired by the facial expression acquisition unit and the learning data created by the learning data creation unit. Dementia estimation device.

3. A dementia estimation device that estimates whether a subject has dementia based on the subject's answers to the questioner's questions, A question acquisition unit that acquires the content of the question from the aforementioned questioner, A sound collection device for collecting the voice of the subject, The sound collection device includes an answer acquisition unit that acquires the subject's answers to the questioner's questions from the subject's voice collected by the sound collection device, A learning data creation unit creates learning data by machine learning the question content of the questioner obtained by the question acquisition unit and the answer content of the subject obtained by the answer acquisition unit. The system includes an estimation unit that uses artificial intelligence to estimate whether the subject has dementia, based on the question content obtained by the question acquisition unit, the answer content obtained by the subject, and the learning data created by the learning data creation unit. Dementia estimation device.

4. A dementia estimation device that estimates whether a subject has dementia based on the position of the subject's center of gravity, A terminal device having an acceleration sensor, A center of gravity calculation unit that calculates the center of gravity position of the subject wearing the terminal device, and a learning data creation unit that creates learning data by machine learning the center of gravity position of the subject calculated by the center of gravity calculation unit. The system includes an estimation unit that uses artificial intelligence to estimate whether the subject has dementia, based on the center of gravity position of the subject calculated by the center of gravity calculation unit and the learning data created by the learning data creation unit. Dementia estimation device.

Images