Stroke examination system, stroke examination method, and procedure recording medium

Abstract

The stroke examination system (500) examines a subject for signs of stroke. The stroke examination system (500) includes: an acquisition unit (101) that acquires personal data information related to the subject's brain disease; a decision unit (102) that determines the priority of each of a plurality of examination items related to stroke based on the acquired personal data information; a plurality of examination units (103) that each of the plurality of examination units (103) performs an examination of each of the plurality of examination items, and the plurality of examination units (103) perform the examination of the plurality of examination items in descending order of the determined priority; and a diagnosis unit (104) that outputs diagnostic information related to the signs of stroke of the subject based on the examination results.

Description

Technical Field

[0001] This disclosure relates to a stroke examination system, a stroke examination method, and a procedure recording medium for examining signs of stroke in subjects. Background Technology

[0002] It is known that prompt treatment of stroke can prevent serious sequelae and increase the chances of a cure. Therefore, it is desirable to perform immediate examinations for stroke precursors when a stroke is suspected. For example, a stroke detection method has been developed that allows for easy examination using information terminals such as smartphones, which are commonly used in recent years, enabling immediate detection of stroke precursors (see, for example, Patent Document 1).

[0003] (Existing technical literature)

[0004] (Patent Document)

[0005] Patent Document 1: International Publication No. 2018 / 053521 Summary of the Invention

[0006] The problem the invention aims to solve

[0007] However, from the perspective of performing appropriate checks, the stroke detection method disclosed in the aforementioned Patent Document 1 is inadequate.

[0008] In view of the above problems, the purpose of this disclosure is to provide a stroke examination system, etc., capable of performing appropriate examinations.

[0009] The means used to solve the problem

[0010] To achieve the above objectives, one embodiment of the stroke examination system disclosed herein examines a subject for signs of stroke. The stroke examination system comprises: an acquisition unit that acquires personal data information related to the subject's brain disease; a decision unit that determines the priority of each of a plurality of stroke-related examination items based on the acquired personal data information; a plurality of examination units, each of which performs an examination of each of the plurality of examination items, the plurality of examination units performing the examinations in descending order of the determined priority; and a diagnosis unit that outputs diagnostic information related to the stroke signs of the subject based on the examination results.

[0011] Furthermore, in one form of the stroke examination method disclosed herein, personal data information related to the medical records of the examinee’s previous brain diseases is obtained, and based on the obtained personal data information, the priority of each of a plurality of examination items related to stroke is determined, and each of the plurality of examination items is performed in descending order of the determined priority.

[0012] In addition, these general or specific forms can be realized by systems, devices, integrated circuits, computer programs or computer-readable CD-ROMs and other recording media, or by any combination of systems, devices, integrated circuits, computer programs and recording media.

[0013] Invention Effects

[0014] This disclosure provides a stroke examination system capable of performing appropriate examinations, etc. Attached Figure Description

[0015] Figure 1 This is a schematic diagram illustrating an example of the configuration of a stroke detection system according to an embodiment.

[0016] Figure 2 This is a block diagram illustrating the functional configuration of the stroke examination system according to the embodiment.

[0017] Figure 3 This is a block diagram comparing the functional configuration of the stroke examination system according to the implementation method with the functional configuration of a smartphone as a form of portable terminal device.

[0018] Figure 4 This is an explanatory diagram of a 3-axis sensor and a 3-axis angular velocity sensor found in a smartphone, which is a type of portable terminal device.

[0019] Figure 5 This is a flowchart illustrating a working example of the stroke examination system according to the embodiment.

[0020] Figure 6 Figure 1 is an example of an operation screen of the stroke examination system according to the embodiment.

[0021] Figure 7A Figure 2 is an example of an operation screen of the stroke examination system according to the embodiment.

[0022] Figure 7B Figure 3 is an example of an operation screen of the stroke examination system according to the embodiment.

[0023] Figure 8AFigure 4 is an example of an operation screen of the stroke examination system according to the embodiment.

[0024] Figure 8B Figure 5 shows an example of the operation screen of the stroke examination system according to the embodiment.

[0025] Figure 9A Figure 6 shows an example of the operation screen of the stroke examination system according to the embodiment.

[0026] Figure 9B Figure 7 is an example of an operation screen of the stroke examination system according to the embodiment.

[0027] Figure 10 An example of a neural network used to check for facial paralysis when implementing the embodiment is shown.

[0028] Figure 11A An example of a neural network is shown that takes into account the effects of facial image rotation during the capture for checking facial paralysis when implementing the embodiment.

[0029] Figure 11B An example of a neural network is shown that, when implementing an embodiment, takes into account the effects of the skew between the upper (forehead) and lower (jaw) parts of the face in the facial image taken to check for facial paralysis.

[0030] Figure 12 An example is shown of correcting tilt that occurs when capturing an image of the subject's face during the implementation of the method.

[0031] Figure 13 The image shows the subject holding the stroke examination device with their arm extended, having their face photographed, and simultaneously undergoing examinations for Barré's sign (a mild symptom of numbness in the hands and feet that occurs during the raising and lowering of the limbs) and facial paralysis. Detailed Implementation

[0032] (Basic insights of this invention)

[0033] In recent years, most people carry high-performance information terminals (smartphones, tablets, and personal computers) capable of information processing. On the other hand, as explained in the "Background Art" section above, it is necessary to quickly check for stroke precursors in individuals suspected of having a stroke. If a simple check for stroke precursors can be performed on-site using an information terminal when a stroke is suspected, appropriate action can be taken immediately according to the urgency of the situation. Therefore, as shown in Patent Document 1, a stroke detection method has been developed that allows for immediate detection of stroke precursors by using a simple check with an information terminal.

[0034] However, as mentioned above, if a stroke is suspected and the information terminal is operated by someone without the necessary knowledge, it could lead to delays in treatment. Specifically, in situations where every second counts, such as sequentially performing multiple checks, time may be wasted; or, for example, if the person being checked and the operator are different, errors may occur because it is unclear whether the instructions are addressed to the person being checked or the operator, ultimately resulting in wasted time.

[0035] In view of the above problems, this disclosure provides a stroke examination system, etc., which can perform multiple examination items in an appropriate order, and can also give appropriate operation instructions for each examination item based on whether the operator and the examinee are the same.

[0036] (Public Summary)

[0037] The summary of this disclosure is as follows.

[0038] One aspect of this disclosure relates to a stroke examination system that examines a subject for signs of stroke. The stroke examination system comprises: an acquisition unit that acquires personal data information related to the subject's brain disease; a decision unit that determines the priority of each of a plurality of stroke-related examination items based on the acquired personal data information; a plurality of examination units, each of which performs an examination of each of the plurality of examination items, the plurality of examination units performing the examinations in descending order of the determined priority; and a diagnosis unit that outputs diagnostic information related to the subject's signs of stroke based on the examination results.

[0039] This stroke screening system prioritizes multiple examination items based on personal data related to the examinee's brain disease and performs each examination item accordingly. If the determined priority corresponds to a level of usefulness, the examination unit can perform each of the multiple examination items in descending order of usefulness, where usefulness refers to the usefulness of each examination item for the examinee. Furthermore, because useful results can be obtained relatively early after the start of the examination, diagnosis can proceed without waiting for subsequent results. Therefore, from a time-saving perspective, it can appropriately perform examinations for strokes with high urgency.

[0040] Furthermore, personal information can also include, for example, information related to the examinee's past medical records of brain diseases.

[0041] Accordingly, information related to the examinee's past brain disease medical records can be used as personal data to determine the priority of each of the multiple examination items related to stroke.

[0042] Furthermore, for example, multiple examination items may include at least one of the following: examination items related to facial paralysis of the subject, examination items related to Barrett's sign of the subject, examination items related to dysarthria of the subject, and examination items related to gait impairment of the subject.

[0043] Accordingly, since each of the multiple examination items is a different one among the examination items related to the subject's facial paralysis, the subject's Barrett's sign, the subject's dysarthria, and the subject's gait disorder, it is possible to determine the priority of these examination items and perform the examinations according to that priority.

[0044] Furthermore, for example, if the personal data includes the examinee's medical records with specific symptoms, the decision-making unit may prioritize the examination items related to the examinee's specific symptoms over the other examination items.

[0045] Therefore, when there are existing medical records showing specific symptoms that are easy to identify in the examinee, the priority of examinations to be performed can be increased based on the presence or absence of those specific symptoms. Thus, if a specific symptom that is anticipated to be a sign of stroke recurs, that examination can be prioritized.

[0046] Furthermore, for example, if the personal data includes the examinee's medical records with specific symptoms, the decision-making unit may prioritize the examination items related to the examinee's specific symptoms over the other examination items.

[0047] Therefore, when a subject is found to have specific symptoms, the priority of upcoming examinations can be lowered based on the presence or absence of those specific symptoms. Thus, when it is difficult to determine whether a specific symptom indicating stroke is a newly discovered symptom, other examinations can be prioritized.

[0048] Furthermore, for example, if the personal data includes the examinee's medical history of facial paralysis, the decision-making department may prioritize the examination items related to the examinee's facial paralysis among multiple examination items as lower than the priority of other examination items.

[0049] Therefore, when facial paralysis has been detected in the subject, the priority of upcoming examinations can be reduced based on the presence or absence of this facial paralysis. Thus, when it is difficult to determine whether facial paralysis, a symptom of stroke, is newly discovered, examinations other than those related to facial paralysis can be prioritized.

[0050] Furthermore, for example, if the personal data includes the examinee's medical records showing symptoms of Barrett's disease, the decision-making department may prioritize the examination items related to the examinee's symptoms of Barrett's disease over the other examination items.

[0051] Therefore, when Barrett's sign has been detected in the subject, the priority of examinations to be performed can be reduced based on the presence or absence of this sign. Thus, when it is difficult to determine whether a Barrett's sign, which is considered a sign of stroke, is a newly discovered sign, examinations other than those related to Barrett's sign can be prioritized.

[0052] Furthermore, for example, if the personal data includes the examinee's medical history of dysarthria, the decision-making unit may prioritize the examination items related to the examinee's dysarthria among the multiple examination items as lower than the priority of other examination items.

[0053] Therefore, when a subject is found to have dysarthria, the priority of upcoming examinations can be reduced based on the presence or absence of the dysarthria. Thus, when it is difficult to determine whether a dysarthria presenting as a sign of stroke is a newly discovered dysarthria, examinations other than those related to dysarthria can be prioritized.

[0054] Furthermore, for example, if the personal data includes the examinee's medical records showing gait impairment, the decision-making unit may prioritize the examination items related to the examinee's gait impairment over the other examination items.

[0055] Therefore, when a subject is found to have gait impairment, the priority of upcoming examinations can be reduced based on the presence or absence of this gait impairment. Thus, when it is difficult to determine whether a gait impairment, which is a sign of stroke, is a newly discovered gait impairment, examinations other than those related to gait impairment can be prioritized.

[0056] Furthermore, for example, it is also possible that, among multiple inspection departments, the inspection department that inspects inspection items with a priority below a specified value does not perform the inspection.

[0057] Therefore, if the determined priority corresponds to the level of usefulness, the inspection department can perform the inspection of each of the multiple inspection items in descending order of usefulness. Here, usefulness refers to the usefulness of performing each inspection item for the inspected object. This allows useful inspection results to be obtained earlier after the inspection begins, thus eliminating the need to perform subsequent inspections of lower priority items (those below a specified priority value).

[0058] Furthermore, for example, it may also include a storage unit that stores at least one of the subject's history of brain disease and health diagnosis information including information related to brain disease, and the acquisition unit obtains at least one of the history of brain disease and health diagnosis information from the storage unit as personal data information.

[0059] Accordingly, at least one of the medical records and health diagnosis information obtained from the storage department can be used as personal data information.

[0060] Furthermore, for example, it may also include an identification unit to identify the subject being inspected and output identification information, and an acquisition unit to obtain personal data information corresponding to the output identification information.

[0061] Accordingly, for the inspected object identified by the identification unit, personal data information corresponding to the inspected object can be obtained.

[0062] Furthermore, personal information may also be information related to the results of preliminary examinations performed on the examinee.

[0063] Accordingly, information related to the results of preliminary examinations performed on the examinee can be used as personal data.

[0064] Furthermore, in one embodiment of the stroke examination method disclosed herein, personal data information related to the medical records of the examinee’s previous brain diseases is obtained, and based on the obtained personal data information, the priority of each of a plurality of examination items related to stroke is determined, and each of the plurality of examination items is performed in descending order of the determined priority.

[0065] This stroke examination method can achieve the same effect as the stroke examination system described above.

[0066] Furthermore, one embodiment of this disclosure relates to a program recording medium that is a computer-readable, non-transitory recording medium containing a program for causing a computer to execute the stroke examination method described above.

[0067] Such a program recording medium can achieve the same effect as the stroke examination system described above using a computer.

[0068] Furthermore, another aspect of this disclosure relates to a stroke examination system that examines stroke symptoms in a subject. The stroke examination system includes: a judgment unit that determines whether the operator of the stroke examination system is the subject; an examination unit that performs prescribed examination items related to stroke, performing the prescribed examination items in a first mode if the operator is determined to be the subject, and performing the prescribed examination items in a second mode different from the first mode if the operator is determined not to be the subject; and a diagnosis unit that outputs diagnostic information related to stroke symptoms in the subject based on the examination results.

[0069] This stroke examination system, when the judgment unit determines that the operator is the subject of examination, performs the prescribed examination items in Mode 1. Mode 1 is for the subject to operate the stroke examination system to perform the prescribed examination items themselves. When the judgment unit determines that the operator is not the subject of examination, it performs the prescribed examination items in Mode 2. Mode 2 is for an operator other than the subject to operate the stroke examination system to perform the prescribed examination items. Accordingly, the possibility of wasting examination time can be reduced. Here, wasting examination time is caused by operational errors due to not knowing whether the operation instructions for performing the examination are issued to the subject or the operator. Therefore, from the perspective of examination time, stroke examinations can be performed appropriately.

[0070] The embodiments of this disclosure will now be described with reference to the accompanying drawings.

[0071] Furthermore, the embodiments described below are all general or specific examples illustrating this disclosure. The numerical values, shapes, materials, constituent elements, the arrangement and connection methods of constituent elements, steps, and the order of steps shown in the following embodiments are all examples and are not intended to limit this disclosure. Moreover, constituent elements in the following embodiments that are not described in the independent technical solutions will be described as arbitrary constituent elements.

[0072] Furthermore, the diagrams are schematic diagrams, not rigorous illustrations. In each diagram, substantially identical components are assigned the same symbols, and repetitive explanations are omitted or simplified.

[0073] Furthermore, in this specification, terms such as parallelism indicating the relationship between elements, terms such as rectangle indicating the shape of elements, and numerical values ​​and numerical ranges are not merely rigorous expressions, but also indicate substantially equivalent ranges, for example, implying that there may be an error of a few percent.

[0074] (Implementation Method)

[0075] [constitute]

[0076] First, regarding the overview of the stroke detection system in the implementation method, using Figures 1 to 4 Please provide an explanation. Figure 1 This is a schematic diagram illustrating an example of the configuration of a stroke examination system according to an embodiment. Additionally, Figure 2 This is a block diagram illustrating the functional configuration of the stroke examination system according to the embodiment. Additionally, as an example of the functional configuration of a portable terminal device, Figure 3 This is a block diagram illustrating a case where a portable terminal device, such as a smartphone, constitutes the stroke examination system described in the embodiment.

[0077] like Figure 1 As shown, the stroke examination system 500 in this embodiment includes a stroke examination device 100 implemented by an information terminal and a server device 200.

[0078] The stroke examination device 100 includes sensors corresponding to various examination items for detecting stroke precursors. The stroke examination device 100, for the operator of the stroke examination system 500 (i.e., the operator of the stroke examination device 100), sequentially provides instructions as required for the examination and drives these sensors to perform the examination for stroke precursors in the subject. Furthermore, although the stroke examination device 100 is implemented via an information terminal, it can be other devices if it has the configuration to perform the various functions described below. For example, the stroke examination device 100 could also be a dedicated device for individuals diagnosed with a high risk of stroke during a health check, used when such individuals suspect a stroke.

[0079] Server device 200 is a device connected to stroke examination device 100 via a network such as the Internet. Here, server device 200 is a storage device used to store information used by stroke examination device 100. Server device 200 can be implemented by a cloud computer set up on a network, or by an edge computer within a local area network that stroke examination device 100 can communicate with. Furthermore, server device 200 can also be replaced by a storage unit built into stroke examination device 100 that stores the same information. In other words, stroke examination system 500 can also be implemented with only one information terminal.

[0080] like Figure 2 As shown, the stroke examination device 100 includes an acquisition unit 101, a decision unit 102, an examination unit 103, a diagnosis unit 104, a sensor unit 105, a transceiver unit 106, a judgment unit 107, an output unit 108, and a storage unit 109. In this embodiment, the acquisition unit 101, decision unit 102, examination unit 103, diagnosis unit 104, and judgment unit 107 are implemented by the CPU (Central Processing Unit) and memory of the control unit 110, and by executing a program stored in the memory.

[0081] The output unit 108 consists of a display unit and a sound output unit. The display unit consists of a display or the like, and the sound output unit consists of a speaker or the like.

[0082] The storage unit 109 serves as a memory and has functions such as storing programs executed by the CPU of the control unit 110 and personal data information related to the brain disease of the examinee.

[0083] The acquisition unit 101 is a processing unit that acquires personal data information related to the brain disease of the examinee. In this embodiment, the personal data information related to the brain disease of the examinee refers to information stored in the storage unit 201 of the server device 200. This information may be, for example, the examinee's electronic medical record card or data related to the results of a health diagnosis. Therefore, the server device 200 is a server device that is preferably installed in a medical institution. The acquisition unit 101 performs processing such as transforming the acquired personal data information and decrypting encrypted information as needed, and outputs the information to the decision unit 102 after processing the information into a usable form. Furthermore, the acquisition unit 101 acquires physical quantities detected by various sensors included in the sensor unit 105 as personal data information. In the stroke examination system 500, these acquired physical quantities are used for preliminary examinations of multiple examination items. Therefore, the acquisition unit 101 outputs the acquired physical quantities as the results of the preliminary examinations to the decision unit 102.

[0084] The decision unit 102 determines the priority of each of a pre-set set of multiple stroke-related examination items based on the personal data information output by the acquisition unit 101. The decision unit 102 can also determine the priority of each of the multiple examination items based on statistical information, which is information about the correlation between stroke and the examinee's lifestyle (place of residence, gender, age group, activity level, dietary tendencies, etc.). Specifically, the decision unit 102 can also infer the location of the stroke from the statistical information based on the examinee's lifestyle, and determine the basic priority of each examination item based on the inference result. Here, the examinee's personal data information is not reflected in the determined basic priority.

[0085] The decision unit 102 determines the final priority of each of multiple examination items by calculating priorities based on the examinee's personal information. In the final priority calculation process, if the examinee has a history of persistent symptoms from previous brain diseases, examination items that may not detect symptoms even based on new stroke precursors are given a lower priority than other examination items. Furthermore, the decision unit 102 prioritizes examination items related to the onset of symptoms based on the results of preliminary examinations included in the personal information.

[0086] For example, during a preliminary examination, if the 6-axis sensor included in the sensor unit 105 detects vibration in the stroke examination device 100, there is a possibility of Barrett's syndrome occurring. Therefore, the examination items related to Barrett's syndrome are given a higher priority than other examination items. Similarly, regarding each examination item, since the sensing for the preliminary examination is preset, the preliminary examination is performed before issuing the action instruction for the examination to the operator or the examinee. Furthermore, regarding examination items based on such a preliminary examination, if it is found that there are consistent symptoms from the examinee's previous brain diseases, the priority is reduced. In other words, in the process of increasing or decreasing priority, the process of decreasing priority is performed first. Thus, the decision unit 102 determines the final priority based on the basic priority by performing at least one of decreasing or increasing priority according to personal data information.

[0087] The inspection unit 103 is a processing unit that obtains physical quantities from various sensors included in the sensor unit 105 to inspect for stroke precursors in the subject. The sensors for obtaining physical quantities according to the inspection items and the timing are predetermined. When the inspection unit 103 obtains the physical quantity from the sensor corresponding to the performed inspection item, it outputs the inspection result corresponding to that physical quantity. Thus, the inspection unit 103 in this disclosure can be considered as multiple inspection units 103 that perform multiple inspection items respectively.

[0088] The examination results output from each examination unit 103 can be either a result indicating that a certain examination item is considered a stroke precursor or not, or a result expressed on a scale, such as an 80% probability of a stroke precursor occurring. Furthermore, based on the examination results output from each of the multiple examination units 103, a comprehensive examination result is output to the diagnostic unit. This comprehensive examination result can be, for example, an average value expressed on a scale, a value for an examination item considered a stroke precursor, or at least one of the multiple examination items being considered either a stroke precursor or not. The output of the examination results by the examination unit 103 is obtained, for example, by inputting the obtained physical quantities into a learned machine learning model. This learned machine learning model uses at least one of the physical quantities considered as stroke precursors and those not considered as stroke precursors as teacher data. Therefore, each of the multiple examination units 103 has a learned machine learning model, which is a model that has undergone optimal learning.

[0089] The diagnostic unit 104 is a processing unit that outputs diagnostic information related to the symptoms of stroke in the examined subject based on the examination results. The diagnostic information output by the diagnostic unit 104 may include, for example, at least one of the following: image information showing the examination results, or a notification message indicating that the examination results have been output externally. The image information showing the examination results output from the diagnostic unit 104 may be displayed on a smartphone screen, for example. The operator of the stroke examination system 500 can see the displayed image information and take necessary actions. Furthermore, the notification message indicating that the examination results have been output externally from the diagnostic unit 104 is sent verbatim to a medical institution via a network. The medical institution then initiates its own response based on the received notification message.

[0090] The sensor unit 105 comprises a group of various sensors included in the stroke detection device 100. The sensor unit 105 includes, for example, a camera, microphone, touch screen, fingerprint sensor, proximity sensor, GPS, 6-axis sensor (3-axis accelerometer and 3-axis angular velocity sensor), magnetic sensor, and brightness sensor.

[0091] The transceiver unit 106 is a communication module that connects the stroke examination device 100 to external devices to enable communication via a network. The transceiver unit 106 is used when the stroke examination device 100 communicates with the server device 200, and when the stroke examination device 100 communicates with a receiving device of a medical institution to receive notification information.

[0092] The judgment unit 107 is a processing unit that determines whether the operator of the stroke examination system is the subject of the examination. The judgment unit 107 makes the above judgment based on the information input to the stroke examination system 500. This operation will be described later.

[0093] Storage unit 201 is a storage device such as a semiconductor memory. Information extracted from the examinee's electronic medical record card and information extracted from the examinee's health diagnosis results are stored in storage unit 201.

[0094] The transceiver unit 202 is a communication module that connects the server device 200 to external devices to enable communication via a network. The transceiver unit 202 is used when the server device 200 communicates with the stroke detection device 100, etc.

[0095] like Figure 3As shown, a portable terminal device such as a smartphone consists of a control unit 301, a display unit 302, a storage unit 303, various sensor groups (GPS 304, 3-axis sensor 305, 3-axis angular velocity sensor 306, proximity sensor 307, magnetic sensor 308, ambient light sensor 309, and microphone 310, etc.), a camera 311, a speaker 312, a communication unit 313, a touch screen 314, a fingerprint sensor 315, a face recognition sensor group 316, a battery 317, and a power supply unit 318, etc.

[0096] The control unit 301 is capable of comprehensive control of the smartphone and includes a CPU (not shown in the accompanying drawings) and storage elements (such as SRAM). The control unit 301 and... Figure 2 The control unit 110 in the middle has at least the functions of the constituent elements such as the acquisition unit 101, the decision unit 102, the inspection unit 103, and the diagnosis unit 104.

[0097] Display unit 302 consists of Figure 2 A portion of the output unit 108 is displayed based on information received from the control unit 301.

[0098] The storage unit 303 stores the OS (Operating System), various applications, and various data used by the various applications, which are read and executed by the control unit 301. It may also store some or all of the information stored in the storage unit 201 of the server device 200, namely, personal data related to the brain disease of the examinee.

[0099] Department of Communications 313 and Figure 2 The transceiver unit 106 in the middle uses wireless communication technologies such as LTE (Long Term Evolution, registered trademark) or 5G to wirelessly connect to a communication base station (not shown) operated by a telecommunications operator, and then connects to the Internet via the communication base station. Furthermore, the communication unit for external communication is not essential to this disclosure, but it does not exclude portable terminal devices that connect via Wi-Fi (registered trademark) or wired LAN.

[0100] The 314 touchscreen can receive signals from... Figure 2 The operator of the stroke examination system 500, i.e. the operator of the stroke examination device 100, inputs the screen (display unit 302) and can send signals based on the input (e.g., which part of the screen was touched, and how much force was applied when touching) to the control unit 301.

[0101] Various sensor groups and Figure 2Corresponding to the sensor unit 105, it includes: a GPS 304 for detecting the smartphone's position on Earth; a 3-axis sensor 305 for setting X, Y, and Z axes for the smartphone and detecting the acceleration of each axis; a 3-axis angular velocity sensor 306 for detecting the angular velocity of the rotation direction for each axis set by the 3-axis sensor 305; a proximity sensor 307 for detecting approaching objects (e.g., the face of the smartphone); a magnetic sensor 308 for detecting the Earth's magnetic field and indicating orientation; and an ambient light sensor for detecting the brightness around the smartphone. A light sensor 309; a microphone 310 for collecting ambient sound and voice; a camera 311 for taking pictures of the front or back of the smartphone; a speaker 312 for emitting sound; a fingerprint sensor 315 for user authentication, etc.; and a facial recognition sensor group 316 for facial authentication (actually, it is a combination of an infrared camera, a illuminator, and a dot projector, etc., not shown in the figure, and the sensor group shown in the figure (proximity sensor 307, ambient light sensor 309, camera 311), which works as a sensor for facial authentication).

[0102] Here, utilizing Figure 4 The functions of the 3-axis sensor 305 (accelerometer) and the 3-axis angular velocity sensor are explained. Most smartphones can measure the acceleration of the device itself when accelerating in a straight line, relative to the X, Y, and Z axes (3-axis sensor). At the same time, they can also measure the acceleration acting in the direction that causes the device to rotate (expressed as X-axis angular velocity, Y-axis angular velocity, and Z-axis angular velocity; the combination of these three is called a 3-axis angular velocity sensor).

[0103] For example, when a subject uses the stroke examination device 100 to take a picture of their face, the values ​​detected by these sensors can be used to determine the angle (posture) at which the stroke examination device 100 is taking the picture of the subject's face. For example, when taking a selfie while holding the stroke examination device 100 vertically, it is possible to determine whether the stroke examination device 100 is tilted relative to the horizontal line, and further, if it is tilted, it is possible to determine the angle of tilt.

[0104] [Work]

[0105] Next, the operation of the stroke detection system 500 described above will be investigated. Figures 5 to 9B Please provide an explanation. Figure 5 This is a flowchart illustrating a working example of the stroke examination system according to the embodiment.

[0106] In this embodiment, the system is configured to output appropriate instructions even when the operator and the subject of the stroke examination system 500 are different. When the stroke examination system 500 begins operation, it first determines whether the operator and the subject are the same (S100). Figure 6 Figure 1 is an example of an operation screen of the stroke examination system according to the embodiment. This figure shows the image displayed on the screen when the procedure related to stroke examination in the stroke examination device 100 is running.

[0107] like Figure 6 As shown, during the operation of the stroke examination system 500, a selection screen is displayed to the operator of the stroke examination device 100, allowing the operator to choose whether they are in a position to examine the subject (upper half of the display screen) or the subject is in a position to perform a self-examination (lower half of the display screen). If the upper half is selected, it indicates that the operator is in a position to examine the subject; that is, the selection result "the subject and the operator are different" is input into the system. Conversely, if the lower half is selected, it indicates that the subject is in a position to perform a self-examination; that is, the selection result "the subject and the operator are the same" is input into the system. The judgment unit 107 then determines whether the operator and the subject are the same based on this input. Alternatively, the judgment unit 107 may also determine whether the operator and the subject are the same by inputting data from biometric authentication sensors, such as a fingerprint sensor, included in the sensor unit of the stroke examination device 100. Furthermore, the judgment unit 107 can also determine whether the operator and the subject of examination are the same, provided that the owner of the information terminal of the stroke examination device 100 is the operator. If it is determined that the operator and the subject of examination are not the same (S100 "No"), the process proceeds to step S201; if it is determined that the operator and the subject of examination are the same (S100 "Yes"), the process proceeds to step S101. Since steps S201 and S101 essentially perform the same work, step S101 will be explained, while the explanation of step S201 will be omitted.

[0108] In step S101, the acquisition unit 101 obtains statistical information from an external statistical information server (not shown) via the network through the transceiver unit 106. The obtained statistical information is then output to the decision unit 102. After step S101, the process proceeds to step S102. After step S201, the process proceeds to step S202. Since step S202 essentially performs the same work as step S102, step S102 will be described, while the description of step S202 will be omitted.

[0109] In step S102, the acquisition unit 101 obtains personal data information from the server device 200 via the transceiver unit 106 and the network, and also receives the personal data information from the sensor unit 105. The obtained personal data information is output to the decision unit 102. After step S102, the process proceeds to step S103. After step S202, the process proceeds to step S203. Since step S203 essentially performs the same operation as step S103, step S103 will be described, while the description of step S203 will be omitted.

[0110] In step S103, the decision unit 102 determines the priority of each of the multiple inspection items based on statistical information and personal data. The method for determining the priority is as described above. After step S103 is completed, the process proceeds to step S104. After step S203 is completed, the process proceeds to step S204.

[0111] In step S104, the decision unit 102 activates the corresponding inspection unit 103 among the multiple inspection units 103 according to the order of decision priority from high to low. At this time, in step S104, that is, when the operator and the object being inspected are the same, each inspection is performed in the first mode.

[0112] On the other hand, in step S204, the decision unit 102 activates the corresponding inspection unit 103 among the multiple inspection units 103 according to the order of decision priority from high to low. At this time, in step S204, that is, when the operator and the object being inspected are not the same, each inspection is performed in the second mode. Figure 7A Figure 2 is an example of the operation screen of the stroke examination system according to the embodiment. Additionally, Figure 7B Figure 3 is an example of an operation screen of the stroke examination system according to the embodiment. Figure 7A This shows the image displayed on the screen when performing examinations related to Barrett's syndrome in Mode 1. Additionally, Figure 7B This shows the display screen where the image is shown when the same examination items related to Barrett's syndrome are performed in mode 2.

[0113] exist Figure 7A In the process, an instruction is displayed for the operator (who is the same as the person being examined): "Please hold the terminal and extend your arm forward." Here, the terminal refers to the stroke examination device 100. Then, after this instruction is displayed, the angle of the stroke examination device 100 obtained from a 6-axis sensor or the like is detected.

[0114] On the other hand, Figure 7BIn this embodiment, an instruction is displayed for the operator (different from the subject of examination): "Please use the terminal to take a picture of the subject with their arm outstretched forward." Here, the terminal refers to the stroke examination device 100. Then, after this instruction is displayed, the image obtained from a camera or the like is examined. Thus, depending on whether the operation is performed by the operator or the subject of examination, it is necessary to use different images (instructions for the operation) and different sensors. However, in this embodiment, it is possible to appropriately distinguish between these different instructions and sensors. In this embodiment, it can be said that the use of the first examination unit corresponding to the first mode and the second examination unit in the second mode can be differentiated based on whether the operator and the subject of examination are the same.

[0115] Furthermore, as shown in steps S104 and S204, according to the determined priority from high to low, the corresponding inspection unit 103 among the multiple inspection units 103 is activated in either mode 1 or mode 2. The result of activating the inspection unit 103 is the inspection of stroke precursors performed as follows: Figure 8A as well as Figure 8B As shown. Figure 8A Figure 4 is an example of the operation screen of the stroke examination system according to the embodiment. Additionally, Figure 8B Figure 5 shows an example of the operation screen of the stroke examination system according to the embodiment. Figure 8A The diagram shows the images displayed as each of the examination items is performed sequentially from left to right in the figure. These examination items include those related to facial paralysis, dysarthria, and Barrett's sign. Additionally, Figure 8B The diagram illustrates the situation where, when each of the examinations from left to right in the diagram is performed sequentially, the image is displayed on the screen in the following circumstances: the priority of the examination item related to the subject's facial paralysis is set to be lower than the priority of the other examination items, and the priority of the examination item related to the subject's facial paralysis is below a specified value.

[0116] like Figure 8A as well as Figure 8BAs shown, since the priority of examinations related to the subject's facial paralysis is lower, the display shows the scenario where other examinations are performed first. These other examinations include those related to the subject's dysarthria and those related to the subject's Barrett's sign. Furthermore, since the priority of examinations related to the subject's facial paralysis is lower than a predetermined priority value, the display shows the scenario where this examination itself is not performed. This priority value is considered to be a value where, taking into account time or the impact on the results, it is better to suspend the examination and output the diagnostic result first. In this way, the output of the diagnostic result can be made faster by suspending examinations with lower priority as needed. Additionally, the predetermined values ​​here can be set based on experiments or experience.

[0117] Furthermore, in this embodiment, there are cases where a camera is used as a sensor to acquire images. In such cases, there may be situations where an inappropriate image cannot be obtained due to symptoms such as Barrett's syndrome. For example, if the stroke examination device 100 cannot be held sufficiently, the orientation of the subject may rotate within the image plane or within a plane intersecting the image plane. In such cases, the image processing unit (not shown) assembled in the examination unit 103 can rotate the image and simulate the generation of a normal image. This configuration will be described later.

[0118] Additionally, there are cases where the image appears as described above simply because the user is unfamiliar with operating the stroke examination device 100. In such cases, the displayed message can be changed to rotate the stroke examination device 100 and capture an appropriate image. Alternatively, a rotating analog marker (a clear 3D image showing the faces and orientations of a coin or dice, etc.) linked to the 6-axis sensor can be displayed on the screen. For example, instructions such as "Please rotate the terminal so that the heads of the coin are facing you" can be given to adjust the orientation of the analog marker, thus rotating the stroke examination device 100 to achieve the aforementioned purpose.

[0119] return Figure 5 After step S104 or step S204 is completed, step S105 is performed. In step S105, the diagnostic unit 104 outputs diagnostic information to the display screen and external devices.

[0120] In this way, multiple examination items can be performed in an appropriate order, and appropriate operating instructions can be given for each examination item based on whether the operator and the examinee are consistent. Therefore, the stroke examination system 500 can quickly obtain examination results while reducing time loss.

[0121] (Regarding examinations related to facial paralysis in the examinee)

[0122] The following describes a method for detecting facial paralysis when performing the above-described embodiments.

[0123] Figure 9A Figure 6 is an example of an operation screen of the stroke examination system according to the embodiment. Additionally, Figure 9B Figure 7 is an example of the operation screen of the stroke examination system according to the embodiment. Additionally, Figure 10 This paper illustrates a method for detecting facial paralysis using a neural network called deep learning. The following data sets are input as teacher data into a multi-level neural network with intermediate layers: facial images with facial paralysis and data sets showing the correct interpretation of the paralysis corresponding to each image; and facial images without facial paralysis and data sets showing the correct interpretation of the non-paralysis corresponding to each image. Figure 10 In the middle, the first and second images from top to bottom show facial paralysis, while the third image does not. At this point, through backpropagation algorithms and other operations, the weight coefficients of each node in the intermediate layer are adjusted to suit the data set.

[0124] By performing this operation, the neural network learns the features of facial paralysis and the features of facial paralysis. Once a facial image that has not been used for learning is input, it can distinguish whether the image has facial paralysis or not.

[0125] like Figure 10 As shown, when enabling the neural network to learn, positive images are desired as teacher data. Here, Figure 11A An example of a neural network is shown that takes into account the effects of facial image rotation during the capture for checking facial paralysis when implementing the embodiment. For example... Figure 11A As shown, for images or images of a stroke patient's face that are prone to skewness (for example, if there is residual paralysis in the arm, the stroke examination device 100 may rotate to the left or right, resulting in rotation of the image or image), the neural network learns using corrected teacher data. Furthermore, Figure 11B An example of a neural network is shown that, when implementing an embodiment, takes into account the effects of the skew between the upper (forehead) and lower (jaw) parts of the face in the facial image taken to check for facial paralysis. For example... Figure 11B As shown, a scenario is envisioned in which a patient has difficulty keeping the stroke examination device 100 in a vertical position, and facial images of the upper part (forehead) and lower part (jaw) of the face are prepared in advance as teacher data for the neural network to learn.

[0126] Thus, if multiple neural networks are generated using facial image data corresponding to the anticipated distortion that occurs during shooting, then when detecting facial paralysis, by inputting the image that will be the object of detection into each of these multiple neural networks, it is possible to properly detect the presence or absence of facial paralysis even if the face reflected in the image is not photographed from the front.

[0127] Furthermore, while the example described above illustrates how teacher data, which has been designed and corrected from different photographic angles, is individually input into each of multiple neural networks for learning, it is also possible to input all teacher data, including the corrected teacher data, into a single neural network to generate a single neural network.

[0128] In the above explanation, although using, as Figure 11A The neural network shown corrects for the rotation of the stroke examination device 100, but is not limited to it. The stroke examination device 100 can also obtain a tilt angle relative to the horizontal based on the sensing values ​​output by the sensor unit 105 (3-axis sensor 305, 3-axis angular velocity sensor 306) of the stroke examination device 100 during image capture. This angle is, for example, as shown in... Figure 9A The angle α shown is the angle difference, which is the angle difference between the vertical direction of the subject and the vertical direction of the stroke examination device 100.

[0129] Here, Figure 12 An example is shown where a tilt correction process is performed when capturing an image of the subject's face during the implementation of an embodiment. For example... Figure 12 As shown, angle α can be calculated based on the sensing value output by sensor unit 105 during image capture. Alternatively, the sensing value during face capture can be retained and then calculated based on the sensing value when restoring the tilted face image using an image rotation processing unit (e.g., affine transformation). Alternatively, angle α can be calculated based on the sensing value during face capture and input to a facial paralysis detection unit (e.g., before recording the face image into the stroke examination device 100 or into the facial paralysis detection unit (e.g., [missing information]). Figure 10 Before the neural network shown is capable of detecting facial paralysis, tilt is corrected in the image rotation processing unit. These correction processes can be performed in the stroke examination device 100 or sent to the server device 200 for cloud-based processing. Furthermore, it is also possible to perform... Figure 9B The same processing is applied to the angle difference between the orientation of the subject in the horizontal plane of the stroke examination device 100 and the orientation of the stroke examination device 100.

[0130] (Examinations related to Barrett's syndrome in the examinee)

[0131] The following describes a method for detecting Barrett's symptoms when performing the above-described implementation.

[0132] When a doctor diagnoses numbness in the arm, the subject is asked to raise both arms forward and is assessed on whether they can maintain this position for a specified time (e.g., 5 seconds).

[0133] It is suggested to use a method that utilizes a terminal device to perform examinations such as Barrett's sign (see, for example, Patent Document 1). In this method: the subject is held with both arms extended forward, and the scene is captured on video. Using the position of the wrist, elbow, and shoulder of one arm as a reference, the other arm is checked for drooping.

[0134] In this embodiment, regarding the horizontal position of the arm, positional information of both healthy individuals and those with arm paralysis can be accumulated, and these individuals and those with arm paralysis symptoms can be used as learning data for the neural network to learn. Alternatively, appropriate thresholds can be set based solely on the elbow, shoulder, and wrist positions (in this case, what changes occur in the elbow position relative to the shoulder and wrist positions, or, in a "lined" forward-extended arm posture, what threshold is shown for the wrist position from its initial downward drooping position), and a judgment can be made. In either case, a judgment can be made by replacing the diagnostic processing of a doctor with dynamic images captured by a camera and image analysis derived from those dynamic images.

[0135] However, this method requires a third party to act as the operator of the photographing of the subject, or the camera to be fixed on a designated platform so that a self-portrait can be taken from a shooting angle that includes the subject's arm.

[0136] Next, the method for detecting Barrett's signs in the subject will be explained using the 3-axis sensor 305 of the stroke examination device 100. Figure 13 This illustrates the process of taking a picture of the subject's face while holding the stroke detection device with their arm extended, simultaneously performing examinations for Barrett's sign and facial paralysis. Figure 13As shown, the subject holds the stroke examination device 100 with one or both hands, and maintains the arm or arms holding the device in a forward-extended position for a certain period of time. If paralysis occurs in either arm, the paralyzed arm will droop, causing the stroke examination device 100 to tilt relative to a horizontal position. This tilt angle is determined based on a predetermined threshold. Alternatively, the sensory values ​​of healthy individuals and individuals with arm paralysis performing the same action can be used as teacher data to train the neural network. The detected sensory values ​​are input into the trained neural network, and the system outputs whether arm paralysis is present or absent.

[0137] Furthermore, since the arm paralysis is detected while the subject maintains their position on the stroke examination device 100 with one or both arms horizontally extended for a specified time, and the image from the stroke examination device 100 is viewed with the arm extended, facial paralysis can be detected simultaneously by capturing the subject's face with the camera. By simultaneously detecting Barrett's sign and facial paralysis, it is expected that the detection time for TIA (Transient Ischemic Attack), a type of stroke precursor, can be shortened.

[0138] (Regarding the examination of the examinee's articulation disorders)

[0139] The following describes a method for detecting articulation disorders when performing the above-described embodiments.

[0140] In this embodiment, the examination of articulation disorders means a speech test, which refers to an examination of whether a given text can be spoken fluently. The speech data of the subject repeatedly reading the prepared text is stored in the storage unit 303 via microphone 310. The prepared text is preferably a prescribed text that provides a standard speech sample, or it is preferably a text that repeatedly utters plosive sounds (“PA”, “KA”, and “TA”) for a specified duration. Speech recognition technology is used to determine whether the subject's speech is good or bad. At this time, as preprocessing, speech and other sound detection, statistical analysis of speech data, and signal filtering processing for feature extraction may also be performed. As an example, the raw speech data and / or any derived features are provided as input to a neural network for further feature extraction.

[0141] The assessment of articulation disorders involves having the subject read a prepared text aloud. The subject's speech data is then input into a neural network. If the speech is "clear and fluent," it is judged as normal; if the speech contains "several unclear pronunciations," it is judged as mild to moderate articulation disorder; and if the speech is "so unclear that it is difficult to understand, or if the subject is unable to produce sound," it is judged as severe articulation disorder.

[0142] (Other implementation methods)

[0143] Although the implementation methods have been described above, this disclosure is not limited to the above implementation methods.

[0144] Furthermore, although the above embodiments have exemplified the constituent elements of a stroke examination system, the functions of each constituent element of a stroke examination system can be arbitrarily assigned to multiple parts constituting a stroke examination system.

[0145] Furthermore, in the above embodiments, each component can be implemented by executing a software program suitable for that component. Each component can also be implemented by a program execution unit such as a CPU or processor reading and executing a software program recorded on a recording medium such as a hard disk or semiconductor memory.

[0146] Furthermore, each component can be implemented in hardware. Each component can be, for example, a circuit (or integrated circuit). These circuits can form a single circuit as a whole, or they can each form a separate circuit. Moreover, each of these circuits can be a general-purpose circuit or a special-purpose circuit.

[0147] Furthermore, the general or specific forms of this disclosure can also be implemented by a system, apparatus, method, integrated circuit, computer program, or a computer-readable CD-ROM or other recording medium. It can also be implemented by any combination of the system, apparatus, method, integrated circuit, computer program, and recording medium.

[0148] Furthermore, any modifications that can be conceived by those skilled in the art to be applied to this embodiment, or any combination of constituent elements and functions in the embodiment without departing from the spirit of this disclosure, are all included within the scope of this disclosure.

[0149] Industrial applicability

[0150] This disclosure is useful in performing appropriate examinations for stroke.

[0151] Symbol Explanation

[0152] 100 Stroke Detection Device

[0153] 101 Acquisition Department

[0154] 102 Decision Department

[0155] 103 Inspection Department

[0156] 104 Diagnostic Department

[0157] 105 Sensor Department

[0158] 106 Receiving and Dispatch Department

[0159] 107 Judgment Department

[0160] 108 Output Section

[0161] Storage Departments 109, 201, and 303

[0162] 110, 301 Control Department

[0163] 200 server devices

[0164] 202 Receiving and Dispatch Department

[0165] 302 Display Section

[0166] 305 3-axis sensor

[0167] 306 3-axis angular velocity sensor

[0168] 310 microphone

[0169] 311 camera

[0170] 312 speakers

[0171] 313 Ministry of Communications

[0172] 500 Stroke Detection System

Claims

1. A stroke detection system for examining stroke signs in subjects. The stroke detection system has the following features: The acquisition department obtains at least personal data information related to the medical records of the subject's previous brain diseases; The decision-making unit determines the priority of each of the multiple examination items related to stroke based on the obtained personal data information. Among the priorities determined by the decision-making unit, the examination item that finds that the subject has constant symptoms is given a lower priority than the other examination items. Multiple inspection departments, each of which performs an inspection of each of the multiple inspection items, wherein the multiple inspection departments perform the inspection of the multiple inspection items in descending order of determined priority; as well as The diagnostic department, based on the examination results, outputs diagnostic information related to stroke symptoms for the examined subject. Among the plurality of inspection departments, the inspection department that inspects inspection items with a priority value below a specified value does not perform the inspection.

2. The stroke detection system as described in claim 1, The plurality of examination items are at least one of the following: examination items related to facial paralysis of the subject, examination items related to Barrett's syndrome of the subject, examination items related to dysarthria of the subject, and examination items related to gait impairment of the subject.

3. The stroke detection system as described in claim 1, When the personal data includes the examinee's medical records showing specific symptoms, the decision-making unit prioritizes the examination items related to the examinee's specific symptoms over the other examination items.

4. The stroke detection system as described in claim 1, When the personal data includes the examinee's medical records showing specific symptoms, the decision-making unit prioritizes the examination items related to the examinee's specific symptoms over the other examination items.

5. The stroke detection system as described in claim 4, When the personal data includes the examinee's medical history of facial paralysis, the decision-making unit prioritizes the examination items related to the examinee's facial paralysis among the multiple examination items, lowering their priority compared to the other examination items.

6. The stroke detection system as described in claim 4, When the personal data includes the examinee's medical records showing symptoms of Barrett's disease, the decision-making unit prioritizes the examination items related to the examinee's symptoms of Barrett's disease among the multiple examination items, lowering their priority compared to the other examination items.

7. The stroke detection system as described in claim 4, When the personal data includes the examinee's medical history of dysarthria, the decision-making unit prioritizes the examination items related to the examinee's dysarthria among the multiple examination items as lower than the priority of the other examination items.

8. The stroke detection system as described in claim 4, When the personal data includes the examinee's medical history of gait impairment, the decision-making unit prioritizes the examination items related to the examinee's gait impairment among the multiple examination items as lower than the priority of the other examination items.

9. The stroke detection system as described in claim 1, The stroke examination system further includes a storage unit that stores at least one of the following: the subject's history of brain disease and health diagnostic information including information related to brain disease. The acquisition unit obtains at least one of the disease history and the health diagnosis information from the storage unit as the personal data information.

10. The stroke detection system as described in claim 1, The stroke examination system further includes an identification unit to identify the subject of examination and output identification information. The acquisition unit obtains the personal data information corresponding to the output identification information.

11. The stroke detection system according to any one of claims 1 to 10, The personal data information further includes information related to the results of preliminary examinations performed on the subject of the examination.

12. A method for examining stroke, wherein in the method for examining stroke, The computer performs the following processing: Obtain personal data related to the examinee's past medical records of brain diseases. Based on the obtained personal data, the priority of each of the multiple examination items related to stroke is determined in such a way that the examination items that reveal that the subject already has consistent symptoms have a lower priority than other examination items. The checks are performed on each of the multiple check items in descending order of priority. in, Inspections will not be performed on items whose priority is below a specified value among the multiple inspection items.

13. A program recording medium, The program recording medium contains a program that causes a computer to execute the stroke examination method of claim 12.

14. A stroke detection system for examining stroke signs in a subject. The stroke detection system has the following features: The acquisition department obtains at least personal data information related to the medical records of the subject's previous brain diseases; The decision-making unit determines the priority of each of the multiple examination items related to stroke based on the obtained personal data information. Among the priorities determined by the decision-making unit, the examination item that finds that the subject has constant symptoms is given a lower priority than the other examination items. The judgment unit determines whether the operator of the stroke examination system is the subject of the examination. Multiple inspection units, each of which performs an inspection of each of the multiple inspection items, wherein if the operator is determined to be the inspected object, each of the multiple inspection units performs an inspection of each of the multiple inspection items in a first mode in descending order of determined priority, wherein the first mode is a mode for issuing operation instructions to the inspected object; if the operator is determined not to be the inspected object, a second mode, different from the first mode, performs an inspection of each of the multiple inspection items in descending order of determined priority, wherein the second mode is a mode for issuing operation instructions to the operator other than the inspected object; as well as The diagnostic department, based on the examination results, outputs diagnostic information related to stroke symptoms for the examined subject. Among the plurality of inspection departments, the inspection department that inspects inspection items with a priority value below a specified value does not perform the inspection.

Images