Detection system

Abstract

To provide a detection system capable of accurately distinguishing data arising from teeth grinding from data not arising from teeth grinding.SOLUTION: A detection system includes a detector for sensing teeth grinding, and an information processor which determines the existence of teeth grinding based on the data indicating results of sensing by the detector. The detector comprises: a pressure sensor provided on a mouthpiece; a myoelectric sensor attached to a human's cheek; and a controller which outputs data obtained by synchronizing the output from the pressure sensor with the output from the myoelectric sensor.SELECTED DRAWING: Figure 9

Description

[Technical field]

[0001] The present disclosure relates to detection systems. [Background technology]

[0002] Known methods for detecting human teeth grinding include a method that uses a mouthpiece equipped with a pressure sensor (eg, Patent Document 1) and a method that uses an electromyographic sensor (eg, Patent Document 2). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 6634567 [Patent Document 2] Patent No. 6618482 Summary of the Invention [Problem to be solved by the invention]

[0004] Whether using only a pressure sensor or only an electromyographic sensor, there were cases where the data obtained from the sensor could not be distinguished between data caused by teeth grinding and data not caused by teeth grinding. For example, with a pressure sensor, it was difficult to distinguish between pressure caused by teeth grinding and pressure caused by the tongue or lips hitting the pressure sensor. Also, with an electromyographic sensor, closing the eyelids tightly sometimes produced an output similar to that of teeth grinding.

[0005] The present disclosure has been made in consideration of the above-mentioned problems, and aims to provide a detection system that can distinguish between data due to teeth grinding and data that is not due to teeth grinding with higher accuracy. [Means for solving the problem]

[0006] A detection system according to one aspect of the present disclosure includes a detection device that performs sensing regarding teeth grinding, and an information processing device that determines the presence or absence of teeth grinding based on data indicating the results of sensing by the detection device, wherein the detection device includes a pressure sensor disposed in a mouthpiece, an electromyographic sensor attached to a person's cheek, and a control unit that outputs data that synchronizes the output of the pressure sensor and the output of the electromyographic sensor. [Brief description of the drawings]

[0007] [Figure 1] FIG. 1 is a block diagram showing the main configuration of a detection system having a detection device. [Diagram 2] FIG. 2 is a diagram showing an example of the outer shape of the detection device. [Diagram 3] FIG. 3 is a schematic exploded perspective view showing an example of the configuration of main components of the first portion. [Figure 4] FIG. 4 is a schematic side view showing a mechanism for bonding the first component and the second component. [Diagram 5] FIG. 5 is a schematic plan view showing an example of the configuration of main components of the first portion. [Figure 6] FIG. 6 is a schematic diagram showing an example of how the detection device is attached to a human. [Figure 7] FIG. 7 is a schematic diagram showing the placement of the first portion in the mouth shown in FIG. [Figure 8] FIG. 8 is a schematic diagram showing the attachment positions of the first electrode, the second electrode, and the third electrode. [Figure 9] FIG. 9 is a diagram showing an example of a time-series relationship between a human bite force indicated by an output from a pressure sensor and a myoelectric potential indicated by an output from an electromyographic sensor. [Figure 10] FIG. 10 is a diagram showing a first pattern, a second pattern, and a third pattern as examples of outputs that occur when there is no bruxism. [Figure 11] FIG. 11 is a flowchart showing an example of a process flow related to the operation of the detection device. [Figure 12]FIG. 12 is a flowchart showing an example of a process flow related to the operation of the detection device, which is different from the flow shown in FIG. [Figure 13] FIG. 13 is a flowchart showing an example of a process flow related to the operation of the detection device, which is different from the examples shown in FIGS. [Figure 14] FIG. 14 is a schematic diagram illustrating an example of the teacher data. [Figure 15] FIG. 15 is a schematic diagram illustrating an example of feature amount data. [Figure 16] FIG. 16 is a two-dimensional graph that shows a schematic diagram of machine learning for classifying the presence or absence of bruxism. [Figure 17] FIG. 17 is a two-dimensional graph that shows a schematic representation of machine learning for classifying the type of bruxism by a subject. [Figure 18] FIG. 18 is a schematic two-dimensional graph showing machine learning relating to classification of types of bruxism by a subject different from that shown in FIG. [Figure 19] FIG. 19 is a diagram showing an example of the relationship between the recognition target data and the recognition information indicating the output of the recognition unit based on the recognition target data. [Figure 20] FIG. 20 is a flowchart showing the flow of processing related to generation of model data. [Figure 21] FIG. 21 is a flowchart showing the flow of processing related to estimation of the occurrence of bruxism based on the discrimination target data. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Hereinafter, each embodiment of the present disclosure will be described with reference to the drawings. Note that the disclosure is merely an example, and those who are skilled in the art can easily come up with appropriate modifications while maintaining the gist of the invention are naturally included in the scope of the present disclosure. In addition, in order to make the explanation clearer, the drawings may be schematic in terms of the width, thickness, shape, etc. of each part compared to the actual embodiment, but they are merely examples and do not limit the interpretation of the present disclosure. In addition, in this specification and each figure, elements similar to those described above with respect to the previous figures may be given the same reference numerals, and detailed explanations may be omitted as appropriate.

[0009] 1 is a block diagram showing a main configuration of a detection system 100 having a detection device 10. The detection system 100 has the detection device 10 and an information processing device 60.

[0010] FIG. 2 is a diagram showing an example of the outer shape of the detection device 10. As shown in FIG. 2, the detection device 10 has a first portion P1, a second portion P2, and a third portion P3. The first portion P1 includes a mouthpiece 21, a first sensor element 221, a second sensor element 222, a third sensor element 223, an FPC 25, and the like. A more detailed configuration of the first portion P1 will be described later in the description of the pressure sensor 20. The second portion P2 includes an FPC 30. The third portion P3 includes a housing 41, a first electrode 42, a second electrode 43, and a third electrode 44. The first electrode 42, the second electrode 43, and the third electrode 44 are made of, for example, silver-silver chloride (Ag-AgCl), but may also be made of gold, platinum, silver, carbon, or the like.

[0011] 1 includes a sensor unit 11, a control unit 12, and a communication unit 13. The sensor unit 11 includes a pressure sensor 20 and an electromyographic sensor 40. The pressure sensor 20 and the electromyographic sensor 40 will be described below in order. First, the pressure sensor 20 will be described with reference to FIGS. 3 to 7.

[0012] 3 is a schematic exploded perspective view showing an example of the main component configuration of the first portion P1. The mouthpiece 21 is formed by bonding a first component 21a and a second component 21b together. A first sensor element 221, a second sensor element 222, a third sensor element 223, and an FPC 25 are sandwiched between the first component 21a and the second component 21b.

[0013] The first sensor element 221, the second sensor element 222, and the third sensor element 223 are configured to generate an electrical effect in response to pressure. As a specific example, the first sensor element 221, the second sensor element 222, and the third sensor element 223 are piezoelectric sensors. As a more specific example, thin plate-like ceramic elements that generate a piezoelectric effect are arranged as the first sensor element 221, the second sensor element 222, and the third sensor element 223. The first sensor element 221, the second sensor element 222, and the third sensor element 223 individually generate electric charges in response to the pressure applied thereto.

[0014] The FPC 25 is a flexible thin-film substrate on which wiring is formed. As a specific example, the FPC 25 is a flexible printed circuit board (FPC: Flexible Printed Circuits), but may be another structure that functions in a similar manner. The first sensor element 221, the second sensor element 222, and the third sensor element 223 are wirings formed on the FPC 25, and are each connected to individual wirings.

[0015] Fig. 4 is a schematic side view showing a mechanism for bonding the first component 21a and the second component 21b. The sensor element 22 shown in Fig. 4 is any one of the first sensor element 221, the second sensor element 222, and the third sensor element 223. Hereinafter, when the sensor element 22 is described, it is a description that collectively includes the first sensor element 221, the second sensor element 222, and the third sensor element 223.

[0016] 4, one surface of the FPC 25 is bonded to the first component 21a via an adhesive 26, and the other surface of the FPC 25 is bonded to the second component 21b. In the example shown in FIG. 4, the sensor element 22 is provided on the other surface of the FPC 25, but may be provided on one surface of the FPC 25.

[0017] Mouthpiece 21 is a mouthpiece that corresponds to the curved shape of the bifurcated portion of FPC 25 where first sensor element 221, second sensor element 222, and third sensor element 223 are provided. As shown in Fig. 4, FPC 25 is bonded and fixed inside mouthpiece 21 so that a gap is generated between FPC 25 and mouthpiece 21.

[0018] Mouthpiece 21 is formed, for example, using a resin having a certain degree of flexibility (for example, resin or silicone), but is not limited thereto, and the specific composition can be appropriately changed as long as it is configured to transmit pressure to first sensor element 221, second sensor element 222, and third sensor element 223. Mouthpiece 21 is desirably manufactured by obtaining a tooth shape for each wearer and molding it.

[0019] 5 is a schematic plan view showing an example of the main components of the first portion P1. As shown in FIGS. 2, 3, 5, and 7, the FPC 25 has a forked shape at one end where the first sensor element 221 and the second sensor element 222 are provided. The first sensor element 221 and the second sensor element 222 are provided near the tip of the first portion P1. The FPC 25 has a single-strand shape at the end of the first portion P1 that is continuous with the FPC 30 of the second portion P2. The third sensor element 223 is provided near the connection position between the forked portion and the single-strand portion.

[0020] As shown in FIG. 5, for example, most of the adhesive 26 is applied to the area between the first sensor element 221, the second sensor element 222, and the third sensor element 223 that are arranged in the forked portion, but this is not limited to this, and it is sufficient that the adhesive 26 is applied in such a manner that the integrity of the mouthpiece 21 and the FPC 25 is maintained.

[0021] 6 is a schematic diagram showing an example of how the detection device 10 is attached to a human. As shown in FIG. 6, when the detection device 10 is attached to a human, the housing 41 of the third part P3 is attached so that the surface on which the first electrode 42, the second electrode 43, and the third electrode 44 are exposed is close to or in contact with the outer side of the cheek of the human HF. Specifically, for example, conductive gels 42a, 43a, and 44a are provided on the first electrode 42, the second electrode 43, and the third electrode 44 of the housing 41, and a double-sided tape 45 for skin is provided on the other parts of the housing 41, making it possible to attach the housing 41 to the human HF. The material of the double-sided tape is preferably silicone.

[0022] In the embodiment, a double-sided tape 45 for skin is attached to the surface of the housing 41 on the side of the first electrode 42, the second electrode 43, and the third electrode 44. The double-sided tape 45 for skin is attached over one surface of the housing 41 around the first electrode 42, the second electrode 43, and the third electrode 44. The conductive gels 42a, 43a, and 44a are provided so as to cover the outer electrode surfaces of the first electrode 42, the second electrode 43, and the third electrode 44, respectively. When the detection device 10 is attached to a human, the first portion P1 is included in the mouth M of the human. The FPC 30 of the second portion P2 is extended to a length that can connect the first portion P1 and the housing 41 when the detection device 10 is attached to a human.

[0023] The FPC 30, which is a component provided in the second portion P2, is an FPC that continues from the end side of the first portion P1 of the FPC 25. The FPC 30 functions as wiring that connects the first sensor element 221, the second sensor element 222, and the third sensor element 223 with a circuit provided in the housing 41 of the third portion P3. A circuit that converts the output of each of the first sensor element 221, the second sensor element 222, and the third sensor element 223 into individually manageable digital data and the like are provided in the housing 41. The pressure sensor 20 of the embodiment includes the circuit and the like in its configuration.

[0024] 7 is a schematic diagram showing the arrangement of the first portion P1 in the mouth M shown in FIG. 6. As shown in FIG. 7, the first portion P1 is held by a person so that the first sensor element 221 is located near one of the molars M1, which is the innermost tooth in the mouth M, the second sensor element 222 is located near the other molar M2, which is the innermost tooth in the mouth M, and the third sensor element 223 is located near the incisor M3, which is the frontmost tooth in the mouth M. In other words, the shape and extension length of the bifurcated mouthpiece 21 and FPC 25 in the first portion P1 and the arrangement of the first sensor element 221, the second sensor element 222, and the third sensor element 223 are determined in advance so that such a positional relationship between the molars M1, M2, and the incisor M3 and the first sensor element 221, the second sensor element 222, and the third sensor element 223 is established. Of course, since there may be individual differences in the size of a human mouth M and the positions of the molars M1, M2 and incisors M3 in the mouth M, the first part P1 may be customized for each person, or the design of the first part P1 assuming a relationship with a model of a general adult human tooth may be adopted as a standard design. The first part P1 is for the upper jaw and is attached to the teeth of the upper jaw. However, it may also be for the lower jaw, which is attached to the teeth of the lower jaw. Strictly speaking, whether it is for the upper jaw or the lower jaw is determined for each individual person based on the judgment of a dentist. In this way, the mouthpiece 21 may be attached to either the upper or lower jaw inside the human mouth M.

[0025] When the upper and lower teeth of the jaw move to collide with each other at the position of the molar M1, the pressure caused by the collision is detected by the first sensor element 221. When the upper and lower teeth of the jaw move to collide with each other at the position of the molar M2, the pressure caused by the collision is detected by the second sensor element 222. When the upper and lower teeth of the jaw move to collide with each other at the position of the incisor M3, the pressure caused by the collision is detected by the third sensor element 223. Therefore, in the embodiment, the first sensor element 221, the second sensor element 222, and the third sensor element 223 act as a pressure sensor 20 that functions in the human mouth.

[0026] The shapes of the mouthpiece 21 and the FPC 25 shown in Figs. 2 and 7 and the shapes of the mouthpiece 21 and the FPC 25 shown in Figs. 3 to 5 may be adopted. The shapes of the components of the first part P1 shown in Figs. 3 to 5 are schematic for explaining the relative positional relationship between the components of the first part P1, and do not show a specific form. The specific form of the first part P1 is not limited to the form shown in Figs. 2 and 7, and may be a mouthpiece form in which a pressure sensor capable of detecting pressure according to the occlusion of teeth in a human mouth can be arranged. In addition, although the case where three sensor elements 22 functioning as the pressure sensor 20 are illustrated here, pressure sensors corresponding to each of the teeth may be arranged. For example, if there are 14 teeth in the upper jaw, 14 pressure sensors may be arranged.

[0027] Next, the myoelectric sensor 40 shown in FIG. 1 will be described with reference to FIGS.

[0028] FIG. 8 is a schematic diagram showing the attachment positions of the first electrode 42, the second electrode 43, and the third electrode 44. In the third part P3, the housing 41 is attached to the human HF so that the first electrode 42, the second electrode 43, and the third electrode 44 are aligned along the direction D in the area W shown in FIG. 8. The area W is an area located near the ear Y in the vicinity of the cheek of the human HF between the mouth M and the ear Y of the human HF. The direction D is, for example, a direction that is approximately perpendicular to a straight line connecting the ear Y and the mouth M, and approximately overlaps with a straight line from the corner of one of the two eyes of the human HF to the neck. For example, as described with reference to FIG. 6 above, the housing 41 of the third part P3 is attached so that the surface on which the first electrode 42, the second electrode 43, and the third electrode 44 are exposed is placed close to or in contact with the cheek of the human HF, thereby realizing an arrangement in which the first electrode 42, the second electrode 43, and the third electrode 44 are aligned along the direction D in the area W.

[0029] The above-mentioned features of the area W and the direction D are merely general examples, and may be adjusted for each individual person. It is preferable that such adjustment is performed on the assumption that the positions of the first electrode 42, the second electrode 43, and the third electrode 44 of the third portion P3 overlap with the shallow part (shallow layer) of the human masseter muscle with high accuracy.

[0030] The myoelectric sensor 40 shown in FIG. 1 includes a function as a differential amplifier using a first electrode 42, a second electrode 43, and a third electrode 44. One of the first electrode 42, the second electrode 43, and the third electrode 44 is a reference electrode (REF). One of the first electrode 42, the second electrode 43, and the third electrode 44 that is not a reference electrode is an exploring electrode. One of the first electrode 42, the second electrode 43, and the third electrode 44 that is not a reference electrode or an exploring electrode is a ground electrode (GND). The ground electrode is an electrode that indicates a ground potential (0V). Ideally, it is desirable for the reference electrode to indicate the same potential (0V) as the ground potential, but in reality, it is an electrode that indicates a potential that is affected by a noise component (noise) at the position where the third portion P3 is arranged. The exploring electrode is an electrode that is arranged in the area W to capture a myoelectric potential and indicate a potential corresponding to the myoelectric potential. The myoelectric potential refers to an electrical signal generated in a muscle corresponding to the superficial part (superficial layer) of the above-mentioned human masseter muscle.

[0031] In reality, the above-mentioned noise may be mixed into the potential of the searching electrode. Therefore, in the embodiment, the reference electrode detects the same noise, and the potential corresponding to the myoelectric potential can be detected with higher accuracy based on the difference between the potential of the searching electrode and the potential of the reference electrode. More specifically, an electric signal corresponding to the difference between the potential of the searching electrode and the potential of the ground electrode is input to one of two inputs of an amplifier composed of an operational amplifier or the like. Also, an electric signal corresponding to the difference between the potential of the reference electrode and the potential of the ground electrode is input to the other of the two inputs of the amplifier. As a result, the electric signals of the same phase among the two inputs of the amplifier cancel each other out, and signals that are different between one and the other, such as reverse phase signals, are amplified. Therefore, according to the myoelectric sensor 40 of the embodiment, the myoelectric potential can be captured with higher accuracy.

[0032] A circuit for converting the output of the myoelectric sensor 40 serving as the amplifier described above into digital data and the like are provided inside the housing 41. The myoelectric sensor 40 of the embodiment includes the circuit and the like in its configuration.

[0033] 1 performs processing based on a signal output from the sensor unit 11. The processing performed by the control unit 12 will be described below with reference to FIGS. 1 and 9 to 13.

[0034] FIG. 9 is a diagram showing an example of a time-series relationship between the bite force of a person indicated by the output of the pressure sensor 20 and the myoelectric potential indicated by the output of the myoelectric sensor 40. Of the four lines in the graph shown in FIG. 9, the lower three lines L2, L3, and L4 indicate the strength of the output of the pressure sensor 20, i.e., the strength of the bite force of a person. In addition, the lower three lines L2, L3, and L4 each indicate the individual output of the first sensor element 221, the second sensor element 222, and the third sensor element 223. Here, the line L2 indicates the output of the first sensor element 221, the line L3 indicates the output of the second sensor element 222, and the line L4 indicates the output of the third sensor element 223. Of the four lines, the upper line L1 indicates the strength of the output of the myoelectric sensor 40, i.e., the strength of the myoelectric potential. In Fig. 9, the dashed lines between the three lower lines L2, L3, and L4 and the upper line L1 separate the range of the graph showing the bite force from the range of the graph showing the myoelectric potential. The vertical axis of the graphs in Fig. 9 and Fig. 10, which will be described later, indicates the strength of the output. The horizontal axis of the graphs in Fig. 9 and Fig. 10 indicates the passage of time.

[0035] During periods T1 and T2 in the graph shown in FIG. 9, one of the three lines L2, L3, and L4 at the bottom, which indicate the strength of the bite force, (line L2) indicates a significantly stronger bite force than the other two lines (lines L3 and L4). This indicates that a situation occurs in which the upper and lower teeth of the jaw are close enough to abut at one of the positions of the molar M1, molar M2, or incisor M3, and a pressing force is applied to the sensor element 22. If line L2 indicates the output of the first sensor element 221, then a situation occurs in which the upper and lower teeth of the jaw are close enough to abut at the molar M1, and a pressing force is applied to the sensor element 22. This situation can occur when teeth grinding occurs. In addition, the upper line L1, which indicates the strength of the myoelectric potential, also shows significantly larger fluctuations in output during periods T1 and T2 compared to other periods. This suggests that the movement of the masseter muscle, that is, a force in the direction of drawing the lower jaw to the upper jaw, is acting. Such an indication provides more reliable confirmation that bruxism is occurring.

[0036] In this manner, in the embodiment, the occurrence of bruxism can be identified based on data in which the output of the pressure sensor 20 and the output of the myoelectric sensor 40 are synchronized in time. In contrast, it is difficult to accurately identify the occurrence of such bruxism using only one of the output of the pressure sensor 20 or the output of the myoelectric sensor 40. Below, a reference example in which a false detection occurs using only one of the output of the pressure sensor 20 or the output of the myoelectric sensor 40 will be described with reference to FIG.

[0037] FIG. 10 shows a first pattern, a second pattern, and a third pattern as examples of outputs that occur when there is no teeth grinding. The first pattern is a pattern when a person tightly closes their eyelids. The second pattern is a pattern when a person's tongue hits their teeth in the mouth. The third pattern is a pattern when a person's lips hit their teeth. "Electromyographic sensor output" in FIG. 10 shows the strength of the output of a configuration similar to that of the electromyographic sensor 40. "Pressure sensor output" in FIG. 10 shows the output of a configuration similar to that of the pressure sensor 20 having the first sensor element 221, the second sensor element 222, and the third sensor element 223. In the reference example, only one of the "electromyographic sensor output" or the "pressure sensor output" is obtained.

[0038] As shown in the "first pattern" in Figure 10, the "electromyography sensor output" shows significantly large fluctuations in output even when a person tightly closes their eyelids, similar to when teeth grinding occurs. Therefore, in the reference example in which only the "electromyography sensor output" is obtained, it is difficult to distinguish between when teeth grinding occurs and when a person tightly closes their eyelids.

[0039] As shown in the "second pattern" and "third pattern" in Fig. 10, the output of the pressure sensor 20 also fluctuates significantly when the tongue or lips hit the teeth. Therefore, in the reference example in which only the "pressure sensor output" is obtained, it is difficult to distinguish between the occurrence of teeth grinding and the occurrence of the tongue or lips hitting the teeth.

[0040] In contrast, in the embodiment, even if a situation similar to the "first pattern" occurs, the output of the pressure sensor 20 does not change significantly. As a result, in the embodiment, it is possible to distinguish between cases where teeth grinding occurs and cases where a person tightly closes their eyelids. Also, in the embodiment, even if a situation similar to the "second pattern" and "third pattern" occurs, the output of the myoelectric sensor 40 does not change significantly. As a result, in the embodiment, it is possible to distinguish between cases where teeth grinding occurs and cases where the tongue or lips hit the teeth. Therefore, in the embodiment, it is possible to distinguish between teeth grinding and events other than teeth grinding with a high degree of accuracy.

[0041] 9, in order to identify bruxism, it is important to synchronize the output of the pressure sensor 20 with the output of the myoelectric sensor 40 in time. In the embodiment, the control unit 12 of the detection device 10 shown in FIG. 1 generates data in which the output of the pressure sensor 20 and the output of the myoelectric sensor 40 are synchronized in time.

[0042] In the embodiment, the pressure sensor 20 generates an output at a frequency of 8 times per second (8 Hz), and the myoelectric sensor 40 generates an output at a frequency of 512 times per second (512 Hz). The control unit 12 acquires the output of the pressure sensor 20 and the output of the myoelectric sensor 40, and generates manageable data for each unit time.

[0043] As shown in FIG. 1, the control unit 12 has a data processing program 12a and a memory 12b. The data processing program 12a is a software program for generating manageable data for each unit time based on the output of the pressure sensor 20 and the output of the myoelectric sensor 40. The control unit 12 includes an arithmetic circuit for reading and executing the data processing program 12a. The memory 12b functions as a storage area capable of temporarily storing data indicating the output of the pressure sensor 20 and the output of the myoelectric sensor 40, as well as data generated using the data processing program 12a. The control unit 12 may be a circuit in which the function of the data processing program 12a is implemented in the form of hardware and which has a storage area for functioning as the memory 12b.

[0044] The communication unit 13 shown in FIG. 1 performs processing related to communication with an external device such as an information processing device 60. Specifically, the communication unit 13 has a circuit for functioning as a NIC (Network Interface Controller) or the like. The communication unit 13 communicates with the external device using a predetermined communication protocol. The predetermined communication protocol may be a known communication protocol such as the Internet Protocol and various protocols for implementing the protocol, or may be a dedicated protocol defined between the communication unit 13 and the information processing device 60. The communication line between the communication unit 13 and the information processing device 60 may be wired, wireless, or a combination of wired and wireless, and may include a public communication line as a part of the communication line.

[0045] In the embodiment, the control unit 12 acquires, for example, output from the pressure sensor 20 and output from the electromyogram sensor 40, synchronizes the output from the pressure sensor 20 and the output from the electromyogram sensor 40 every unit time, and transmits them to the information processing device 60 via the communication unit 13.

[0046] In the embodiment, the control unit 12 and the communication unit 13 are provided within the housing 41 (see Figures 2 and 6), but the specific manner in which the control unit 12 and the communication unit 13 are implemented in the detection device 10 is arbitrary within the scope in which their functions are realized.

[0047] Furthermore, a battery or the like that supplies and stores power necessary for the operation of the sensor unit 11, the control unit 12, and the communication unit 13 may be provided inside the housing 41. Furthermore, the housing 41 may further be provided with an interface or the like that enables a power line for supplying the power to be connected to the housing 41.

[0048] Next, the process flow relating to the operation of the detection device 10 mounted as shown in Fig. 6 will be described with reference to Fig. 11 to Fig. 13. The process described with reference to Fig. 11 to Fig. 13 is realized by the control unit 12 executing the data processing program 12a. In the flowcharts of Fig. 11 to Fig. 13, the operation of the components is described as "ON" and the non-operation is described as "OFF".

[0049] 11 is a flowchart showing an example of a process flow related to the operation of the detection device 10. After starting the operation, the control unit 12 operates the myoelectric sensor 40 at a low frequency cycle (step S1). In addition, in the process of step S1, the control unit 12 sets the pressure sensor 20 and the communication unit 13 to a state in which they are not operated.

[0050] 11 to 13, the operation at a low frequency cycle refers to an operation in which sensing is performed at a time cycle that is 10 times the above-mentioned unit time, for example, and an output is generated. Note that the relationship between the low frequency cycle and the unit time is not limited to this, and can be changed as appropriate. However, according to the operation at a low frequency cycle, the time interval at which sensing is performed is longer than in the operation at a high frequency cycle described later.

[0051] The control unit 12 checks the output of the myoelectric sensor 40 that operates at a low frequency cycle by the process of step S1 (step S2). As a result of checking the output of the myoelectric sensor 40 by the process of step S2, the control unit 12 determines whether the output of the myoelectric sensor 40 exceeds a threshold value (step S3). If it is determined that the output of the myoelectric sensor 40 does not exceed the threshold value (step S3; No), the process proceeds to step S1 unless the operation of the detection device 10 is terminated (step S4; No). That is, the operation state of the myoelectric sensor 40 at a low frequency cycle and the non-operation state of the pressure sensor 20 and the communication unit 13 are continued. An example of a condition that is treated as the end of operation in the process of step S4 is when the user who is the wearer of the detection device 10 performs an operation to terminate the detection device 10 using a terminal app executed on the terminal 60. More specifically, it is assumed that the user activates the terminal app before going to sleep, and performs an operation to terminate the terminal app after waking up. The terminal may be configured to automatically terminate the app after a preset time has elapsed (e.g., 10 hours) since the app was started. The same conditions for terminating the operation may also be applied to the process of step S8, step S14, step S24, and step S28, which will be described later.

[0052] The threshold in the process of step S3 and the process of step S7 described later is, for example, a first threshold Th1 shown in Fig. 9. When an output that exceeds the range of the first threshold Th1 occurs, it is determined that the output of the myoelectric sensor 40 exceeds the threshold. When an output that falls within the range of the first threshold Th1 occurs, it is determined that the output of the myoelectric sensor 40 is less than the threshold. Note that an output that overlaps with a line that defines the range of the first threshold Th1 does not have to be considered as either, or may be included in either.

[0053] On the other hand, if it is determined in step S3 that the output of the myoelectric sensor 40 exceeds the predetermined threshold (step S3; Yes), the control unit 12 operates the myoelectric sensor 40 at a high-frequency cycle (step S5). Furthermore, in the process of step S5, the control unit 12 operates the pressure sensor 20 and the communication unit 13. The operating cycle of the pressure sensor 20 in the process of step S5 is a high-frequency cycle, similar to that of the myoelectric sensor 40.

[0054] 11 to 13 refers to an operation that generates a sensing output in the cycle of the unit time described above, for example. In the embodiment, sensing is performed for one second depending on the output whether it is a low-frequency cycle or a high-frequency cycle, but the extent to which sensing is performed in each cycle can be changed as appropriate.

[0055] The control unit 12 synchronizes the output of the myoelectric sensor 40, which operates at a high frequency cycle, with the output of the pressure sensor 20, and transmits the synchronized output to the information processing device 60 at a predetermined cycle (step S6). The predetermined cycle is, for example, the same cycle as the high frequency cycle (every second in the embodiment), but is not limited thereto and can be changed as appropriate. The data generated in this manner is transmitted via the communication unit 13.

[0056] The control unit 12 also determines whether the output of the myoelectric sensor 40 has remained below the threshold for a predetermined period of time (step S7). If it is determined that the output of the myoelectric sensor 40 has not remained below the threshold for the predetermined period of time (step S7; No), the process proceeds to step S6 unless the operation of the detection device 10 ends (step S8; No). That is, the operating state of the pressure sensor 20 and the myoelectric sensor 40 at a high frequency period, the generation of data, and the transmission of data via the communication unit 13 continue.

[0057] The predetermined time in the description with reference to Figures 11 to 13 is, for example, 5 minutes, but is not limited to this and can be changed as appropriate. However, it is desirable for the predetermined time to be significantly longer than the unit time of the high frequency period.

[0058] On the other hand, if it is determined in step S7 that the output of the myoelectric sensor 40 has been below the threshold for a predetermined period of time (step S7; Yes), the process proceeds to step S4. If the operation of the detection device 10 has not ended (step S4; No), the process proceeds to step S1. That is, the detection device 10 transitions to an operating state of the myoelectric sensor 40 at a low frequency period and a non-operating state of the pressure sensor 20 and the communication unit 13.

[0059] When the operation of the detection device 10 is completed in step S4 (step S4; Yes) and when the operation of the detection device 10 is completed in step S8 (step S8; Yes), the processing of the control unit 12 is completed.

[0060] Fig. 12 is a flowchart showing an example of a process flow related to the operation of the detection device 10, which is different from Fig. 11. After starting the operation, the control unit 12 operates the pressure sensor 20 at a low frequency cycle (step S11). In addition, in the process of step S1, the control unit 12 sets the myoelectric sensor 40 and the communication unit 13 to a state in which they are not operated.

[0061] The control unit 12 checks the output of the pressure sensor 20 operating at a low frequency cycle by the process of step S11 (step S12). As a result of checking the output of the pressure sensor 20 by the process of step S12, the control unit 12 determines whether the output of the pressure sensor 20 exceeds a threshold value (step S13). If it is determined that the output of the pressure sensor 20 does not exceed the threshold value (step S13; No), the process proceeds to step S11 unless the operation of the detection device 10 is terminated (step S14; No). That is, the operating state of the pressure sensor 20 at a low frequency cycle and the non-operating states of the electromyography sensor 40 and the communication unit 13 are continued.

[0062] The threshold value in the process of step S13 and the process of step S17 described later is, for example, a second threshold value Th2 shown in Fig. 9. When an output that exceeds the second threshold value Th2 occurs, it is determined that the output of the pressure sensor 20 exceeds the threshold value. When an output that remains below the second threshold value Th2 occurs, it is determined that the output of the pressure sensor 20 is less than the threshold value. Note that an output that overlaps with the line that defines the second threshold value Th2 does not have to be considered as either, or may be included in either.

[0063] On the other hand, if it is determined in step S13 that the output of the pressure sensor 20 exceeds the predetermined threshold (step S13; Yes), the control unit 12 operates the pressure sensor 20 at a high frequency cycle (step S15). Furthermore, in the process of step S15, the control unit 12 operates the myoelectric sensor 40 and the communication unit 13. The operation cycle of the myoelectric sensor 40 in the process of step S15 is a high frequency cycle, similar to that of the pressure sensor 20.

[0064] The control unit 12 synchronizes the output of the myoelectric sensor 40, which operates at a high frequency cycle, with the output of the pressure sensor 20, and transmits the synchronized output to the information processing device 60 at a predetermined cycle (step S16).

[0065] The control unit 12 also determines whether the output of the pressure sensor 20 has remained below the threshold for a predetermined period of time (step S17). If it is determined that the output of the pressure sensor 20 has not remained below the threshold for the predetermined period of time (step S17; No), the process proceeds to step S16 unless the operation of the detection device 10 is terminated (step S18; No). That is, the operating state of the pressure sensor 20 and the electromyography sensor 40 at a high frequency period, the generation of data, and the transmission of data via the communication unit 13 are continued.

[0066] On the other hand, if it is determined in step S17 that the output of the pressure sensor 20 has been below the threshold for a predetermined period of time (step S17; Yes), the process proceeds to step S14. If the operation of the detection device 10 has not ended (step S14; No), the process proceeds to step S11. That is, the detection device 10 transitions to a low-frequency cycle operation state of the pressure sensor 20 and a non-operation state of the myoelectric sensor 40 and the communication unit 13.

[0067] When the operation of the detection device 10 is completed in step S14 (step S14; Yes) and when the operation of the detection device 10 is completed in step S18 (step S18; Yes), the processing of the control unit 12 is completed.

[0068] Fig. 13 is a flowchart showing an example of a process flow related to the operation of the detection device 10, which is different from Fig. 11 and Fig. 12. After starting the operation, the control unit 12 operates the pressure sensor 20 and the myoelectric sensor 40 (step S21). In addition, in the process of step S21, the control unit 12 sets the communication unit 13 to a state in which it is not operated.

[0069] The operation cycle of the pressure sensor 20 and the myoelectric sensor 40 in the process of step S21 may be a low-frequency cycle or a high-frequency cycle.

[0070] The control unit 12 generates data in which the output of the pressure sensor 20 and the output of the myoelectric sensor 40, which are operated by the process of step S21, are synchronized, and temporarily stores the data in the memory 12b (step S22). The temporary storage of the data in the memory 12b by the process of step S22 holds, for example, only the most recent 5 minutes of data, and discards data older than 5 minutes, but the length of the holding period can be changed as appropriate.

[0071] The control unit 12 determines whether either the pressure sensor 20 or the myoelectric sensor 40 satisfies the condition (step S23). In the process of step S23 and the process of step S28 described below, for example, in the case of the pressure sensor 20, the condition is deemed to be satisfied when the output of the pressure sensor 20 exceeds a predetermined threshold, similar to the determination of step S13 described above. In the case of the myoelectric sensor 40, the condition is deemed to be satisfied when the output of the myoelectric sensor 40 exceeds a predetermined threshold, similar to the determination of step S3 described above.

[0072] If it is determined in step S23 that none of the sensors satisfies the condition (step S23; No), the process proceeds to step S21 unless the operation of the detection device 10 is terminated (step S24; No). That is, the operation of the pressure sensor 20 and the myoelectric sensor 40 and the temporary storage of data in the memory 12b are continued.

[0073] On the other hand, if it is determined in step S23 that any of the sensors satisfies the condition (step S23; Yes), the control unit 12 operates the communication unit 13 (step S25). Also, the control unit 12 transmits the data temporarily stored in the memory 12b in the process of step S22 to the information processing device 60 via the communication unit 13 (step S26).

[0074] If the pressure sensor 20 and the myoelectric sensor 40 are operating at a low frequency cycle in the process of step S21, the operation cycle of the pressure sensor 20 and the myoelectric sensor 40 becomes a high frequency cycle at the time of step S25.

[0075] The control unit 12 synchronizes the output of the myoelectric sensor 40 and the output of the pressure sensor 20 and transmits them to the information processing device 60 at a predetermined cycle (step S27).

[0076] The control unit 12 also determines whether the time during which both the pressure sensor 20 and the myoelectric sensor 40 do not satisfy the conditions has continued for a predetermined time (step S28). If it is determined that the time during which both sensors do not satisfy the conditions has not continued for the predetermined time (step S28; No), the process proceeds to step S27 unless the operation of the detection device 10 is terminated (step S29; No). That is, the operating states of the pressure sensor 20 and the myoelectric sensor 40, data generation, and data transmission via the communication unit 13 at a high frequency cycle continue.

[0077] On the other hand, if it is determined in step S28 that the time during which both the pressure sensor 20 and the myoelectric sensor 40 do not satisfy the condition has continued for a predetermined time (step S28; Yes), the process proceeds to step S24. If the operation of the detection device 10 has not ended (step S24; No), the process proceeds to step S21. That is, the detection device 10 transitions to a state in which the pressure sensor 20 and the myoelectric sensor 40 operate and the data is temporarily stored in the memory 12b.

[0078] When the operation of the detection device 10 is completed in step S24 (step S24; Yes) and when the operation of the detection device 10 is completed in step S29 (step S29; Yes), the processing of the control unit 12 is completed.

[0079] As described above, with reference to Figs. 11 to 13, three processing flows of the control unit 12 are illustrated, and the control unit 12 of the embodiment operates according to any one of these three. Which processing flow is applied may be determined in advance, or may be provided to be selectable. For example, the detection device 10 may be provided to be able to change the processing flow applied to the control unit 12 via an operation on an operation unit such as the switch 46 shown in Fig. 6. Also, the detection device 10 may be provided to be able to change the processing flow applied to the control unit 12 via an operation on the operation unit 63 of the information processing device 60 and communication between the communication unit 62 and the communication unit 13.

[0080] Next, the configuration of the information processing device 60 and the processes performed by the information processing device 60 will be described with reference to FIG. 1 and FIG. 14 to FIG.

[0081] The information processing device 60 includes a data acquisition unit 61, an operation unit 63, a display unit 64, a calculation unit 70, and a storage unit 80. The data acquisition unit 61 acquires data output from the sensor unit 11 from the detection device 10. The data acquisition unit 61 shown in FIG. 1 includes a communication unit 62. The communication unit 62 performs processing related to communication with an external device such as the detection device 10. The communication unit 62 has a circuit for functioning as a NIC, similar to the communication unit 13. In the embodiment, a communication protocol adopted by the information processing device 60 and a communication protocol adopted by the communication unit 13 are selected so that communication is established between the information processing device 60 and the communication unit 13. The data acquisition unit 61 acquires data transmitted from the detection device 10 via the communication unit 13 through communication between the communication unit 13 and the communication unit 62. In FIG. 1, teacher data D1 and identification target data D4 are shown as data acquired by the data acquisition unit 61.

[0082] The operation unit 63 accepts input from a user to the information processing device 60. Specifically, the operation unit 63 has a human interface such as a keyboard, a mouse, etc., and accepts the input operation contents performed by the user via the human interface. The operation unit 63 may be a touch panel or the like provided integrally with the display unit 64.

[0083] The display unit 64 performs display output according to the processing contents performed by the information processing device 60. Specifically, the display unit 64 has at least one display device such as a liquid crystal display or an organic electroluminescence (EL: ElectroLuminescence) display, and performs display output using such a display device. The display output contents by the display unit 64 correspond to the processing contents by the calculation unit 70.

[0084] The calculation unit 70 performs various information processing performed by the information processing device 60. Specifically, the calculation unit 70 has a calculation circuit such as a CPU (Central Processing Unit). The calculation unit 70 of the embodiment functions as a feature acquisition unit 71, a learning unit 73, and a discrimination unit 75. In FIG. 1, a filter 72 used when functioning as the feature acquisition unit 71, a model learning program 74 used when functioning as the learning unit 73, and a bruxism discrimination program 76 used when functioning as the discrimination unit 75 are described in the calculation unit 70. The filter 72, the model learning program 74, and the bruxism discrimination program 76 may be implemented in the calculation unit 70 as hardware, or may be stored in the calculation unit 70 as so-called software programs so as to be readable and referable. In addition, such software programs may be stored in a storage unit 80, and a configuration may be adopted in which the calculation unit 70 reads them from the storage unit 80 when executing processing corresponding to each function.

[0085] The storage unit 80 stores software programs and data used in information processing by the calculation unit 70. Specifically, the storage unit 80 has at least one storage device such as a hard disk drive (HDD: Hard Disk Drive) or a solid state drive (SSD: Solid State Drive). In Fig. 1, teacher data D1, feature amount data D2, model data D3, and classification target data D4 are shown as data stored in the storage unit 80.

[0086] Hereinafter, the information processing by the calculation unit 70 and the data stored in the storage unit 80 will be described with reference to FIGS.

[0087] FIG. 14 is a schematic diagram showing an example of teacher data D1. As shown in FIG. 14, teacher data D1 includes myoelectric data 81, pressure data 82, synchronization information AD1, and identification information AD2. The myoelectric data 81 is a column in which parameters corresponding to the output of the myoelectric sensor 40 are registered. The pressure data 82 is a column in which parameters corresponding to the output of the pressure sensor 20 are registered. The pressure data 82 shown in FIG. 14 includes three columns of "first pressure data", "second pressure data", and "third pressure data". Each of the three columns corresponds to the output of a different sensor. For example, the "first pressure data" corresponds to the output of the first sensor element 221, the "second pressure data" corresponds to the output of the second sensor element 222, and the "third pressure data" corresponds to the output of the third sensor element 223. That is, the number of columns included in the pressure data 82 corresponds to the number of sensors provided in the pressure sensor 20. The myoelectric data 81 and pressure data 82 are output from the sensor unit 11 and transmitted from the detection device 10 via the communication unit 13 , and are acquired by the information processing device 60 via the communication unit 62 of the data acquisition unit 61 .

[0088] The synchronization information AD1 is a column in which a parameter of time information indicating the time point of acquisition of the output of the myoelectric sensor 40 registered in the myoelectric data 81 and the output of the pressure sensor 20 registered in the pressure data 82 is registered. Each of the multiple records in the table shown in FIG. 14 has different parameters of time information registered in the field of the synchronization information AD1. That is, the teacher data D1 shown in FIG. 14 is data in a format that allows information for each time point at which the output of the pressure sensor 20 and the myoelectric sensor 40 is acquired to be managed on a record-by-record basis. The format of the data shown in FIG. 15 and FIG. 19 described later is also the same as the format of the teacher data D1. In other words, the myoelectric data (e.g., myoelectric data 81) and the pressure data (e.g., pressure data 82) that can be associated with each other by time information corresponding to the acquisition timing of the data as in the synchronization information AD1 can be said to be data in which the output of the pressure sensor 20 and the output of the myoelectric sensor 40 are synchronized. The output of the pressure sensor 20 and the output of the myoelectric sensor 40 are synchronized at the time of data transmission via the communication unit 13 under the control of the control unit 12. The timing of adding the synchronization information AD1 may be before the data is transmitted via the communication unit 13 or after the data is transmitted via the communication unit 13.

[0089] The process of attaching the synchronization information AD1 to the data may be performed by the control unit 12, but is performed by the data acquisition unit 61 of the information processing device 60 in this embodiment.

[0090] The identification information AD2 is a column in which parameters indicating the relationship between the combination of the output of the pressure sensor 20 and the output of the electromyography sensor 40 indicated by each record included in the teacher data D1 and teeth grinding are registered. Specifically, the identification information AD2 shown in Fig. 14 includes four columns: "Correct answer (presence or absence of teeth grinding)", "Correct answer (type of teeth grinding)", "Correct answer (location of teeth grinding)", and "Correct answer (intensity)".

[0091] A parameter registered in a field included in the column of "correct answer (presence or absence of teeth grinding)" indicates whether teeth grinding is occurring at the time of the record including the parameter. Hereinafter, when a "parameter of A" is described, it means a parameter registered in a field included in the column of "A". For example, if a combination of a parameter of electromyography data 81 and a parameter of pressure data 82 included in a certain record is a combination of parameters that occurs when "teeth grinding is occurring" such as periods T1 and T2 in FIG. 9, the parameter of "correct answer (presence or absence of teeth grinding)" in that record is "1". Conversely, if a combination of a parameter of electromyography data 81 and a parameter of pressure data 82 is a combination of parameters that occurs when "teeth grinding is not occurring", the parameter of "correct answer (presence or absence of teeth grinding)" in that record is "0".

[0092] The "correct answer (type of teeth grinding)" parameter indicates the type of teeth grinding. For records in which the "correct answer (presence or absence of teeth grinding)" parameter is "0", "0" is registered as the "correct answer (type of teeth grinding)" parameter. In other words, for records in which teeth grinding is not occurring, there is no type of teeth grinding to indicate because teeth grinding is not occurring in the first place, and so "0" is registered.

[0093] The number of variations of the parameters of the "correct answer (type of teeth grinding)" that are not "0" corresponds to the types of teeth grinding that can be identified by the detection system 100. For example, if there are four types of teeth grinding that can be identified by the detection system 100, a total of four types of parameters, such as "1", "2", "3", and "4", are expected as parameters corresponding to the types of teeth grinding. Therefore, in this case, the parameter of the "correct answer (type of teeth grinding)" is set to any one of "0", "1", "2", "3", or "4".

[0094] For example, a parameter of "1" for the "correct answer (type of teeth grinding)" indicates that the type of teeth grinding is gliding. A parameter of "2" for the "correct answer (type of teeth grinding)" indicates that the type of teeth grinding is tapping. A parameter of "3" for the "correct answer (type of teeth grinding)" indicates that the type of teeth grinding is clenching. A parameter of "4" for the "correct answer (type of teeth grinding)" indicates that the type of teeth grinding is nothing. The relationship between the numbers and the types of teeth grinding is merely an example and is not limited to this, and further numbers and types may be added. The types of teeth grinding may reflect medical information that will be recognized currently and in the future.

[0095] Gliding is a type of teeth grinding where the upper and lower teeth grind against each other. Gliding tends to occur easily while people are sleeping, and can be accompanied by a sound when it occurs. If the sound is expressed as an onomatopoeia, it would be something like "grrr." People who experience gliding tend to unconsciously move their teeth quickly and widely.

[0096] Tapping is a type of teeth grinding in which the upper and lower teeth hit each other. People who experience tapping tend to move their lower jaw up and down. The sound that accompanies tapping can be expressed in onomatopoeic words as "click-click."

[0097] Clenching is teeth grinding in which the upper and lower teeth are strongly clenched. It tends to occur relatively often in people who are used to clenching their teeth for work, sports, etc. When clenching occurs, a large burden is placed on the jaw. In addition, people who tend to clench may unconsciously clench their teeth while sleeping, and may feel stiffness when yawning when waking up. Such people may also tend to feel tired when waking up.

[0098] Nothing is a type of teeth grinding that involves grinding certain parts of the teeth together. People who are prone to nothing can do it while sleeping.

[0099] The parameter "correct answer (teeth grinding location)" indicates the location where teeth grinding occurs. For records where the parameter "correct answer (presence or absence of teeth grinding)" is "0", "0" is registered as the parameter "correct answer (teeth grinding location)".

[0100] Specifically, a parameter of the "correct answer (teeth grinding location)" that is not "0" indicates which of the multiple sensors functioning as the pressure sensor 20 has been identified as the sensor in which teeth grinding is occurring. More specifically, the number of variations of the "correct answer (teeth grinding type)" parameters that are not "0" corresponds to the number of sensors functioning as the pressure sensor 20. For example, if there are three sensors functioning as the pressure sensor 20, namely, the first sensor element 221, the second sensor element 222, and the third sensor element 223, it is assumed that one or more of the three types of parameters, such as "1", "2", and "3", are registered in the "correct answer (teeth grinding location)" field as parameters indicating the location where teeth grinding is occurring. The "correct answer (teeth grinding location)" parameter included in a record at a time when teeth grinding is occurring in multiple locations includes multiple values.

[0101] For example, a parameter of "correct answer (teeth grinding location)" of "1" indicates that teeth grinding occurs at the location where the first sensor element 221 is provided, i.e., near the molar M1 in FIG. 7. A parameter of "correct answer (teeth grinding location)" of "2" indicates that teeth grinding occurs at the location where the second sensor element 222 is provided, i.e., near the molar M2 in FIG. 7. A parameter of "correct answer (teeth grinding location)" of "3" indicates that teeth grinding occurs at the location where the third sensor element 223 is provided, i.e., near the incisor M3 in FIG. 7. Based on these examples, if the parameters of "correct answer (teeth grinding location)" are "1, 3", teeth grinding occurs near the molar M1 and incisor M3 in FIG. 7.

[0102] The "correct answer (intensity)" parameter indicates the intensity of teeth grinding. For records in which the "correct answer (presence or absence of teeth grinding)" parameter is "0", "0" is registered as the "correct answer (intensity)" parameter. Specifically, the larger the value of the "correct answer (intensity)" parameter that is not "0", the more intense the teeth grinding that is occurring. The range of values ​​that the "correct answer (intensity)" parameter can take is up to the discretion of the user.

[0103] The myoelectric data 81 and the pressure data 82 of the teacher data D1 are registered by the data acquisition unit 61 acquiring data output from the sensor unit 11 from the detection device 10. The synchronization information AD1 of the teacher data D1 is automatically added by processing by the circuit or the calculation unit 70 included in the data acquisition unit 61 as the data acquisition unit 61 acquires the data. The identification information AD2 of the teacher data D1 is manually added by a person who can diagnose bruxism, such as a dentist or a person who has received guidance from a dentist. When adding the identification information AD2, for example, under a situation where the display unit 64 is displaying and outputting (see FIG. 9, for example) the myoelectric data 81 and the pressure data 82, the person individually judges from what time point to what time point bruxism has occurred. The person inputs according to the judgment result via the operation unit 63. As a result, the identification information AD2 is added.

[0104] By adding the identification information AD2 to the myoelectric data 81 and the pressure data 82, the teacher data D1 functions as teacher data in so-called machine learning.

[0105] The feature amount acquiring unit 71 generates feature amount data D2 from the teacher data D1. Hereinafter, the configuration of the feature amount data D2 and the processing contents by the feature amount acquiring unit 71 will be described with reference to FIG.

[0106] FIG. 15 is a schematic diagram showing an example of feature data D2. The content of the feature data D2 shown in FIG. 15 corresponds to the content of the teacher data D1 shown in FIG. 14. As shown in FIG. 15, the teacher data D1 includes myoelectric feature data 83, pressure feature data 84, and synchronization information AD3. The myoelectric feature data 83 is registered with a parameter of a feature corresponding to the output of the myoelectric sensor 40. The myoelectric feature data 83 includes a first myoelectric feature AD4, a second myoelectric feature AD5, a third myoelectric feature AD6, and a fourth myoelectric feature AD7. The first myoelectric feature AD4, the second myoelectric feature AD5, the third myoelectric feature AD6, and the fourth myoelectric feature AD7 indicate features derived from the myoelectric data 81 through different processes.

[0107] The parameter of the first myoelectric feature AD4 is the root mean square (RMS) of the parameter of the myoelectricity after preprocessing. The preprocessing will be described later. Specifically, the parameter of the first myoelectric feature AD4 is derived based on the following formula (1). Note that m(t) indicates myoelectric data after preprocessing at a certain timing, time t. i(t) indicates the number of the myoelectric data vector (one-dimensional array) at the time t. n in formulas (1) to (4) corresponds to the output frequency of the myoelectric sensor 40. In the embodiment, as described above, the output of the myoelectric sensor 40 is 512 times per second (512 Hz), so n=512. Therefore, m(i) on the left side of formula (1) can take values ​​from m(1) to m(512). Each of the values ​​from m(1) to m(512) corresponds to each of the outputs of the myoelectric sensor 40 obtained 512 times (512 Hz) for one second.

number

[0108] The parameters of the myoelectric data 81 shown in FIG. 14 are merely schematic, and in practice, parameters indicating data corresponding to the output frequency of the myoelectric sensor 40 are registered as described above. The parameters of the second myoelectric feature value AD5 indicate values ​​obtained by rectifying and smoothing the preprocessed myoelectric parameters (ARV: Average Rectified Value). Specifically, a low-pass filter H2(ω) is applied to the absolute value of the myoelectric parameter, and the value passed through the low-pass filter is registered as the parameters of the second myoelectric feature value AD5 represented by the following formula (2). The cutoff frequency of the low-pass filter H2(ω) is, for example, 2 Hz, but can be changed as appropriate. Here, the value passed through the low-pass filter H2(ω) corresponds to data filtered to a specific frequency component from the output of the myoelectric sensor 40.

number

[0109] The parameters of the third myoelectric feature AD6 indicate values ​​that pass through a band-pass filter H3(ω) applied to the pre-processed myoelectric parameters. The values ​​that pass through the band-pass filter are expressed as in the following formula (3). The passband of the band-pass filter H3(ω) is, for example, from 50 Hz to 200 Hz, but can be changed as appropriate. Here, the values ​​that pass through the band-pass filter H3(ω) correspond to data that has been filtered to a specific frequency component from the output of the myoelectric sensor 40.

number

[0110] The parameters of the fourth myoelectric feature AD7 indicate values ​​that pass through a bandpass filter H4(ω) applied to the preprocessed myoelectric parameters. The bandpass filters H3(ω) and H4(ω) have different passbands. The values ​​that pass through the bandpass filters are expressed by the following formula (4). The passband of the bandpass filter H4(ω) is, for example, from 200 Hz to 300 Hz, but can be changed as appropriate. Here, the values ​​that pass through the bandpass filter H4(ω) correspond to data that has been filtered to a specific frequency component from the output of the myoelectric sensor 40.

number

[0111] In addition, when the parameters of the third myoelectric feature AD6 are expressed in the frequency domain, M3(ω)=H3(ω)M(ω). M(ω) and M3(ω) are obtained by Fourier transforming m(t) and m3(t), respectively. In addition, when the parameters of the fourth myoelectric feature AD7 are expressed in the frequency domain, M4(ω)=H4(ω)M(ω). In these, M(ω) indicates the frequency domain of the myoelectric data 81 after preprocessing.

[0112] The pressure feature amount data 84 includes a first pressure feature amount AD8 and a second pressure feature amount AD9. The first pressure feature amount AD8 and the second pressure feature amount AD9 each indicate a feature amount derived from the pressure data 82 via a different process.

[0113] The parameter of the first pressure feature AD8 indicates a feature normalized by the difference between the parameter value of the pressure data 82 and the value of the steady state p0. The parameter value of the first pressure feature AD8 is expressed by the following formula (5). p(t) indicates the parameter value of the pressure data 82 at a certain timing, time t. The value of the steady state p0 is a value corresponding to the output of the pressure sensor 20 in a static state where no force is applied to the first portion P1.

number

[0114] The parameter of the second pressure feature AD9 indicates a value obtained by applying a moving average filter H6(ω) to the parameter value (p1) of the first pressure feature AD8 and passing the moving average filter. The parameter value of the second pressure feature AD9 can be said to be a value obtained by taking a time average over a unit time (e.g., 1 second) that is a separator between multiple records of the teacher data D1. The value passed through the moving average filter is expressed as in the following formula (6). n p (t) indicates the number of the pressure data vector (one-dimensional array) at time t. n in formula (6) corresponds to the output frequency of the pressure sensor 20. In the embodiment, as described above, the output of the pressure sensor 20 is 8 times per second (8 Hz), so n=8. Here, the value that has passed through the moving average filter H6(ω) corresponds to data that has been filtered to a specific frequency component from the output of the pressure sensor 20.

number

[0115] When the parameters of the second pressure feature AD9 are expressed in the frequency domain, P2(ω)=H6(ω)P1(ω), where P1(ω) represents the frequency domain of the first pressure feature AD8 and corresponds to the Fourier transform of p1(t).

[0116] 14 and the like are roughly written as numerical values, the actual myoelectric data 81 and pressure data 82 can be regarded as waveforms showing high and low outputs, as described with reference to Fig. 9. Therefore, the components included in such waveforms can be expressed in frequency domains such as the above-mentioned formulas (2), (3), (4), (6), M3(ω)=H3(ω)M(ω), M4(ω)=H4(ω)M(ω), and P2(ω)=H6(ω)P1(ω), and the components included in specific frequency domains thus expressed can be targets for acquisition by the feature acquisition unit 71.

[0117] The first pressure feature amount AD8 and the second pressure feature amount AD9 are derived individually for each of the columns included in the pressure data 82 of the teacher data D1 (see FIG. 14). That is, the first pressure feature amount AD8 and the second pressure feature amount AD9 are derived individually for each sensor functioning as the pressure sensor 20. For example, the parameters of the first pressure feature amount AD8 and the second pressure feature amount AD9 of the "pressure sensor 1" shown in FIG. 15 are derived from the parameters registered in the column of the "pressure sensor 1" of the pressure data 82 shown in FIG. 14. The parameters of the first pressure feature amount AD8 and the second pressure feature amount AD9 of the "pressure sensor 2" shown in FIG. 15 are derived from the parameters registered in the column of the "pressure sensor 2" of the pressure data 82 shown in FIG. 14. The parameters of the first pressure feature amount AD8 and the second pressure feature amount AD9 of the "pressure sensor 3" shown in FIG. 15 are derived from the parameters registered in the column of the "pressure sensor 3" of the pressure data 82 shown in FIG. 14.

[0118] The synchronization information AD3 of the feature data D2 (see FIG. 15) is a parameter corresponding to the synchronization information AD1 of the teacher data D1 (see FIG. 14). That is, the record of the feature data D2 shown in FIG. 15 is a record indicating the feature of the myoelectric data 81 and the pressure data 82 included in the record of the teacher data D1 in which the same parameters as the parameters of the synchronization information AD3 attached to the record are attached to the synchronization information AD1.

[0119] The feature acquisition unit 71 performs preprocessing on the parameters of the myoelectric data 81 before deriving the first myoelectric feature amount AD4, the second myoelectric feature amount AD5, the third myoelectric feature amount AD6, and the fourth myoelectric feature amount AD7. The above-mentioned "preprocessed myoelectric parameters" refer to the parameters of the myoelectric data 81 after the preprocessing. Specifically, the parameters of the myoelectric data 81 can be considered as parameters of an electromyogram signal. The effective frequency band of the electromyogram signal is from 5 Hz to 500 Hz. The feature acquisition unit 71 applies a bandpass filter in the effective frequency band to the parameters of the myoelectric data 81, and treats the data that has passed through the bandpass filter as "preprocessed myoelectric parameters."

[0120] Furthermore, since muscle tissues are different for each human subject on whom sensing is performed by the detection device 10, the myoelectric potential signal measured by each subject is different even when the same movement is performed. In addition, even for the same subject, the measured value may change depending on the physical condition such as the condition of the skin. Therefore, the skin resistance may change due to an experiment that spans days, or due to the attachment and detachment of the electrodes of the myoelectric sensor 40, and the measured value may also change. From this viewpoint, it is desirable to use a normalized myoelectric potential value. For example, normalization can be performed by the %MVC (Maximum Voluntary Contraction) method. In this case, as a preparation for sensing by the detection device 10, a process of having the subject perform MVC and normalizing the parameters of the myoelectric data 81 by setting the value of the integrated electromyogram (IEMG) at that time as a value of 100% may be further included in the pre-processing.

[0121] 1 derives a first myoelectric feature amount AD4, a second myoelectric feature amount AD5, a third myoelectric feature amount AD6, and a fourth myoelectric feature amount AD7 of myoelectric feature amount data 83 based on parameters of the preprocessed myoelectricity obtained through the above-mentioned preprocessing. The feature amount acquisition unit 71 also derives a first pressure feature amount AD8 and a second pressure feature amount AD9 of pressure feature amount data 84 based on parameters of pressure data 82. The feature amount acquisition unit 71 generates records of feature amount data D2 for each record of teacher data D1 to which synchronization information AD1 is added, and generates feature amount data D2 by adding synchronization information AD3 to each record of feature amount data D2.

[0122] The filter 72 is, for example, a software program that functions as various filters used in preprocessing and the second myoelectric feature amount AD5, the third myoelectric feature amount AD6, the fourth myoelectric feature amount AD7, and the second pressure feature amount AD9, but is not limited thereto. For example, the feature amount acquisition unit 71 may be a dedicated circuit, and the filter 72 may be provided as a digital filter. In the embodiment, the filter 72 is a software program that further includes processing content related to deriving the first myoelectric feature amount AD4 and the first pressure feature amount AD8, but the filter 72 and a software program for processing by the feature amount acquisition unit 71 that does not use a filter may be provided separately.

[0123] The learning unit 73 performs machine learning to derive model data D3 based on the identification information AD2 of the teacher data D1 (see FIG. 14) and the feature amount data D2 (see FIG. 15). Specifically, for example, the learning unit 73 functions by the calculation unit 70 executing a model learning program 74. The learning unit 73 performs supervised machine learning based on a support vector machine (SVM). The SVM is a pattern recognition model for classifying patterns based on a combination of multiple parameters. Note that the applicable machine learning method is not limited to SVM. Any supervised machine learning method such as a decision tree or logistic regression can be applied.

[0124] In the embodiment, the identification information AD2 is used as a basis for classification to realize supervised machine learning. That is, the identification information AD2 indicates which combination of parameters corresponds to which pattern in terms of "presence or absence of teeth grinding," "type of teeth grinding," "location of teeth grinding," and "intensity of teeth grinding." In addition, since the teacher data D1 and the feature data D2 can be linked by the synchronization information AD1 and synchronization information AD3, the relationship between the myoelectric feature data 83 and the pressure feature data 84 and the identification information AD2 can be specified. Therefore, in the supervised machine learning using the SVM of the embodiment, a model is generated to enable classification of which pattern the combination of the parameters included in the myoelectric feature data 83 and the parameters included in the pressure feature data 84 corresponds to in terms of "presence or absence of teeth grinding," "type of teeth grinding," "location of teeth grinding," and "intensity of teeth grinding." The model is the model data D3 shown in FIG. 1. Hereinafter, an outline of the machine learning will be described with reference to FIG. 16 to FIG. 18.

[0125] FIG. 16 is a diagram showing a schematic two-dimensional graph of machine learning related to classification of the presence or absence of teeth grinding. The vertical axis in FIG. 16 and FIG. 17 and FIG. 18 described later indicates a pressure feature amount. The pressure feature amount referred to here is either the first pressure feature amount AD8 or the second pressure feature amount AD9. The vertical axis in FIG. 16, FIG. 17, and FIG. 18 indicates a myoelectric feature amount. The myoelectric feature amount referred to here is either the first myoelectric feature amount AD4, the second myoelectric feature amount AD5, the third myoelectric feature amount AD6, or the fourth myoelectric feature amount AD7. A point DP in FIG. 16, FIG. 17, and FIG. 18 indicates a combination of a pressure feature amount and a myoelectric feature amount of a certain record included in the feature amount data D2. In FIG. 16, FIG. 17, and FIG. 18, one of the multiple black points is given the symbol DP, and the other black points also indicate a combination of a pressure feature amount and a myoelectric feature amount of a certain record included in the feature amount data D2. Therefore, when a point DP is written, it indicates a combination of a pressure feature and a myoelectric feature of a certain record included in the feature data D2.

[0126] In Fig. 16, Fig. 17, and Fig. 18, the relationship between one pressure feature and one myoelectric feature is shown for the sake of clarity. In reality, however, processing is performed assuming a {(q×α)+β}-dimensional vector space based on a combination of the first pressure feature AD8 and the second pressure feature AD9 of each of the multiple sensors and the first myoelectric feature AD4, the second myoelectric feature AD5, the third myoelectric feature AD6, and the fourth myoelectric feature AD7. Here, q is the number of types of parameters included in the pressure feature data 84. Also, α is the number of multiple sensors functioning as the pressure sensor 20. Also, β is the number of types of parameters included in the myoelectric feature data 83. Therefore, in the embodiment, q=2, α=3, and β=4. From this, in the embodiment, a {(q×α)+β}={(2×3)+4}=10-dimensional vector space is assumed.

[0127] 16, as an example of classification by SVM, classification based on the value of "correct answer (presence or absence of teeth grinding)" among classifications included in the identification information AD2, i.e., classification of "presence or absence of teeth grinding", will be described. In Figs. 16 and 17, a point DP within the boundary SP1 is a point DP where the value of "correct answer (presence or absence of teeth grinding)" is 1, i.e., a point DP corresponding to a combination of a pressure feature and an electromyographic feature when teeth grinding is occurring. Note that while the boundary SP1 is shown as a two-dimensional frame line in Fig. 16, it is actually a hyperplane in a {(q×α)+β}-dimensional vector space.

[0128] Although the classification is not included in the identification information AD2, other classifications are possible by further including information indicating other classifications in the teacher data. For example, the inside of the boundary EP1 shown in FIG. 16 corresponds to the first pattern described with reference to FIG. 10. Also, the inside of the boundary EP2 shown in FIG. 16 corresponds to the second pattern described with reference to FIG. 10. When a parameter indicating whether or not it is one of these patterns is further included in the identification information AD2, it is also possible to perform machine learning regarding classifications such as the boundary EP1 and the boundary EP2.

[0129] It is also possible to exclude some data that is clearly not related to teeth grinding from among the data included in the feature data D2. As described with reference to FIG. 9, in teeth grinding, characteristic outputs are generated from both the pressure sensor 20 and the myoelectric sensor 40. On the other hand, the point DP with a smaller myoelectric feature value than the boundary line BL1 shown in FIG. 16 is clearly poor in output from the myoelectric sensor 40. Also, the point DP with a smaller pressure feature value than the boundary line BL2 shown in FIG. 16 is clearly poor in output from the pressure sensor 20. From this perspective, at least one of the point DP with a smaller myoelectric feature value than the boundary line BL1 and the point DP with a smaller pressure feature value than the boundary line BL2 may be excluded from the machine learning target. In this case, the processing load related to the machine learning can be further reduced. Also, even if it is based on data obtained when applying the variable operation control of the detection device 10 based on thresholds (for example, the first threshold value Th1 and the second threshold value Th2) as described with reference to FIG. 11 to FIG. 13, sufficiently high-precision machine learning can be realized. Incidentally, including all points DP as targets for machine learning contributes to the generation of model data D3 that can more reliably classify points DP that are to be excluded as not corresponding to teeth grinding, so it is acceptable to include all points DP as targets for machine learning.

[0130] Next, machine learning for identifying more detailed classifications when bruxism occurs will be described with reference to FIG.

[0131] FIG. 17 is a diagram showing machine learning for classification of types of bruxism by a certain subject in a two-dimensional graph. For example, a point DP within the boundary SP1 and boundary SP11 shown in FIG. 17 corresponds to gliding among types of bruxism. A point DP within the boundary SP1 and boundary SP12 corresponds to nothing among types of bruxism. A point DP within the boundary SP1 and boundary SP13 corresponds to tapping among types of bruxism. A point DP within the boundary SP1 and not included in any of the boundaries SP11, SP12, and SP13 corresponds to clenching among types of bruxism. Such classification of types of bruxism can be realized based on the parameters of the "correct answer (type of bruxism)" among the parameters included in the identification information AD2, for example. The boundaries SP11, SP12, and SP13 are hyperplanes in a {(q×α)+β}-dimensional vector space, similar to the boundary SP1.

[0132] In addition, parameters other than the "correct answer (type of teeth grinding)" can also be considered in classifying the types of teeth grinding. For example, the vector direction indicated by the intensity direction PW shown in FIG. 17 corresponds to the magnitude of the value of the "correct answer (intensity)". Therefore, the boundary SP11 is a classification in which the intensity of teeth grinding is relatively strong. The boundary SP13 is a classification in which the intensity of teeth grinding is relatively weak. The boundary SP12 is a classification in which the intensity of teeth grinding that may occur is intermediate. Such a correlation between the intensity of teeth grinding and the classification of the types of teeth grinding can be further included in the learning related to the classification of the types of teeth grinding. In a similar way, a correlation can also be established between the "correct answer (location of teeth grinding)", that is, a sensor that shows an output according to pressure among the multiple sensors that function as the pressure sensor 20, and the classification of the types of teeth grinding. Such a correlation between the location of teeth grinding and the classification of the types of teeth grinding can be further included in the learning related to the classification of the types of teeth grinding.

[0133] Furthermore, the teacher data D1 and the feature data D2 are not limited to data from one subject. Machine learning based on data from multiple subjects may be performed, and learning may be performed to realize classification of points DP according to the subject's tendencies. As the subject's tendencies, for example, learning may be performed in which multiple subjects with the same or similar physical characteristics are regarded as "(a set of) subjects showing similar tendencies." Examples of information indicating physical characteristics include information indicating the subject's age, sex, height, weight, etc.

[0134] FIG. 18 is a diagram showing, in a two-dimensional graph, a schematic diagram of machine learning related to classification of types of bruxism by a subject different from that in FIG. 17. The distribution of points DP shown in FIG. 18 is different from the distribution of points DP shown in FIG. 17. This is because the points DP shown in FIG. 18 are teacher data D1 and feature data D2 different from the teacher data D1 and feature data D2 that are the basis of the points DP shown in FIG. 17. The subject from which the teacher data D1 that is the basis of the points DP shown in FIG. 17 was obtained is different from the subject from which the teacher data D1 that is the basis of the points DP shown in FIG. 18 was obtained. Therefore, the boundary SP2 indicates a spatial region different from the boundary SP1. In FIG. 18, the points DP within the boundary SP2 are points DP where the value of "correct answer (presence or absence of bruxism)" is 1, that is, points DP corresponding to a combination of pressure feature value and myoelectric feature value when bruxism is occurring.

[0135] In addition, a point DP within the boundary SP2 and boundary SP21 shown in FIG. 18 corresponds to the type of bruxism, gliding. In addition, a point DP within the boundary SP2 and boundary SP22 corresponds to the type of bruxism, nothing. In addition, a point DP within the boundary SP2 and boundary SP23 corresponds to the type of bruxism, tapping. In addition, a point DP within the boundary SP2 and not included in any of the boundaries SP21, SP22, and SP23 corresponds to the type of bruxism, clenching. In addition, the boundary SP2, as well as the boundaries SP21, SP22, and SP23, are hyperplanes in the {(q×α)+β}-dimensional vector space, similar to the boundary SP1.

[0136] As described with reference to FIGS. 16 to 18, the learning unit 73 shown in FIG. 1 generates model data D3 as an output of supervised machine learning based on SVM using the identification information AD2 of the teacher data D1 (see FIG. 14) and the feature amount data D2 (see FIG. 15) as input (teacher data). The model data D3 includes teeth grinding estimation model data 85 and teeth grinding detail identification model data 86. The teeth grinding estimation model data 85 is model data used to identify whether the data indicates the occurrence of teeth grinding. The teeth grinding detail identification model data 86 is model data used to identify the type of teeth grinding (type, location, strength) when the occurrence of teeth grinding is estimated. The teeth grinding estimation model data 85 and the teeth grinding detail identification model data 86 are referenced in the processing by the identification unit 75.

[0137] The identification unit 75 shown in FIG. 1 identifies whether data acquired by the data acquisition unit 61 and not having the identification information AD2 attached thereto is data that is suspected to indicate the occurrence of teeth grinding. Furthermore, when the occurrence of teeth grinding is suspected, the identification unit 75 identifies the type of teeth grinding (type, location, strength). Data acquired by the data acquisition unit 61 and not having the identification information AD2 attached thereto is, for example, the identification target data D4. Specifically, the identification unit 75 functions by, for example, the calculation unit 70 executing and processing the teeth grinding identification program 76. The identification unit 75 receives the identification target data D4 as input data, refers to the model data D3, and outputs a data classification based on the SVM.

[0138] FIG. 19 is a diagram showing an example of the relationship between the identification target data D4 and the identification information OP indicating the output of the identification unit 75 based on the identification target data D4. The identification target data D4 includes myoelectric data 87 and pressure data 88. The myoelectric data 87 is a column in which parameters corresponding to the output of the myoelectric sensor 40 are registered, similar to the myoelectric data 81. The pressure data 88 is a column in which parameters corresponding to the output of the pressure sensor 20 are registered, similar to the pressure data 82. Therefore, the pressure data 88 shown in FIG. 19 includes three columns of "first pressure data", "second pressure data", and "third pressure data", similar to the pressure data 82 shown in FIG. 14. The identification target data D4 is data acquired by the data acquisition unit 61, similar to the teacher data D1, including the fact that the synchronization information AD1 described with reference to FIG. 14 is added. However, unlike the teacher data D1 described with reference to FIG. 14, the identification target data D4 does not include the identification information AD2. In other words, the identification target data D4 is not teacher data, but is data to be identified in data classification using model data based on machine learning.

[0139] The identification unit 75 classifies each record of the identification target data D4 by classifying the points DP by SVM with reference to the model data D3. Specifically, the identification unit 75 generates first data as data similar to the myoelectric feature data 83 from the myoelectric data 87 of the identification target data D4, and generates second data as data similar to the pressure feature data 84 from the pressure data 88, in the same way that the feature acquisition unit 71 generates the feature data D2 from the teacher data D1. The first data is, in other words, data that is referred to as "myoelectric feature" in the description with reference to Figs. 16 to 18. The second data is, in other words, data that is referred to as "pressure feature" in the description with reference to Figs. 16 to 18.

[0140] The identification unit 75 treats a combination of an electromyographic feature and a pressure feature indicated by each record of the generated first data and second data as a point DP and represents it in a {(q×α)+β}-dimensional vector space. The identification unit 75 also refers to the model data D3 and applies a boundary indicating the classification of the point DP to the {(q×α)+β}-dimensional vector space. The boundary indicating the classification of the point DP here is, for example, the boundary SP1, boundary SP11, boundary SP12, and boundary SP13 shown in FIG. 17 or the boundary SP2, boundary SP21, boundary SP22, and boundary SP23 shown in FIG. 18. The identification unit 75 estimates whether or not bruxism is occurring based on the relationship between the point DP and the boundary indicating the classification of the point DP, and specifies the details of the bruxism (type, location, strength) when the occurrence of bruxism is estimated.

[0141] In an embodiment, the identification unit 75 adds, for example, identification information OP shown in Fig. 19 to the identification target data D4 as an output indicating the result of estimating whether bruxism is occurring and identifying the details of the bruxism (type, location, intensity) if the occurrence of bruxism is estimated. The identification information OP includes four columns: "Result (presence or absence of bruxism)", "Result (type of bruxism)", "Result (location of bruxism)", and "Result (intensity)".

[0142] The "Result (presence or absence of teeth grinding)" parameter indicates the result of the identification unit 75 estimating whether teeth grinding is occurring based on the relationship between boundary SP1 or boundary SP2 and point DP. The rules for the "Result (presence or absence of teeth grinding)" parameter are the same as those for the "Correct answer (presence or absence of teeth grinding)" parameter. Therefore, the "Result (presence or absence of teeth grinding)" parameter included in the record corresponding to point DP estimated to be "where teeth grinding is occurring" is set to "1". The "Result (presence or absence of teeth grinding)" parameter included in the record corresponding to point DP estimated to be "where teeth grinding is not occurring" is set to "0".

[0143] The "Result (Type of Teeth Grinding)" parameter indicates the result of the identification unit 75 identifying the type of teeth grinding when teeth grinding is estimated to be occurring based on the relationship between boundary SP1, boundary SP11, boundary SP12, boundary SP13 or boundary SP2, boundary SP21, boundary SP22, boundary SP23 and point DP. The rules for the "Result (Type of Teeth Grinding)" parameter are the same as the rules for the "Correct Answer (Type of Teeth Grinding)" parameter.

[0144] The "Result (bruxism location)" parameter indicates the result of the identification unit 75 identifying the sensor that provided an output indicating the occurrence of pressure. The rules for the "Result (bruxism location)" parameter are the same as those for the "Correct answer (bruxism location)" parameter.

[0145] The "Result (Bruxism Intensity)" parameter indicates the result of the identification unit 75 determining the intensity of bruxism when bruxism is estimated to be occurring based on the relationship between the intensity direction PW (see Figs. 17 and 18) and point DP. The rules for the "Result (Bruxism Intensity)" parameter are the same as those for the "Correct Answer (Bruxism Intensity)" parameter.

[0146] The word "estimate" in the phrase "estimating whether bruxism is occurring" is intended to indicate that it is medical personnel such as doctors who make the diagnosis of bruxism, and not the detection system 100. That is, the detection system 100 provides information including the identification information OP to medical personnel as information that helps to diagnose whether a subject using the detection device 10 is bruxism. The information is provided, for example, by display output by the display unit 64. The medical personnel can diagnose bruxism of the subject with the help of the information provided by the detection system 100. In this way, it is the medical personnel who make the final diagnosis, and the word "estimate" indicates that the diagnosis is not completed at the time when the identification information OP is output by the detection system 100, but is merely at the estimation stage.

[0147] Also, a rule is determined in advance as to whether the boundary SP1 shown in FIG. 17 or the boundary SP2 shown in FIG. 18 is adopted, or whether model data D3 corresponding to other subjects, not limited to the boundary SP1 or the boundary SP2, is adopted. For example, assume that individual model data D3 is prepared in advance for each of a plurality of subjects. In this case, a rule may be set such that the subject's personal information and the model data D3 are associated in advance, and the input indicating the personal information is made via the operation unit 63 at the time of estimation and identification by the identification unit 75. This allows the model data D3 specialized for each subject to be applied. Also, a rule may be adopted that associates the model data D3 with information indicating the subject's physical characteristics. Examples of information indicating the physical characteristics include information indicating the age, sex, height, weight, etc. of the subject from whom the teacher data D1 used in generating the model data D3 was obtained. This allows the identification unit 75 to perform processing using the model data D3 of subjects with similar physical characteristics, without preparing individual model data D3 for each subject.

[0148] In the above description of the processing by the classification unit 75, the classification unit 75 derives features in the same manner as the feature acquisition unit 71, but the process of deriving features from the classification target data D4 among the processes performed by the classification unit 75 may be performed by the feature acquisition unit 71. Even when the classification unit 75 derives features, the feature acquisition unit 71 and the classification unit 75 may share reference to the filter 72.

[0149] Among the various matters related to the detection system 100 described above, an outline of the flow of processes performed by the information processing device 60 will be described with reference to FIGS. 20 and 21. FIG.

[0150] 20 is a flowchart showing a process flow for generating model data D3. First, myoelectric data and pressure data are acquired as teacher data (step S31). Specifically, teacher data D1 is acquired by the data acquisition unit 61. Next, teacher data is created (step S32). Specifically, identification information AD2 is assigned to the teacher data D1 acquired in step S31.

[0151] After the process of step S32, the feature is derived (step S33). Specifically, the feature acquisition unit 71 derives feature data D2 based on the myoelectric data 81 and pressure data 82 of the teacher data D1 acquired in step S31 and assigned with the identification information AD2 in step S32.

[0152] After the processing of step S33, generation of bruxism estimation model data 85 by machine learning (step S34) and generation of bruxism detailed specific model data 86 by machine learning (step S35) are performed. Specifically, based on the feature amount data D2 derived in step S33, the learning unit 73 performs machine learning as described with reference to Figs. 16 to 18 to generate model data D3, i.e., bruxism estimation model data 85 and bruxism detailed specific model data 86. The processing of steps S34 and S35 may be performed in any order.

[0153] 21 is a flowchart showing a process flow relating to estimation of the occurrence of bruxism based on the discrimination target data D4. First, myoelectric data and pressure data are acquired (step S41). Specifically, the discrimination target data D4 is acquired by the data acquisition unit 61. Next, feature amounts are derived (step S42). Specifically, based on the myoelectric data 81 and pressure data 82 of the discrimination target data D4 acquired in step S41, the discrimination unit 75 or the feature amount acquisition unit 71 generates first data including parameters similar to the first myoelectric feature amount AD4, the second myoelectric feature amount AD5, the third myoelectric feature amount AD6, and the fourth myoelectric feature amount AD7, and second data including parameters similar to the first pressure feature amount AD8 and the second pressure feature amount AD9.

[0154] After the process of step S42, the occurrence or non-occurrence of bruxism is estimated based on the teeth grinding estimation model data 85 (step S43). Specifically, the identification unit 75 treats a combination of an electromyographic feature and a pressure feature indicated by each record of the generated first data and second data as a point DP, and represents it in a {(q×α)+β}-dimensional vector space. The identification unit 75 also refers to the teeth grinding estimation model data 85 and applies a boundary indicating the occurrence or non-occurrence of bruxism (e.g., boundary SP1 or boundary SP2) among the boundaries indicating the classification of the point DP to the {(q×α)+β}-dimensional vector space. The identification unit 75 estimates whether teeth grinding is occurring based on the relationship between the DP and the applied boundary. The result of the estimation by the process of step S43 is reflected in the identification information OP.

[0155] If it is estimated in the process of step S43 that bruxism is occurring (step S44; Yes), the presence or absence of bruxism is estimated based on the bruxism detail identification model data 86 (step S45). Specifically, the identification unit 75 refers to the bruxism detail identification model data 86 and applies boundaries relating to the details of bruxism (for example, boundaries SP1, SP11, SP12, and SP13, or boundaries SP2, SP21, SP22, and SP23) among the boundaries indicating the classification of the point DP to the {(q×α)+β}-dimensional vector space. The identification unit 75 identifies the details of bruxism (type, location, strength) based on the relationship between the point DP and the boundaries applied in the process of step S43. The result of the estimation in the process of step S46 is reflected in the identification information OP.

[0156] After the process of step S45 or if it is estimated in the process of step S43 that bruxism is not occurring (step S44; No), the process proceeds to step S41 unless the operation of the information processing device 60 has ended (step S46; No). When the operation of the information processing device 60 ends (step S46; Yes), the process relating to the estimation of the occurrence of bruxism based on the recognition target data D4 ends.

[0157] 14 to 21, the determination of matters related to teeth grinding using machine learning has been described, but machine learning is not essential for determining matters related to teeth grinding, and rule-based determination may also be used. For example, as described above with reference to Fig. 9 and Fig. 10, the presence or absence of teeth grinding can be sufficiently determined based on the synchronization between the output of the pressure sensor 20 and the output of the myoelectric sensor 40 and the relationship between these outputs and thresholds (for example, a first threshold Th1 and a second threshold Th2).

[0158] As described above, according to the embodiment, the detection system 100 includes a detection device (e.g., detection device 10) that performs sensing related to teeth grinding, and an information processing device (e.g., information processing device 60) that determines the presence or absence of teeth grinding based on data indicating the result of sensing by the detection device. The detection device includes a mouthpiece (e.g., mouthpiece 21) in which pressure sensors (e.g., first sensor element 221, second sensor element 222, third sensor element 223) are arranged, an electromyographic sensor (e.g., electromyographic sensor 40) that is attached to a human cheek, and a control unit (e.g., control unit 12) that outputs data in which the output of the pressure sensor and the output of the electromyographic sensor are synchronized (e.g., electromyographic data 81 and pressure data 82 in teacher data D1, and electromyographic data 87 and pressure data 88 in data to be identified D4). This makes it possible to show the relationship between the output of the pressure sensor and the output of the electromyographic sensor generated at each time point during the period in which the output of the pressure sensor and the output of the electromyographic sensor were generated. Therefore, compared to the case where only a pressure sensor or only an electromyographic sensor is used, data due to teeth grinding and data not due to teeth grinding can be distinguished with higher accuracy.

[0159] Moreover, the information processing device (for example, the information processing device 60) includes a storage unit (for example, the storage unit 80) that stores feature data (for example, feature data D2) and model data (for example, model data D3). The feature data is data in which a myoelectric feature (for example, myoelectric feature data 83) corresponding to the output of a myoelectric sensor (for example, myoelectric sensor 40) and a pressure feature (for example, pressure feature data 84) corresponding to the output of a pressure sensor (for example, the first sensor element 221, the second sensor element 222, and the third sensor element 223) are synchronized in time. The model data is data indicating a correlation between a combination of the myoelectric feature and the pressure feature, and matters related to teeth grinding. The myoelectric feature includes data filtered to a specific frequency component from the output of the myoelectric sensor (for example, the above-mentioned M3(ω)=H3(ω)M(ω) and M4(ω)=H4(ω)M(ω)). The pressure feature quantity includes data that has been filtered to a specific frequency component from the output of the pressure sensor (for example, the above-mentioned P2(ω)=H6(ω)P1(ω)). This allows data to be classified using frequency components that are included in the data and that can be more suitably used for data classification, compared to a case where data is classified based on the sensing data as output. Therefore, data due to teeth grinding and data that is not due to teeth grinding can be distinguished with higher accuracy.

[0160] Furthermore, the matters relating to teeth grinding in the model data (e.g., model data D3) include the type of teeth grinding, and the information processing device (e.g., information processing device 60) includes a learning unit (e.g., learning unit 73) that generates the model data based on teacher data (e.g., teacher data D1) including feature data (e.g., feature data D2) and information indicating the type of teeth grinding (e.g., synchronization information AD1). This allows the information processing device to generate the model data that reflects the correlation between the feature data and the information indicating the type of teeth grinding.

[0161] Furthermore, the myoelectric feature amount (e.g., myoelectric feature amount data 83) includes data that is output from the myoelectric sensor (e.g., the myoelectric sensor 40) and filtered by a band-pass filter. This makes it possible to obtain frequency components that can be more suitably used for data classification in a simpler manner than limiting frequency components using a filter.

[0162] Furthermore, model data (e.g., model data D3) for each of multiple people (e.g., see Figures 17 and 18) who use a detection device (e.g., detection device 10) is stored in a storage unit (e.g., storage unit 80). When determining the presence or absence of teeth grinding based on new data (e.g., identification target data D4), an information processing device (e.g., information processing device 60) determines the presence or absence of teeth grinding based on the model data corresponding to the person using the detection device. This allows a determination to be made based on the model data that is more optimal for the person using the detection device. Therefore, it becomes possible to distinguish data due to teeth grinding from data that is not due to teeth grinding with a higher degree of accuracy.

[0163] Furthermore, a plurality of pressure sensors (e.g., first sensor element 221, second sensor element 222, third sensor element 223) are arranged in a mouthpiece (e.g., mouthpiece 21), and a control unit (e.g., control unit 12) outputs data that allows the outputs of the plurality of pressure sensors to be individually identified (e.g., see FIG. 9). This makes it possible to identify with high accuracy where in the human mouth pressure is occurring, based on the relationship between the respective arrangements of the plurality of pressure sensors assumed in the human mouth (e.g., see FIG. 6) and the respective outputs of the plurality of pressure sensors in the data.

[0164] Furthermore, the pressure sensors (e.g., first sensor element 221, second sensor element 222, third sensor element 223) are provided on a flexible substrate (e.g., FPC 25), which is attached to a mouthpiece (e.g., mouthpiece 21). This allows the output of the pressure sensor to be transmitted via the flexible substrate. Furthermore, the flexibility of the flexible substrate makes it easier to achieve both flexibility in response to the movements of the human mouth and a configuration for more reliable output transmission of the pressure sensor.

[0165] Furthermore, the flexible substrate (e.g., FPC 25) is bonded and fixed inside the mouthpiece (e.g., mouthpiece 21) so as to create a gap between the mouthpiece and the flexible substrate (e.g., see FIG. 4). This makes it easier to prevent pressure on the mouthpiece from acting directly on the flexible substrate. It also makes it easier to increase the flexibility and engagement of the mouthpiece with the unevenness of teeth in the human mouth.

[0166] Furthermore, the myoelectric sensor (e.g., myoelectric sensor 40) has electrodes (e.g., first electrode 42, second electrode 43, and third electrode 44), and gel (e.g., conductive gel 42a, 43a, 44a) and double-sided tape (e.g., double-sided tape for skin 45) are attached to the attachment surface of the myoelectric sensor on which the electrodes are provided. This makes it easier for a person to attach the myoelectric sensor with the electrodes more reliably close to the person's cheek.

[0167] Furthermore, the control unit (e.g., control unit 12) operates the myoelectric sensor (e.g., myoelectric sensor 40) in a first cycle until the output of the myoelectric sensor exceeds a first threshold (e.g., first threshold Th1), and operates the myoelectric sensor in a second cycle after the output of the myoelectric sensor exceeds the first threshold. The second cycle is a cycle with a higher frequency than the first cycle. The first cycle here is the low-frequency cycle in the description with reference to FIG. 11, for example. The second cycle here is the high-frequency cycle in the description with reference to FIG. 11, for example. This allows power saving to be prioritized until an output of the myoelectric sensor that may be an output corresponding to teeth grinding is generated. After an output of the myoelectric sensor that may be an output corresponding to teeth grinding is generated, more accurate sensing can be performed with a more frequent output.

[0168] Furthermore, the control unit (e.g., control unit 12) operates the pressure sensor (e.g., pressure sensor 20) in a third cycle until the output of the pressure sensor exceeds a second threshold (e.g., second threshold Th2), and operates the pressure sensor in a fourth cycle after the output of the pressure sensor exceeds the second threshold. The fourth cycle is a cycle with a higher frequency than the third cycle. The third cycle here is the low-frequency cycle in the description with reference to FIG. 12, for example. The fourth cycle here is the high-frequency cycle in the description with reference to FIG. 12, for example. This allows power saving to be prioritized until an output of the pressure sensor that may be an output corresponding to teeth grinding is generated. In addition, after an output of the pressure sensor that may be an output corresponding to teeth grinding is generated, more accurate sensing can be performed with a more frequent output.

[0169] Furthermore, the detection device 10 includes a communication unit (e.g., communication unit 13) that communicates with an external device (e.g., information processing device 60). A control unit (e.g., control unit 12) does not operate the communication unit when both the first condition and the second condition are not satisfied, and operates the communication unit when at least one of the first condition and the second condition is satisfied. The first condition is that the output of the myoelectric sensor (e.g., myoelectric sensor 40) exceeds a first threshold value (e.g., first threshold value Th1). The second condition is that the output of the pressure sensor (e.g., pressure sensor 20) exceeds a second threshold value (e.g., second threshold value Th2). This allows power saving to be prioritized until an output that may be an output corresponding to teeth grinding from at least one of the myoelectric sensor and the pressure sensor is generated. After an output that may be an output corresponding to teeth grinding is generated, more accurate sensing can be performed by more frequent output.

[0170] In the above embodiment, the learning unit 73 generates the model data D3 as an output of supervised machine learning based on SVM using the identification information AD2 of the teacher data D1 and the feature amount data D2 as input (teacher data), but this is not limited to this. For example, the identification information AD2 may be copied from the teacher data D1 to the feature amount data D2. In this case, the learning unit 73 can perform supervised machine learning by simply referring to the feature amount data D2.

[0171] Furthermore, features obtained by modifying the above-mentioned filters may be additionally included as parameters included in the myoelectric feature data 83 and the pressure feature data 84 in the feature data D2. For example, features corresponding to a median frequency (MDF: Median Frequency), a mean frequency (MNF: Mean Frequency), or the like may be added to the myoelectric feature data 83. Furthermore, features with different time average widths, and features obtained by applying a low-pass filter, a high-pass filter, or a band-pass filter may be included in the pressure feature data 84.

[0172] Furthermore, the algorithm for generating the model data D3 is not limited to SVM. In particular, the detailed bruxism model data 86 may be generated using other algorithms, such as a decision tree or logistic regression.

[0173] In addition, the specific configuration of the pressure sensor 20 is not limited to that based on the piezoelectric effect described above. For example, the specific configuration of the sensor element 22 may be a strain gauge, and a circuit including a Wheatstone bridge may be provided in the housing 41. In this case, the strain generated in each of the first sensor element 221, the second sensor element 222, and the third sensor element 223 generates an output indicating pressure. In addition, it is not essential to provide a plurality of sensor elements such as the first sensor element 221, the second sensor element 222, and the third sensor element 223, and a sensor element having a curved shape that conforms to the shape of a human tooth may be adopted. In addition, the specific configuration of the pressure sensor 20 may be a resistive film pressure sensor. However, it is preferable that the pressure sensor 20 is a film-type pressure sensor.

[0174] Furthermore, other advantageous effects brought about by the aspects described in this embodiment that are obvious from the description in this specification or that can be appropriately thought of by a person skilled in the art are naturally understood to be brought about by the present disclosure. [Explanation of symbols]

[0175] 10. Detection Device 12 Control section 20 Pressure Sensor 21 Mouthpiece 221 First sensor element 222 Second sensor element 223 Third Sensor Element 40 EMG Sensor 60 Information processing equipment 61 Data Acquisition Section 70 Arithmetic section 71 Feature acquisition unit 73 Learning Department 75 Identification section D1 Teacher data D2 Feature data D3 model data D4 Identifiable Data

Claims

1. A detection device that senses bruxism; An information processing device that determines whether or not bruxism occurs based on data indicating a result of sensing by the detection device, The detection device includes: A pressure sensor disposed in the mouthpiece; An electromyographic sensor attached to a person's cheek; A control unit that outputs data in which the output of the pressure sensor and the output of the myoelectric sensor are synchronized. Detection system.

2. A detection device that senses bruxism; An information processing device that determines whether or not bruxism occurs based on data indicating a result of sensing by the detection device, The detection device includes: A mouthpiece having a pressure sensor disposed therein; An electromyographic sensor attached to a person's cheek; a control unit that synchronizes an output of the pressure sensor and an output of the myoelectric sensor and outputs the output as the data; The information processing device includes: A storage unit that stores feature amount data and model data, the feature data is data in which a myoelectric feature corresponding to an output of the myoelectric sensor and a pressure feature corresponding to an output of the pressure sensor are synchronized in time; the model data is data indicating a correlation between a combination of the myoelectric feature amount and the pressure feature amount and matters related to teeth grinding, the myoelectric feature amount includes data obtained by filtering a specific frequency component from an output of the myoelectric sensor, The pressure feature amount includes data obtained by filtering a specific frequency component from an output of the pressure sensor. Detection system.

3. The matters related to bruxism include types of bruxism, The information processing device includes a learning unit that generates the model data based on the feature amount data and teacher data including information indicating a type of teeth grinding. The detection system of claim 2 .

4. The myoelectric feature amount includes data of the output of the myoelectric sensor filtered by a band-pass filter.

4. A detection system according to claim 2 or 3.

5. The model data for each of a plurality of people using the detection device is stored in the storage unit, the information processing device determines the presence or absence of bruxism based on the new data, based on the model data corresponding to a person using the detection device; 4. A detection system according to claim 2 or 3.

Images