JP7243439B2 - DIETARY RISK DETERMINATION METHOD, DIETARY RISK DETERMINATION PROGRAM AND INFORMATION PROCESSOR - Google Patents

Description

The present invention relates to a dietary risk determination method, a dietary risk determination program, and an information processing apparatus.

At nursing care sites, dysphagia may increase the risk of suffocation, so a lot of time is spent accompanying meals. As a method to prevent dysphagia, a method of preventing throat blockage by promoting saliva secretion is used by performing swallowing exercises such as tilting the neck to the left and right and puffing out the mouth before meals. In addition, since the time during which the face is moved during conversation before meals has the same effect as the swallowing exercise, methods such as taking longer conversation time are also used. In recent years, there is also known a method of measuring the swallowing function of a care-receiver by mixing volatile compounds into meals and measuring the presence of volatile compounds in exhalation.

JP 2012-544659 A

However, the above technology requires special equipment such as mixing volatile compounds into meals, measuring the presence of volatile compounds, etc., so it cannot be used easily and cannot be used for general purposes. It is difficult to detect dysphagia in care recipients.

In addition, adding a thickening agent to the meal may be an option in order to prevent the meal from becoming boring. difficult to use.

An object of one aspect of the present invention is to provide a dietary risk determination method, a dietary risk determination program, and an information processing apparatus capable of detecting dysphagia in a care-receiver.

In the first proposal, the dietary risk determination method is such that the computer determines the oral function of the care-receiver that does not meet a preset standard among a plurality of oral functions that constitute a series of actions when swallowing a meal. Perform a process to identify certain missing features. In the risk determination method, the computer determines whether or not the operation corresponding to the insufficient function is performed before eating, based on the image data of the care-receiver. In the risk determination method, the computer performs a process of issuing an alert regarding the risk of dysphagia based on the determination result.

In one aspect, dysphagia of a care recipient can be detected.

FIG. 1 is a diagram illustrating an overall example of an information processing apparatus according to a first embodiment; FIG. 2 is a functional block diagram of the functional configuration of the information processing apparatus according to the first embodiment; FIG. 3 is a diagram illustrating an example of oral cavity functions. FIG. 4 is a diagram illustrating generation of a reference model for each function. FIG. 5 is a diagram for explaining identification of deficient oral cavity functions. FIG. 6 is a diagram for explaining generation of learning data. FIG. 7 is a diagram explaining generation of a learning model. FIG. 8 is a diagram explaining prediction of dysphagia. FIG. 9 is a diagram explaining prediction of dysphagia. FIG. 10 is a flowchart showing the flow of processing. FIG. 11 is a diagram illustrating a hardware configuration example.

Hereinafter, embodiments of the dietary risk determination method, the dietary risk determination program, and the information processing apparatus disclosed in the present application will be described in detail based on the drawings. In addition, this invention is not limited by this Example. Moreover, each embodiment can be appropriately combined within a range without contradiction.

[Description of information processing device]
FIG. 1 is a diagram illustrating an overall example of an information processing apparatus 10 according to a first embodiment. The information processing apparatus 10 shown in FIG. 1 predicts the risk of dysphagia of the care recipient before eating using the care recipient's image data, nursing care diary, etc., and notifies the caregiver, so that the caregiver It is a computer device that reduces the burden on caregivers and reduces the risk of suffocation for caregivers. Specifically, as shown in FIG. 1, the information processing device 10 executes a preliminary phase, a learning phase, and a prediction phase.

For example, in the preliminary phase, the information processing apparatus 10 uses data of a plurality of caregivers to create a reference model that is an index for determining the state of a plurality of oral functions that constitute a series of actions when swallowing a meal. to generate

Subsequently, in the learning phase, the information processing apparatus 10 compares each reference model corresponding to each of a plurality of oral functions with the state of each oral function of the subject (care recipient), and the subject who does not meet the criteria Identify oral function (hereinafter sometimes referred to as deficient function). Then, the information processing apparatus 10 generates a learning model for predicting the dysphagia of the subject for each of the specified deficient functions.

After that, in the prediction phase, the information processing apparatus 10 predicts the state of the subject's lack of function on the day of determination using each learned model. Then, when it is predicted that there is a risk of developing dysphagia, the information processing device 10 notifies the caregiver of an alert.

In this way, the information processing apparatus 10 can identify the missing function for each care-receiver, learn a learning model corresponding to the missing function, and predict the risk of occurrence of dysphagia only for the missing function. can. As a result, the information processing apparatus 10 can detect the dysphagia of the care-receiver while reducing the burden on the care-receiver. Although an example in which the information processing apparatus 10 executes each phase has been described here, the present invention is not limited to this, and separate information processing apparatuses may execute each phase.

[Function configuration]
FIG. 2 is a functional block diagram of the functional configuration of the information processing apparatus 10 according to the first embodiment. As shown in FIG. 2 , the information processing device 10 has a communication section 11 , a storage section 12 and a control section 20 .

The communication unit 11 is a processing unit that controls communication with other devices, such as a communication interface. For example, the communication unit 11 transmits an alert to a caregiver's computer device such as a tablet or a smartphone, or receives an instruction to start processing from a caregiver or administrator.

The storage unit 12 is an example of a storage device that stores various data, programs executed by the control unit 20, and the like, and is, for example, a memory or a hard disk. The storage unit 12 stores a history information DB 13, a measured value DB 14, a learning data DB 15, and a learning result DB 16.

History information DB13 is a database which memorize|stores the image data of a cared person for every cared person. For example, the history information DB 13 stores image data when each of a plurality of oral functions constituting a series of actions when swallowing a meal is executed in association with each care-receiver.

The measured value DB 14 is a database that stores the plurality of oral function measurement results for each care recipient. Specifically, the measurement value DB 14 stores the measurement results of each perioral muscle for range-of-motion training, the measurement results of each swallowing function for swallowing training, the measurement results for each masticatory function for muscle load training, and the like.

FIG. 3 is a diagram illustrating an example of oral cavity functions. As shown in FIG. 3, the oral function includes multiple training items, and each training item is subdivided into multiple functions. Also, for each function, a determination method and determination conditions for determining whether the function is normal are defined.

Range-of-motion training measures the state of the range of motion around the tongue and aviation, and has the functions of perioral muscle A, perioral muscle B, perioral muscle C, and perioral muscle D. The perioral muscle A is determined by the rate at which the tongue can be kept protruding to reach the tip of the chin (or the tip of the nose) within 10 seconds. The perioral muscle B is determined by the rate at which the mouth is kept wide open for 10 seconds, as in the pronunciation of "a". The perioral muscle C is judged by the rate at which the mouth is kept wide open for 10 seconds, as in the pronunciation of "i". The measurement of the perioral muscle C is determined by the rate at which the mouth can be kept wide open for 10 seconds, similar to the pronunciation of "U".

The swallowing training includes swallowing function A and swallowing function B, which is to induce a swallowing reflex and measure the state of swallowing motion. Swallowing function A is measured by touching the anterior palatine arch or the posterior pharyngeal wall for 10 seconds to induce a swallowing reflex, and measuring the number of seconds before the swallowing reflex occurs. Swallowing function B was measured by pressure stimulation of the K-point for 10 seconds to induce a swallowing reflex, and the number of seconds in which the swallowing reflex occurred. The position of the K-point is generally defined as slightly posterior and medial to the rearmost part of the retromolar triangle.

Muscle load training measures the state of muscle strength around the tongue and aircraft, and has masticatory function A, masticatory function B, and masticatory function C. Masticatory function A was determined by the number of times the spoon was pushed from above the tongue 10 times and the tongue was able to push it back. Masticatory function B is determined by the number of times the straw can be resisted by pulling the straw between the lips 10 times. Masticatory function C is determined by the number of times the cheek is puffed up and pressed against the cheek with a finger and the force is resisted.

The learning data DB 15 is a database that stores learning data for each cared person. For example, the learning data DB 15 stores the measurement value of the insufficient function of each cared person and the presence or absence of dysphagia in association with each other. Here, the measured value of insufficient function is used as an explanatory variable, and the presence or absence of dysphagia is used as an objective variable.

The learning result DB 16 is a database that stores learning results. For example, the learning result DB 16 stores various parameters of a learned learning model. By using various parameters stored here, a learned learning model can be constructed.

The control unit 20 is a processing unit that controls the entire information processing apparatus 10, such as a processor. The control unit 20 has an imaging unit 30 , a preprocessing unit 40 , a learning processing unit 50 and a prediction processing unit 60 .

The image capturing unit 30 is a processing unit that captures an image of each cared person using an image capturing device such as a camera. For example, the imaging unit 30 causes each cared person to perform a movement specified by the determination method of each function shown in FIG. Note that the imaging unit 30 takes images every day and stores them in the history information DB 13 .

The preprocessing unit 40 is a processing unit that prepares in advance for learning and prediction. Specifically, the preprocessing unit 40 uses the image data of each function of each care-receiver stored in the history information DB 13 to create a time-series reference model for determining whether each function is normal. Generate.

FIG. 4 is a diagram illustrating generation of a reference model for each function. As shown in FIG. 4, the preprocessing unit 40 analyzes the image data of each function and determines the state of each function. For example, when the pre-processing unit 40 determines the state of the perioral muscle A of "Mr. A" who is the care recipient, the image of "a state in which the tongue protrudes so as to reach the tip of the chin (or the tip of the nose)" is imaged 10 times. Then, the preprocessing unit 40 identifies the coordinates of the mouth and tongue by image analysis for each image data, and identifies the range of motion of the mouth and tongue. After that, the pre-processing unit 40 determines the number of times (for example, 7 times) that the range of motion of the tongue is equal to or greater than the threshold value and that "the tongue is projected so as to reach the tip of the chin (or the tip of the nose)" out of the 10 times. to count. After that, the preprocessing unit 40 sets “7/10=0.7” for the perioral muscle A of Mr. A in the measured value DB 14 .

To explain another example, when determining the state of the swallowing function A of “Mr. Take an image of the state where the Then, the preprocessing unit 40 identifies the coordinates of the mouth and throat by image analysis of the image data, and identifies the range of motion of the mouth and throat. After that, the preprocessing unit 40 counts the time (for example, 3 seconds) after the start of the measurement state during which the range of motion of the mouth and throat is equal to or greater than the threshold and it can be determined that the swallowing reflex has occurred. After that, the preprocessing unit 40 sets "3/10=0.3" for Mr. B's swallowing function A in the measured value DB 14 .

To explain another example, when determining the state of masticatory function B of “Mr. Take an image. Then, the preprocessing unit 40 identifies the coordinates of the straw and lips by image analysis for each image data, and identifies the range of motion of the straw and lips. After that, the pre-processing unit 40 counts the number of times (for example, 8 times) that the distance between the straw and the lips is equal to or greater than the threshold and that the "straw is separated from the lips" can be determined out of the 10 times. After that, the preprocessing unit 40 sets “(10−8)/10=0.2” to the masticatory function B of Mr. A in the measured value DB 14 .

Returning to FIG. 4, in this manner, the preprocessing unit 40 analyzes the image data of each function shown in FIG. , and the result is stored in the measured value DB 14 . After that, the preprocessing unit 40 performs modeling for each function to generate a reference model for each function.

For example, when the preprocessing unit 40 generates the reference model of the perioral muscle A, each cared person is plotted on a graph in which the horizontal axis is the age and the vertical axis is the determination result (calculated ratio) of the perioral muscle A. Plot the measurements of . Then, the preprocessing unit 40 generates an approximate straight line that minimizes the error from each plot value, and determines the approximate straight line as the reference model of the muscles A around the oral cavity. In this manner, the preprocessing unit 40 generates a reference model for each function, stores it in the storage unit 12, and outputs it to the learning processing unit 50. FIG. As for the error, a known method such as the least squares method can be employed, and as for the approximate straight line, another known method such as an approximate curve can be employed.

Returning to FIG. 2, the learning processing unit 50 includes a shortage determination unit 51, a learning data generation unit 52, and a learning unit 53, and generates a learning model for predicting the occurrence of dysphagia for each care recipient. processing unit.

The shortage determination unit 51 is a processing unit that determines a shortage function for each care recipient. Specifically, the lack determination unit 51 compares the reference model and the measured value (actual value) of each function for each care recipient, and determines the lack function, which is the oral function of the subject who does not meet the criteria. .

FIG. 5 is a diagram for explaining identification of deficient oral cavity functions. As shown in FIG. 5 , the deficiency determination unit 51 refers to the measured value DB 14 and acquires measured values of each function such as the perioral muscle A and the swallowing function A measured on a specific date specified in advance. Subsequently, the shortage determination unit 51 determines that the function satisfies the criterion when the measured value is positioned above the reference model, and determines that the function satisfies the criterion when the measured value is positioned below the reference model. judged not to meet the criteria.

In the example of FIG. 5, the deficiency determination unit 51 determines that Mr. A has the perioral muscles A, the perioral muscles C, and the swallowing function A, and the perioral muscles B, the perioral muscles D, and the masticatory function A satisfy the criteria. It indicates that it was determined that the criteria were not met. Then, the deficiency determination unit 51 outputs the perioral muscle B, the perioral muscle D, and the masticatory function A to the learning data generation unit 52 as insufficient functions determined not to satisfy the criteria. In this way, the shortage determination unit 51 identifies the missing functions of each care receiver and outputs them to the learning data generation unit 52 .

The learning data generation unit 52 is a processing unit that generates learning data for learning a learning model corresponding to each insufficient function for each cared person. Specifically, the learning data generation unit 52 uses the measured value of the caregiver's insufficient function stored in the measured value DB 14 as an explanatory variable for each care-receiver's insufficient function, and the learning data generation unit 52 can acquire the care diary or the like. Learning data having occurrence of dysphagia as an objective variable is generated and stored in the learning data DB 15 .

FIG. 6 is a diagram for explaining generation of learning data. As shown in FIG. 6, the learning data generation unit 52 generates learning data corresponding to each of the perioral muscles B, the perioral muscles D, and the masticatory function A, which are identified as insufficient functions for Mr. A. As shown in FIG.

Specifically, the learning data generation unit 52 sets the “explanatory variable, objective variable” for Mr. A’s perioral muscle B as “measured value (0.4) on May 1, record of occurrence of dysphagia ( Yes)”, “measured value on May 2nd (0.7), record of dysphagia (no record)”, “measured value on May 3rd (0.5), record of dysphagia (yes) ” and so on to generate supervised learning data.

Similarly, the learning data generation unit 52 sets the “explanatory variable, objective variable” for Mr. A’s perioral muscle D as “measured value on May 1 (0.5), record of occurrence of dysphagia (present) ”, “measured value on May 2nd (0.5), dysphagia recorded (none)”, “measured value on May 3rd (0.5), dysphagia recorded (yes)”, etc. Generate supervised learning data with

Similarly, the learning data generation unit 52 sets “measured value on May 1 (0.5), occurrence record of dysphagia (yes)” as “explanatory variable, objective variable” for Mr. A’s masticatory function A. , "Measured value on May 2nd (0.5), dysphagia recorded (no)", "Measured value on May 3rd (0.2), dysphagia recorded (yes)", etc. Generate supervised training data that

By doing so, the learning data generation unit 52 can generate learning data only for functions that are lacking, not for all oral functions that constitute a series of actions when swallowing a meal, for each care recipient. can.

The learning unit 53 is a processing unit that learns a learning model for predicting the occurrence of dysphagia for each of the care-receiver's insufficient functions. Specifically, the learning unit 53 acquires learning data corresponding to each insufficient function of each care-receiver from the learning data DB 15, and performs machine learning or the like using the acquired learning data. Then, after completing the learning, the learning unit 53 stores “the cared person, the missing function, the learning result” in the learning result DB 16 . Note that the timing for ending learning can be arbitrarily set, such as when learning using a predetermined number or more of learning data is completed, or when the restoration error becomes less than a threshold.

Here, an example using a logistic regression model (hereinafter sometimes simply referred to as logistic regression) will be described as an example of a learning model. The learning unit 53 inputs the linear regression equation shown in equation (1) to the sigmoid function shown in equation (2), and acquires the value obtained by equation (2) as logistic regression. Here, an explanatory variable is input to X in Equation (1), y is an objective variable, w is a weight, and b is a prespecified constant.

FIG. 7 is a diagram explaining generation of a learning model. As shown in FIG. 7, when the learning unit 53 generates a learning model of Mr. A's perioral muscle B, the explanatory variable "0.4" for May 1 is input to X in Equation (1), The weight w is calculated by inputting the objective variable “yes=1.0” into y. Similarly, the learning unit 53 inputs the explanatory variable “0.7” for May 2nd to X in Equation (1), inputs the objective variable “none=0.0” to y, and calculates the weight w. Then, the explanatory variable “0.5” for May 3 is input to X in Equation (1), and the objective variable “Yes=1.0” is input to y to calculate the weight w.

In this way, the learning unit 53 performs model learning that optimizes the weight w so as to match the learning data, substitutes the formula (1) using the learned weight w into the formula (2), and obtains Generate a logistic regression. For Mr. A, the learning unit 53 generates a logistic regression for each of the perioral muscle B, the perioral muscle D, and the masticatory function A that are identified as lacking functions. In logistic regression, the closer the output value is to 1, the higher the risk of occurrence of dysphagia, the closer the output value is to 0, the lower the risk of occurrence of dysphagia, and when the output value is above the threshold, swallowing It is determined that the occurrence probability of failure is high.

Returning to FIG. 2, the prediction processing unit 60 has a prediction unit 61 and a notification unit 62. A processing unit that predicts the occurrence of dysphagia for each care recipient and executes notification processing according to the prediction result. is.

The prediction unit 61 is a processing unit that predicts the risk of occurrence probability of dysphagia for each care-receiver before meals using each learning model corresponding to the function deficiency of each care-receiver. Specifically, the prediction unit 61 reads out the learning results corresponding to the insufficient functions of the person to be predicted from the learning result DB 16, and constructs a learned learning model corresponding to each insufficient function. Subsequently, the prediction unit 61 calculates the measurement value of each insufficient function from the image data before the meal on the predicted date of the person to be predicted. Then, the prediction unit 61 inputs the measured value of each insufficient function to the learned learning model corresponding to each insufficient function, and acquires each output result. After that, based on each output result, the prediction unit 61 outputs that fact to the notification unit 62 when it is predicted that the probability of occurrence of dysphagia is high.

8 and 9 are diagrams for explaining prediction of dysphagia. As shown in FIG. 8, the prediction of the occurrence of dysphagia is determined by the value of the objective variable corresponding to the deficit measure in the logistic regression. In the case of the explanatory variables shown in (1) of FIG. 8, since the objective variable "1" has a shorter distance from the logistic regression than the objective variable "0", it is predicted that there is dysphagia. On the other hand, in the case of the explanatory variables shown in (2) of FIG. 8, since the objective variable "0" has a shorter distance from the logistic regression than the objective variable "1", it is predicted that there is no dysphagia.

The prediction unit 61 predicts that the prediction target person is at risk of developing dysphagia when it is predicted that one of the insufficient functions of the prediction target person has dysphagia. Specifically, as shown in (1) of FIG. 9 , the prediction unit 61 predicts, for each of the perioral muscle B, the perioral muscle D, and the masticatory function A, which are the insufficient functions of Mr. A, who is the person to be predicted, Dysphagia is predicted based on measurements on the day of prediction and logistic regression. Then, when "dysphagia" is detected for the perioral muscle A, the prediction unit 61 predicts that Mr. A has a high risk of dysphagia.

On the other hand, as shown in (2) of FIG. 9 , the prediction unit 61 performs measurement on the prediction day for each of the perioral muscles A, the swallowing function A, and the masticatory function B, which are insufficient functions of Mr. B, who is the person to be predicted. Predict dysphagia based on values and logistic regression. Then, when "no dysphagia" is detected for any of the missing functions, the prediction unit 61 predicts that Mr. B has a low risk of dysphagia.

The notification unit 62 is a processing unit that issues an alarm to a caregiver when the occurrence of dysphagia is predicted. For example, when the occurrence of dysphagia for Mr. A is predicted, the notification unit 62 transmits a message to the caregiver's smartphone or the like to call attention such as "Mr. A is in danger of dysphagia, be careful."

To give another example, the notification unit 62 instructs a terminal device or AR glasses (augmented reality glasses) possessed by a caregiver to monitor a care-receiver who is likely to develop dysphagia. to output In addition, the notification unit 62 causes the communication robot to coach the care recipient to perform swallowing exercises until the risk of the care recipient having a high probability of developing dysphagia decreases. For example, the communication robot helps the care recipient with meals and monitors dysphagia on behalf of the caregiver.

[Process flow]
Next, a flow of a series of processes executed by the information processing apparatus 10 described above will be described. FIG. 10 is a flowchart showing the flow of processing. Here, a series of flow of the preliminary phase, the learning phase, and the prediction phase will be described, but each phase can also be executed at different timings.

As shown in FIG. 10, when the administrator or the like instructs the preprocessing unit 40 to start processing (S101: Yes), the preprocessing unit 40 collects history information captured by the imaging unit 30 from the history information DB 13 (S102). , a time-series model (reference) is generated for each oral cavity function (S103).

Subsequently, when the administrator or the like instructs to start learning (S104: Yes), the learning processing unit 50 collects the measurement results of the subject (S105), compares each standard with the measurement results, A function is specified (S106). Then, the learning processing unit 50 generates learning data corresponding to each missing function, and generates a logistic regression model (learning model) of each missing function by model learning using each learning data (S107).

After that, when the administrator or the like instructs to start prediction (S108: Yes), the prediction processing unit 60 acquires the target person's measurement value of the day from the measurement value DB 14 (S109), The functional measurements are extracted (S110).

Subsequently, the prediction processing unit 60 predicts dysphagia based on the results obtained by inputting the measured values of each function deficiency into the logistic regression model of each function deficiency (S111). Then, when "dysphagia is present" is predicted for any of the missing functions (S112: Yes), the prediction processing unit 60 issues an alarm to the caregiver (S113). Note that the prediction processing unit 60 executes S114 without executing S113 when "dysphagia" is predicted for none of the missing functions (S112: No).

Thereafter, the prediction processing unit 60 repeats S108 and subsequent steps when continuing the prediction processing (S114: No), and ends the processing when ending the prediction processing (S114: Yes).

[effect]
As described above, the information processing apparatus 10 performs modeling using a record of dysphagia (choking during meals) from a care diary or the like and information (image data, etc.) on facial movement before meals. Then, the information processing apparatus 10 can estimate the risk of dysphagia by inputting the image data before meals. Therefore, the information processing apparatus 10 can detect the dysphagia of the care-receiver. In addition, since the information processing apparatus 10 can detect the risk of occurrence of details of swallowing in advance, it is possible to realize optimal assignment of caregivers to caregivers, thereby reducing the extra burden on the caregivers.

Although the embodiments of the present invention have been described so far, the present invention may be implemented in various different forms other than the embodiments described above.

[Data, figures, etc.]
The data examples, numerical values, display examples, etc. used in the above embodiments are only examples, and can be arbitrarily changed. A plurality of oral cavity functions constituting a series of actions when swallowing a meal is also an example, and other information can also be used.

[Learning model]
In the above embodiment, an example using a logistic regression model as a learning model has been described, but the learning model is not limited to this, and other machine learning such as neural networks and support vector machines can also be adopted. Prediction using a learned model may be performed at any time before meals on the day to be determined.

[Learning data]
In the above embodiment, the execution time and the number of executions of swallowing exercises are used as oral function determination and learning data. can.

[Reference model]
Further, the reference model may be set in advance by statistical processing using information of each care recipient, information of care recipients who do not have dysphagia, and the like. In addition, an example of using the measurement results of each care recipient when generating the reference model has been described, but the present invention is not limited to this. It is also possible to generate a reference model using only the person's measurements.

[Determination by swallowing exercises]
In the above example, an example of determining the risk of filing a dysphagia application using a learning model of insufficient function was described, but it is not limited to this, and determination is made from image data without generating a learning model. You can also

For example, the information processing apparatus 10 captures an image of the subject determined to have insufficient function before eating. Subsequently, the information processing apparatus 10 analyzes the image data before the meal, and uses the coordinates and the like to identify the movement of the mouth and the like, the distance of the movement, and the like. Then, the information processing apparatus 10 pre-stores movements that specify the swallowing target and each oral function, analyzes the pre-meal image data, and determines whether or not the movement of each oral cavity function has been performed. After that, the information processing device 10 notifies the caregiver of an alarm when an unimplemented oral cavity function is detected.

[system]
Information including processing procedures, control procedures, specific names, and various data and parameters shown in the above documents and drawings can be arbitrarily changed unless otherwise specified.

Also, each component of each device illustrated is functionally conceptual, and does not necessarily need to be physically configured as illustrated. That is, the specific forms of distribution and integration of each device are not limited to those shown in the drawings. That is, all or part of them can be functionally or physically distributed and integrated in arbitrary units according to various loads and usage conditions.

Further, each processing function performed by each device may be implemented in whole or in part by a CPU and a program analyzed and executed by the CPU, or implemented as hardware based on wired logic.

[hardware]
FIG. 11 is a diagram illustrating a hardware configuration example. As shown in FIG. 11, the information processing device 10 has a communication device 10a, a HDD (Hard Disk Drive) 10b, a memory 10c, and a processor 10d. 11 are interconnected by a bus or the like.

The communication device 10a is a network interface card or the like, and communicates with other devices. The HDD 10b stores programs and DBs for operating the functions shown in FIG.

The processor 10d reads from the HDD 10b or the like a program that executes the same processing as each processing unit shown in FIG. 2 and develops it in the memory 10c, thereby operating the process of executing each function described with reference to FIG. 2 and the like. For example, this process executes the same function as each processing unit of the information processing apparatus 10 . Specifically, the processor 10d reads a program having functions similar to those of the imaging unit 30, the preprocessing unit 40, the learning processing unit 50, the prediction processing unit 60, and the like, from the HDD 10b and the like. Then, the processor 10d executes the same processes as those of the imaging unit 30, the preprocessing unit 40, the learning processing unit 50, the prediction processing unit 60, and the like.

Thus, the information processing apparatus 10 operates as an information processing apparatus that executes a learning method by reading and executing a program. Further, the information processing apparatus 10 can read the program from the recording medium by the medium reading device and execute the read program, thereby realizing the same function as the embodiment described above. Note that the programs referred to in other embodiments are not limited to being executed by the information processing apparatus 10 . For example, the present invention can be applied in the same way when another computer or server executes the program, or when they cooperate to execute the program.

REFERENCE SIGNS LIST 10 information processing device 11 communication unit 12 storage unit 13 history information DB
14 Measured value DB
15 Learning data DB
16 Learning result database
20 control unit 30 imaging unit 40 preprocessing unit 50 learning processing unit 51 shortage determination unit 52 learning data generation unit 53 learning unit 60 prediction processing unit 61 prediction unit 62 notification unit

Claims (10)

the computer
Among multiple oral functions that constitute a series of actions when swallowing a meal, identify the insufficient function, which is the oral function of the care recipient that does not meet the preset criteria,
Determining whether or not an operation corresponding to the lacking function is being performed before eating based on image data of the care-receiver;
A method for judging a dietary risk, characterized by executing a process of notifying an alert regarding a risk of dysphagia based on a judgment result. The computer executes a process of generating a plurality of reference models corresponding to each of the plurality of oral functions using the accumulated measured values regarding the plurality of oral functions of the plurality of care recipients,
2. The process of specifying, according to claim 1, characterized in that the plurality of reference models and the measured values relating to each of the plurality of oral functions of the care recipient are compared to specify the lacking function. Dietary risk determination method. The computer executes a process of learning a learning model in which the measurement value related to the insufficient function among the plurality of oral functions corresponding to the care recipient is an explanatory variable, and the presence or absence of the dysphagia occurrence is an objective variable,
The determination process is characterized by determining the risk of the occurrence probability of the dysphagia based on the output result obtained by inputting the measurement value related to the insufficient function measured on the determination date into the learning model. The dietary risk determination method according to claim 1 or 2. The determination process is characterized in that, from the image data of the care-receiver captured before eating on the determination date, a movement corresponding to the lacking function is identified, and the measured value is generated based on the movement. The dietary risk determination method according to claim 3. The identifying process identifies the missing function for each care receiver to be determined,
The learning process generates the learning model for each missing function of each care recipient,
5. The dietary risk according to claim 4, wherein the judging process judges the risk of occurrence probability of the dysphagia for each care recipient using the learning model corresponding to the deficient function. judgment method. The notification process includes outputting at least one of an image, sound, and vibration to a terminal device possessed by a caregiver to instruct the care-receiver to monitor the care-receiver whose probability of occurrence of dysphagia is equal to or greater than a threshold. The dietary risk determination method according to any one of claims 1 to 5, characterized by: The notification processing is to instruct the communication robot to coach the care recipient to perform swallowing exercises until the risk of the care recipient having the dysphagia occurrence probability of a threshold value or higher decreases. The method for judging dietary risk according to any one of claims 1 to 5. 8. The dietary risk determination method according to any one of claims 1 to 7, wherein an alert regarding the risk of occurrence of dysphagia is notified based on the execution status of the swallowing exercise corresponding to the lacking function. to the computer,
Among multiple oral functions that constitute a series of actions when swallowing a meal, identify the insufficient function, which is the oral function of the care recipient that does not meet the preset criteria,
Determining whether or not an operation corresponding to the lacking function is being performed before eating based on image data of the care-receiver;
A dietary risk determination program characterized by executing a process of notifying an alert regarding the risk of dysphagia based on the determination result. an identifying unit that identifies, among a plurality of oral functions constituting a series of actions when swallowing a meal, a deficient function that is an oral function of a care-receiver that does not meet a preset standard;
a determination unit that determines whether or not an action corresponding to the lacking function is being performed before eating, based on image data of the care-receiver;
and an information processing device, comprising: a notification unit that provides an alert regarding a risk of dysphagia based on a determination result.

Images