Information processing device, information processing method, and program

Abstract

To provide technology that reduces the time burden required to evaluate motor symptoms in Parkinson's disease patients and enables frequent follow-up by estimating a comprehensive score with high accuracy from a small number of tasks. [Solution] The information processing device 1 includes an image acquisition unit 521 that acquires images of a Parkinson's disease patient performing some of several pre-set tasks from among several motor tasks, a score extraction unit 522 that extracts individual scores corresponding to each of the pre-set tasks from the images, and a total score estimation unit 523 that estimates a total score based on the individual scores of the pre-set tasks. The pre-set tasks are fewer than the total number of motor tasks and are tasks selected by Recursive Feature Elimination (RFE).

Description

Technical Field

[0001] The present invention relates to an information processing apparatus, an information processing method, and a program.

Background Art

[0002] At present, for the motor symptom evaluation of Parkinson's disease patients, the Unified Parkinson's Disease Rating Scale (UPDRS) is used as an internationally standardized evaluation scale. Part 3 of the UPDRS is composed of items related to motor symptoms, and includes a large number of items such as finger tapping, hand movement, pronation-supination movement, foot tapping, leg agility, standing up, posture, and walking.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of Invention

Problems to be Solved by the Invention

[0004] Regarding the motor symptom evaluation of Parkinson's disease, it is necessary to evaluate a large number of items individually, which takes a certain amount of time. In addition, frequent follow-up was difficult because the evaluation took time. There is a demand for reducing the time burden required for evaluating motor symptoms of Parkinson's disease patients and realizing frequent follow-up observation by estimating a comprehensive score with high accuracy from a small number of tasks, but it has been difficult to provide such a function with conventional techniques including the technique of Patent Document 1.

[0005] The present invention has been made in view of such circumstances, and an object thereof is to provide a technique that reduces the time burden required for evaluating motor symptoms of Parkinson's disease patients and enables realization of frequent follow-up observation by estimating a comprehensive score with high accuracy from a small number of tasks. [Means for solving the problem]

[0006] To achieve the above objective, an information processing apparatus according to one aspect of the present invention is: An information processing device for estimating the overall score of motor symptoms in a Parkinson's disease patient, An image acquisition unit that acquires images of a Parkinson's disease patient performing some of several pre-set motor tasks, A score extraction unit extracts individual scores corresponding to each of the aforementioned pre-set tasks from the aforementioned image, A total score estimation unit that estimates the total score based on the individual scores of some of the pre-set tasks, Equipped with, Some of the aforementioned pre-configured tasks are These are tasks that are fewer in number than the total number of the aforementioned motor tasks, and are selected by Recursive Feature Elimination (RFE).

[0007] Each of the information processing method and program according to one aspect of the present invention corresponds to each of the method and program corresponding to the information processing apparatus according to one aspect of the present invention. [Effects of the Invention]

[0008] According to the present invention, it is possible to reduce the time burden required to evaluate the motor symptoms of Parkinson's disease patients and to enable frequent follow-up by estimating a comprehensive score with high accuracy from a small number of tasks. [Brief explanation of the drawing]

[0009] [Figure 1] This is a schematic diagram showing the overall configuration of an estimation system according to one embodiment of the present invention. [Figure 2] Figure 1 shows the installation status of the imaging unit and information processing device in the estimation system. [Figure 3]A diagram showing the installation state of the apparatus main body of the information processing device in the estimation system of Figure 1, and a diagram showing a swivel chair with a lock function. [Figure 4] A block diagram showing an example of the hardware configuration of the information processing device in the estimation system of Figure 1. [Figure 5] A functional block diagram showing an example of the functional configuration of the information processing device of Figure 4 in the estimation system of Figure 1. [Figure 6] A diagram showing a display example of navigation information. [Figure 7] A diagram showing a display example of a comprehensive score estimation result. [Figure 8] A diagram showing a list of motor symptom surveys included in Part 3 of UPDRS. [Figure 9] A diagram showing a comparison of comprehensive score estimation results from a small number of tasks. [Mode for Carrying Out the Invention]

[0010] Hereinafter, embodiments of the present invention will be described with reference to the drawings.

[0011] First, with reference to Figure 1, the overall configuration of the estimation system according to one embodiment of the present invention will be described. Figure 1 is a schematic diagram showing the overall configuration of the estimation system according to one embodiment of the present invention.

[0012] As shown in Figure 1, the estimation system 100 of the present embodiment is a system for estimating motor symptoms of a Parkinson's disease patient, and includes an information processing device 1 and an imaging unit 2 that captures the movement of the Parkinson's disease patient. In addition, the estimation system 100 further includes a swivel chair 3 with a lock function for the Parkinson's disease patient to sit on during imaging, and a table 4 for installing the information processing device 1 and the imaging unit 2.

[0013] The information processing device 1 and the imaging unit 2 are installed on the table 4 as described above, and the swivel chair 3 with a lock function is arranged at a position approximately 2 meters away from the table 4. Note that the arrangement is an example, and since the table 4 may be moved, the distance is not limited to the aforementioned 2 m.

[0014] With reference to Figure 1 as well as Figure 2 and Figure 3, various devices constituting the estimation system 100 (the information processing device 1, the imaging unit 2, the swivel chair 3 with a lock function, and the table 4) will be described. Figure 2 is a diagram showing the installation state of the imaging unit and the information processing device in the estimation system of Figure 1. Further, Figure 3 is a diagram showing the installation state of the device main body of the information processing device in the estimation system of Figure 1, and a diagram showing the swivel chair with a lock function.

[0015] First, the table 4 will be described. For the table 4, a table with casters is employed herein. This table 4 with casters is configured to include a top plate 4a, a frame 4b, leg portions 4c, and casters 4d.

[0016] The top plate 4a is a plate-shaped portion with flat upper and lower surfaces. On the upper surface of such a top plate 4a, a keyboard (one of the input units 16) and a display device (one of the output units 17), which will be described later, are placed. On the lower surface of the top plate 4a, the device main body 40 of the information processing device 1 is arranged in a fixed state. Note that the arrangement of the device main body 40 is an example, and if sufficient space can be secured on the upper surface of the top plate 4a, the device main body may be installed on the upper surface.

[0017] On the lower surface of the top plate 4a, in addition to the device main body 40, a power strip 43 is arranged in a fixed state, and wiring 41 is routed in a bundled form. The AC adapter 42 at the terminal of the wiring 41 is connected to the power strip 43. The imaging unit 2 is attached to the corner of the top plate 4a on the side close to the swivel chair 3 with a lock function.

[0018] The frame 4b is a cylindrical member that supports the tabletop 4a, with its upper end fixed to the underside of the tabletop 4a and its lower end fixed to the bridge portion of the leg portion 4c, which will be described later. The leg portion 4c here has a pair of leg bodies and a bridge portion connecting the middle of this pair of leg bodies, and is formed in an "H" shape in plan view.

[0019] Each of the four ends of the pair of leg bodies is fitted with a caster 4d. The caster 4d has a locking mechanism, and when the lock is released, the table 4 can be moved. When the locking mechanism is locked, the movement of table 4 is restricted, which stabilizes, for example, the shooting by the shooting unit 2.

[0020] Next, I will explain the second photography unit. The imaging unit 2 is comprised of an imaging unit body 2a, a support column 2b, and a table fixing part 2c.

[0021] The main unit 2a of the camera is, for example, a webcam itself, which can capture and visualize the movements of a Parkinson's disease patient. Furthermore, while "creating a video" generally refers to making a video, it also includes creating images that include multiple still images. The main imaging unit 2a is attached to the support column 2b towards the swivel chair 3 with a locking function in which the Parkinson's disease patient sits. The support column 2b is a straight, rod-shaped member that can position the imaging unit body 2a in the vertical direction. The table fixing part 2c is provided at the lower end of the support column part 2b, enabling the tabletop 4a to be fixed to, for example, one corner. The imaging unit 2 is positioned close to the Parkinson's disease patient at table 4 (in other words, the imaging unit 2 is positioned close to the swivel chair 3 with a locking function).

[0022] Next, we will describe the swivel chair 3 with a locking function. The locking swivel chair 3 is a chair in which a Parkinson's disease patient sits, for example during imaging, and comprises an upper part 3a of the chair having a seat, a backrest, and armrests, a lower part 3b of the chair having four legs, and a shaft part 3c positioned between the upper part 3a and the lower part 3b of the chair. The shaft portion 3c has a rotation mechanism that allows the upper part of the chair 3a to rotate in a horizontal plane around the shaft, and a locking mechanism that allows the upper part of the chair 3a to be stopped at a predetermined rotation position (rotation angle).

[0023] The reason for using the swivel chair with a locking function (3) is that when photographing Parkinson's disease patients from various angles, if the patient has difficulty standing up (because Parkinson's disease involves postural and gait disturbances), the chair itself can be rotated, allowing a nurse or caregiver to take the photographs alone. If the chair itself moves freely, the risk of tipping over increases. A locking mechanism allows for safe and easy shooting.

[0024] Finally, I will explain the information processing device 1. The information processing device 1 is comprised of a main unit 40 positioned on the underside of the top plate 4a, and a keyboard (one of the input units 16) and a display device (one of the output units 17) placed on the top surface of the top plate 4a. Information processing device 1 is an information processing terminal operated by the evaluator (user) who conducts the evaluation. Although the information processing device 1 is configured like a personal computer as shown in Figures 1 to 3, it could also be a tablet or smartphone, for example, including the imaging unit 2. The information processing device 1 will have dedicated software (program) installed. It will also be equipped with a unique evaluation algorithm for estimating UPDRS. Furthermore, it will be equipped with a function to guide users on how to take the best images.

[0025] The aforementioned evaluators (users) include not only specialists who specialize in evaluating the motor symptoms of Parkinson's disease patients, but also non-specialists who have not received training in evaluation (such as doctors who are not specialists in hospitals or clinics, nurses who have free time, or caregivers in nursing homes). Furthermore, even with such evaluators, adopting this system will enable the acquisition of consistent evaluation results.

[0026] The hardware configuration of the information processing device 1 will be described below. Figure 4 is a block diagram showing an example of the hardware configuration of the information processing device in the estimation system shown in Figure 1.

[0027] The information processing device 1 comprises a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a bus 14, an input / output interface 15, an input unit 16, an output unit 17, a storage unit 18, a communication unit 19, and a drive 20.

[0028] The CPU 11 executes various processes according to the program recorded in the ROM 12 or the program loaded from the storage unit 18 into the RAM 13. RAM13 also stores data and other information necessary for the CPU11 to perform various processes.

[0029] The CPU 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output interface 15 is also connected to this bus 14. An input / output interface 15 is connected to an input unit 16, an output unit 17, a storage unit 18, a communication unit 19, and a drive 20.

[0030] The input unit 16 is configured, for example, with a keyboard, and is used to input various types of information. The output unit 17 consists of a display such as an LCD (for example, the display device described above) and a speaker, and outputs various information as images and sounds. The memory unit 18 is composed of DRAM (Dynamic Random Access Memory) and stores various types of data. The communication unit 19 communicates with other devices (for example, the imaging unit main body 2a and a server (not shown) which performs automatic image analysis on the server) via a network including the Internet.

[0031] The drive 20 is appropriately equipped with removable media 30, which may consist of a magnetic disk, optical disk, magneto-optical disk, or semiconductor memory. Programs read from the removable media 30 by the drive 20 are installed in the storage unit 18 as needed. Furthermore, the removable media 30 can store various types of data stored in the storage unit 18, just as the storage unit 18 does.

[0032] It should be noted that the servers not shown in the illustrations can have a hardware configuration that is basically the same as that shown in Figure 4. Therefore, a description of the hardware configuration of the servers not shown in the illustrations will be omitted.

[0033] Through the collaboration of various hardware and software components of the information processing device 1 shown in Figure 3, various processes can be executed. As a result, evaluations and other processes described later become possible.

[0034] The functional configuration of the information processing device 1 will be explained with reference to Figure 5. Figure 5 is a functional block diagram showing an example of the functional configuration of the information processing device in Figure 4, which is part of the estimation system in Figure 1.

[0035] As shown in Figure 5, the CPU 11 of the information processing device 1 functions as an estimation unit 51 and an extraction unit 52. Furthermore, one area of ​​the storage unit 18 of the information processing device 1 is provided with an image database 71, an analysis database 72, and a report database 73.

[0036] The estimation unit 51 includes an image capture control unit 511, a display control unit 512, a recording control unit 513, an analysis unit 514, and a report generation unit 515.

[0037] The imaging control unit 511 controls the imaging unit 2. Specifically, the imaging control unit 511 controls the imaging unit body 2a. The shooting control unit 511 saves the captured image data to the shooting DB 71.

[0038] The display control unit 512 executes control to display navigation information indicating the imaging method for Parkinson's disease patient K on the output unit 17 (display device). The shooting method described above includes information regarding the shooting procedure, the shooting distance, and the angle of view. Furthermore, the aforementioned navigation information includes text to provide specific movement instructions to Parkinson's disease patient K, sample images showing the ideal field of view, and instructions to avoid occlusion and background assimilation of Parkinson's disease patient K's body parts. The above-described filming procedure is performed in the following order: right hand finger tapping, left foot tapping, left leg agility, left hand finger tapping, right foot tapping, right leg agility, right hand movement, left hand movement, right hand pronation and supination, left hand pronation and supination, standing up from a chair, posture, and walking. The finger tapping of the right hand, tapping of the left foot, etc., mentioned in the above-mentioned shooting procedure are motor symptom-related items included in Part 3 of the UPDRS.

[0039] The recording control unit 513 executes control to record the image captured by the shooting unit 2 into the storage unit 18 (storage device) based on the shooting method.

[0040] The analysis unit 514 analyzes the recorded images to estimate a score indicating the motor symptoms of Parkinson's disease patient K. Furthermore, the analysis unit 514 applies preprocessing to the recorded images and performs control to estimate the score for each motor symptom item. Furthermore, the analysis unit 514 can upload the recorded images to a server (not shown) and perform automatic analysis on that server. The preprocessing performed by the analysis unit 514 includes cropping, trimming, joint coordinate estimation, extraction of kinematic features, and input to a machine learning model.

[0041] The report generation unit 515 performs control to generate a report based on the estimated score. The report mentioned above includes radar charts, tables, and comparisons with past scores. The report generation unit 515 saves the generated report to the report DB 73. Report DB73 stores templates and past report data necessary for generating reports.

[0042] The function of the estimation unit 51 described above removes the barrier of expertise in evaluating the motor symptoms of Parkinson's disease patient K, enabling simple and stable evaluation even by non-experts.

[0043] The extraction unit 52 has the function of estimating the overall score of motor symptoms of Parkinson's disease patient K, and includes an image acquisition unit 521, a score extraction unit 522, and an overall score estimation unit 523.

[0044] The image acquisition unit 521 executes control to acquire images of Parkinson's disease patient K performing some of the pre-set motor tasks. Note that some of the pre-configured tasks mentioned above are fewer in number than the total number of motor tasks, and are tasks selected by Recursive Feature Elimination (RFE). The aforementioned motor tasks are the motor symptom-related items in Part 3 of the Unified Parkinson's Disease Assessment Scale (UPDRS). These motor symptom-related items include multiple motor tasks such as tapping with the fingers of the right hand, tapping with the fingers of the left hand, tapping with the right foot, tapping with the left foot, agility of the right leg, agility of the left leg, movement of the right hand, movement of the left hand, pronation and supination of the right hand, pronation and supination of the left hand, walking, standing up from a chair, and posture.

[0045] Some pre-configured tasks may include, for example, five or ten tasks (these numbers are examples and are not limited to five or ten). The five tasks mentioned above are tapping with the right leg, agility of the right leg, movement of the left hand, pronation and supination of the right hand, and walking. The ten tasks mentioned above are: left hand finger tapping, right foot tapping, right leg agility, left leg agility, right hand movement, left hand movement, right hand pronation / supination, walking, standing up from a chair, and posture.

[0046] The score extraction unit 522 performs control to extract individual scores corresponding to each of a set of pre-configured tasks from the image.

[0047] The overall score estimation unit 523 performs control to estimate an overall score (the definition of which will be described later) based on the individual scores of some of the pre-set tasks among the multiple motor tasks performed by Parkinson's disease patient K. Furthermore, when there are five tasks as described above, the coefficient of determination R^2 of the overall score estimation by the overall score estimation unit 523 is 0.95 or higher. On the other hand, when there are 10 tasks as described above, the coefficient of determination R^2 of the overall score estimation by the overall score estimation unit 523 is 0.99 or higher. The tasks will be described later with reference to Figure 8, but there are 18 surveys to investigate motor symptoms (18 motor symptom-related items).

[0048] The extraction unit 52, as described above, reduces the time burden required to evaluate the motor symptoms of Parkinson's disease patient K, and enables frequent follow-up by estimating a comprehensive score with high accuracy from a small number of tasks.

[0049] Figure 6 shows an example of how navigation information is displayed. The navigation screen shown in Figure 6 displays the task to be performed (e.g., "Task 1: Finger Tapping: Right"), instructions on how to perform the task, specific instructions for the patient, and sample images showing the ideal field of view. At the bottom of the screen are a progress bar showing the progress and navigation buttons such as "Skip" and "Next."

[0050] In Figure 6, the instructions for performing the task described above are: "Take a side view of the right hand tapping," "Ensure the hand is open when starting," and "Give the following instructions to the person being filmed." Furthermore, the specific instructions given to the patient, as mentioned above, are: "When I give the signal, tap your right index finger and thumb as loudly and quickly as possible in a continuous sequence."

[0051] Next, we will describe the operation of the estimation system 100. First, the imaging unit 2 is controlled by the imaging control unit 511 of the information processing device 1, and then the navigation information is displayed on the output unit 17 (display device) by the display control unit 512. Navigation information includes details of the task to be photographed, the method of photography (distance, angle, etc.), and specific instructions for the patient.

[0052] The shooting procedure is performed in the following order from (1) to (13): (1) tapping of the fingers of the right hand, (2) tapping of the left foot, (3) agility of the left leg, (4) tapping of the fingers of the left hand, (5) tapping of the right foot, (6) agility of the right leg, (7) movement of the right hand, (8) movement of the left hand, (9) pronation and supination of the right hand, (10) pronation and supination of the left hand, (11) standing up from a chair, (12) posture, and (13) walking.

[0053] During each task, Parkinson's disease patient K sits in a swivel chair 3 with a locking function and performs the exercises according to the instructions. For example, in the case of finger tapping, as mentioned above, the instruction "When given the signal, tap your right index finger and thumb as loudly and quickly as possible in a continuous sequence" will be displayed. The photographer will then convey this instruction to the patient and take the picture with the appropriate angle of view.

[0054] The captured images are recorded in the storage unit 18 by the recording control unit 513, and then analyzed by the analysis unit 514. The analysis involves preprocessing steps such as cropping, trimming, joint coordinate estimation, kinematic feature extraction, and input into a machine learning model, and a motor symptom score is estimated for each item of each motor task.

[0055] Figure 7 shows an example of how the overall score estimation results are displayed. As shown in Figure 7, the screen displays patient information, the current score, a radar chart showing scores by category, a detailed UPDRS score table, and more. The report generation unit 515 generates a report containing this information and, if necessary, displays a comparison with past scores.

[0056] The above is a description of a first embodiment, relating to an estimation system 100 for estimating the motor symptoms of a Parkinson's disease patient K. Next, we will describe a second embodiment, which includes an algorithm for estimating a comprehensive score of motor symptoms in Parkinson's disease from a small amount of tasks. In this specification, "overall score" refers to the sum of the scores from multiple motor tasks (specifically, the 13 tasks that can be filmed) in UPDRS Part 3.

[0057] Figure 8 shows a list of motor symptom assessments included in Part 3 of the UPDRS. Of the list of motor symptom assessment items (18 motor symptom-related items) included in Part 3 of the UPDRS, the following 13 items are tasks that can be assessed using video: (1) right hand finger tapping, (2) left hand finger tapping, (3) right foot tapping, (4) left foot tapping, (5) right leg agility, (6) left leg agility, (7) right hand movement, (8) left hand movement, (9) right hand pronation / supination, (10) left hand pronation / supination, (11) walking, (12) standing up from a chair, and (13) posture.

[0058] In this embodiment, a machine learning feature selection method called Recursive Feature Elimination (RFE) is used to automatically select the item that contributes most to predicting the overall score from among the 13 items mentioned above. RFE is a technique that sequentially removes features with low contribution while maintaining the accuracy of the predictive model.

[0059] Figure 9 shows a comparison of the overall score estimation results from a small number of tasks. Specifically, Figure 9 shows the results of estimating the overall score for all 13 items from five tasks selected by RFE (right leg tapping, right leg agility, left hand movement, right hand pronation / supination, walking) and ten tasks (left hand finger tapping, right foot tapping, right leg agility, left leg agility, right hand movement, left hand movement, right hand pronation / supination, walking, standing up from a chair, posture).

[0060] The coefficient of determination (R^2) was 0.95 when using 5 tasks, and 0.99 when using 10 tasks, demonstrating that the overall score can be estimated with high accuracy in both cases. This means that it becomes possible to assess overall motor symptoms using only a few tasks, without having to perform all of them.

[0061] Next, the operation of the algorithm according to the second embodiment will be described. First, the image acquisition unit 521 of the extraction unit 52 acquires images of some Parkinson's disease patients performing pre-set tasks (for example, five tasks). Next, the score extraction unit 522 extracts individual scores corresponding to each task from the images. This extraction of individual scores is performed using the same preprocessing and machine learning model as the analysis unit 514.

[0062] The extracted individual scores are input into a linear regression model to estimate the overall score. The linear regression model is represented by the following equation. Total score = B0 + B1 × X1 + B2 × X2 + ... + B n ×X n Here, B0 is the intercept, B1~B n These are the coefficients for each task, X1~X n This represents the individual score for each task. These coefficients are learned using pre-collected training data (data with known scores for all 13 items).

[0063] The estimated overall score is incorporated into the report and displayed by the report generation unit 515. Patients and healthcare professionals can use this information to make decisions about treatment strategies and evaluate their effectiveness.

[0064] As explained with reference to Figures 1 to 9, the first embodiment can be used as a motor symptom evaluation system for Parkinson's disease patient K in various settings such as medical institutions, nursing homes, and home healthcare. Furthermore, in the second embodiment, the combination of the estimation system and the overall score estimation algorithm enables frequent monitoring of the patient's progress, allowing for a more detailed understanding of changes in the disease state. This is expected to contribute to the optimization of treatment strategies and the improvement of the patient's quality of life.

[0065] Although one embodiment of the present invention has been described above, the present invention is not limited to the embodiments described above, and any modifications, improvements, etc. that can achieve the objectives of the present invention are considered to be included in the present invention.

[0066] Furthermore, the system configurations shown in Figures 1 to 3, and the hardware configuration of the information processing device 1 shown in Figure 4, are merely illustrative examples for achieving the objectives of the present invention and are not particularly limited.

[0067] Furthermore, the functional block diagram shown in Figure 5 is merely illustrative and not particularly limiting. In other words, it is sufficient that the estimation system in Figures 1 to 3 has the functionality to execute the various processes described above as a whole, and the functional blocks and databases used to realize this functionality are not particularly limited to the example in Figure 5.

[0068] Furthermore, the location of the functional blocks and database is not limited to Figure 5, but can be any location.

[0069] For example, the series of processes described above can be executed by hardware or by software. Furthermore, a single functional block may consist of hardware alone, software alone, or a combination of both.

[0070] When a series of processes are executed by software, the programs that make up that software are installed on a computer or other device from a network or storage medium. The computer may be a computer that is built into dedicated hardware. Furthermore, a computer can be any computer capable of performing various functions by installing various programs, such as a server, a general-purpose smartphone, or a personal computer.

[0071] Such recording media containing programs may consist not only of removable media (not shown) distributed separately from the main unit of the device to provide the program to the user, but also of recording media provided to the user in a state where they are pre-installed in the main unit of the device.

[0072] In this specification, the step of describing a program to be recorded on a recording medium includes not only processes that are performed chronologically in that order, but also processes that are not necessarily performed chronologically, but are executed in parallel or individually. Furthermore, in this specification, the term "system" refers to an overall system composed of multiple devices, means, etc.

[0073] In summary, the estimation system and information processing device to which the present invention applies only need to have the following configuration, and can take various forms. In other words, the estimation system to which the present invention is applied (for example, the estimation system 100 in Figure 1) is: An estimation system for estimating the motor symptoms of Parkinson's disease patients (for example, Parkinson's disease patient K in Figure 5), The aforementioned imaging unit (for example, imaging unit 2 in Figure 1) captures the movements of the Parkinson's disease patient, Information processing device (information processing device 1, etc., in Figures 1 to 5), In an estimation system having, The aforementioned information processing device is The imaging control unit that controls the imaging unit (e.g., the imaging control unit 511 in Figure 5), A display control unit (for example, the display control unit 512 in Figure 5) executes control to display navigation information indicating the method of photographing the Parkinson's disease patient on a predetermined display device (for example, the output unit 17 in Figure 1), A recording control unit (e.g., recording control unit 513 in Figure 5) executes control to record the image captured by the shooting unit based on the shooting method into a predetermined storage device (e.g., shooting DB71 in Figure 5), An analysis unit (for example, the analysis unit 514 in Figure 5) analyzes the recorded images to estimate a score indicating the motor symptoms of the Parkinson's disease patient, Having that will suffice.

[0074] In this way, we can remove the barriers of expertise in evaluating the motor symptoms of Parkinson's disease patients and provide a system that enables simple and stable evaluation even by non-specialists.

[0075] Furthermore, the imaging unit and the information processing device are mounted on a table with casters (such as caster 4d in Figure 1) (for example, table 4 in Figure 1). In this way, the caster-equipped table ensures the mobility of the estimation system, allowing for flexible use in various environments such as patient rooms, examination rooms, and rehabilitation rooms.

[0076] Furthermore, the system includes a swivel chair with a locking mechanism for the Parkinson's disease patient to sit in during imaging (for example, the swivel chair with a locking mechanism shown in Figure 1, model 3). Thus, the swivel chair with a locking function allows for safe and easy imaging from various angles, even for patients who have difficulty standing up, reducing the risk of falls. Furthermore, the operator does not need to move the patient, allowing for efficient imaging even by a single person.

[0077] Furthermore, the aforementioned shooting method includes information regarding the shooting procedure, the shooting distance, and the angle of view (see the sample image shown in Figure 6). In this way, standardizing the shooting conditions makes it possible to obtain consistent, high-quality image data, regardless of the photographer's experience or skill. This leads to improved analysis accuracy and increased reliability of evaluation results.

[0078] Furthermore, the navigation information includes text (see, for example, Figure 6) that provides specific instructions for movement to the Parkinson's disease patient. Thus, standardized instructions ensure clear and consistent explanations to patients, improving the reproducibility of task execution. This reduces variability in evaluations and increases the reliability of comparisons over time.

[0079] Furthermore, the navigation information includes a sample image showing the ideal field of view (see, for example, Figure 6). Thus, visual guidance from sample images allows photographers to intuitively understand the appropriate field of view, reducing training costs and enabling rapid mastery. Furthermore, it reduces analysis errors caused by variations in field of view.

[0080] Furthermore, the navigation information includes instructions to avoid occlusion and background assimilation of body parts of the Parkinson's disease patient. By avoiding occlusion and background assimilation in this way, the accuracy of image recognition is improved, and errors due to feature extraction failures are reduced. This allows for more accurate score estimation.

[0081] Furthermore, the aforementioned shooting procedure is performed in the following order: right hand finger tapping (see, for example, Figure 6), left foot tapping, left leg agility, left hand finger tapping, right foot tapping, right leg agility, right hand movement, left hand movement, right hand pronation and supination, left hand pronation and supination, standing up from a chair, posture, and walking. This efficient imaging sequence minimizes changes in the patient's posture and movement, shortening the imaging time. It also reduces patient fatigue and allows for more natural-looking motion assessment.

[0082] Furthermore, the analysis unit applies preprocessing (for example, the aforementioned cropping, trimming, joint coordinate estimation, kinematic feature extraction, and input to a machine learning model) to the recorded images and estimates the score for each motor symptom item (see, for example, Figure 9). In this way, image preprocessing reduces the influence of environmental conditions (such as lighting) and standardizes image quality, thereby improving analysis accuracy and reproducibility.

[0083] Furthermore, the analysis unit uploads the recorded images to a server (such as a server not shown in the illustration above) and performs automatic analysis on that server. In this way, by utilizing the computing resources of a cloud server, for example, advanced analysis algorithms can be executed without being limited by the processing power of the local environment. Furthermore, updating the model on the server makes it easier to improve the accuracy of the entire system.

[0084] Furthermore, the preprocessing performed by the analysis unit includes cropping, trimming, joint coordinate estimation, extraction of kinematic features, and input to a machine learning model. In this way, multi-stage preprocessing allows for the efficient extraction of features directly related to the evaluation of motor symptoms from images, enabling highly accurate score estimation with reduced effects of noise and variability.

[0085] Furthermore, the system includes a report generation unit (for example, the report generation unit 515 in Figure 5) that generates a report based on the estimated score. The report includes radar charts (see, for example, Figure 7), tables, and comparisons with past scores. In this way, the visually easy-to-understand report format allows healthcare professionals and patients themselves to easily grasp the overall picture of symptoms and how they change over time. This enables more appropriate evaluation of treatment effectiveness and determination of treatment plans.

[0086] The information processing device to which the present invention applies only needs to have the following configuration, and various embodiments can be taken. That is, the information processing device to which the present invention applies (information processing device 1, etc., shown in Figures 1 to 5) is: An information processing device for estimating the overall motor symptom score of a Parkinson's disease patient (for example, Parkinson's disease patient K in Figure 5) (for example, the sum of the scores of multiple motor tasks (specifically, 13 tasks that can be recorded) in UPDRS Part 3), An image acquisition unit (for example, the image acquisition unit 521 in Figure 5) acquires images of the Parkinson's disease patient performing some of the pre-set motor tasks among several motor tasks, A score extraction unit (for example, the score extraction unit 522 in Figure 5) extracts individual scores corresponding to each of the aforementioned pre-set tasks from the aforementioned image, A total score estimation unit (for example, the total score estimation unit 523 in Figure 5, etc.) estimates the total score based on the individual scores of some of the pre-set tasks, Equipped with, Some of the aforementioned pre-configured tasks are A number of tasks less than the total number of the aforementioned motor tasks, and selected by Recursive Feature Elimination (RFE). That should suffice.

[0087] In this way, the time burden required to evaluate the motor symptoms of Parkinson's disease patients can be reduced, and frequent follow-up can be realized by estimating the overall score with high accuracy from a small number of tasks.

[0088] Furthermore, the aforementioned motor tasks are the motor symptom-related items of Part 3 of the Unified Parkinson's Disease Assessment Scale (UPDRS) (see, for example, Figure 8). In this way, by linking with standardized evaluation scales, it is possible to achieve more efficient evaluations while maintaining consistency with conventional evaluations.

[0089] Furthermore, the aforementioned multiple motor tasks include tapping with the fingers of the right hand (see, for example, Figure 6), tapping with the fingers of the left hand, tapping with the right foot, tapping with the left foot, agility of the right leg, agility of the left leg, movement of the right hand, movement of the left hand, pronation and supination of the right hand, pronation and supination of the left hand, walking, standing up from a chair, and posture. In this way, by clearly defining the entire range of motion tasks that can be filmed, it becomes possible to establish an estimation basis from a small number of tasks.

[0090] Furthermore, some of the pre-configured tasks include five tasks (see, for example, Figure 9), The five tasks mentioned above are tapping with the right leg, agility of the right leg, movement of the left hand, pronation and supination of the right hand, and walking. By reducing the number of tasks to five, evaluation time can be significantly shortened while maintaining sufficient prediction accuracy.

[0091] Furthermore, the coefficient of determination R^2 of the overall score estimation by the overall score estimation unit is 0.95 or higher (see, for example, Figure 9). In this way, by guaranteeing high prediction accuracy (over 95%), we achieve clinically reliable estimation.

[0092] Furthermore, some of the pre-configured tasks include 10 tasks (see, for example, Figure 9), The ten tasks mentioned above are: left hand finger tapping, right foot tapping, right leg agility, left leg agility, right hand movement, left hand movement, right hand pronation and supination, walking, standing up from a chair, and posture. By setting the number of tasks to 10 in this way, it becomes possible to estimate the overall score with higher accuracy.

[0093] Furthermore, the coefficient of determination R^2 of the overall score estimation by the overall score estimation unit is 0.99 or higher (see, for example, Figure 9). In this way, by guaranteeing extremely high prediction accuracy (over 99%), it is possible to obtain evaluation results equivalent to those obtained when performing almost all tasks. [Explanation of Symbols]

[0094] 1... Information processing unit, 2... Imaging unit, 3... Swivel chair with locking function, 4... Table, 11... CPU, 12... ROM, 13... RAM, 14... Bus, 15... Input / Output interface, 16... Input unit, 17... Output unit, 18... Storage unit, 19... Communication unit, 20... Drive, 30... Removable media, 51... Estimation unit, 52... Extraction unit, 71... Imaging DB, 72... Analysis DB, 73... Report DB, 100... Estimation system, 511... Imaging control unit, 512... Display control unit, 513... Recording control unit, 514... Analysis unit, 515... Report generation unit, 521... Image acquisition unit, 522... Score extraction unit, 523... Overall score estimation unit

Claims

1. An information processing device for estimating the overall score of motor symptoms in a Parkinson's disease patient, An image acquisition unit that acquires images of a Parkinson's disease patient performing some of several pre-set motor tasks, A score extraction unit extracts individual scores corresponding to each of the aforementioned pre-set tasks from the aforementioned image, A total score estimation unit that estimates the total score based on the individual scores of some of the pre-set tasks, Equipped with, Some of the aforementioned pre-configured tasks are The number of tasks is less than the total number of the aforementioned multiple motor tasks, and the tasks are selected by Recursive Feature Elimination (RFE). Information processing device.

2. The aforementioned motor tasks are motor symptom-related items in Part 3 of the Unified Parkinson's Disease Assessment Scale (UPDRS). The information processing apparatus according to claim 1.

3. The aforementioned multiple motor tasks include right hand finger tapping, left hand finger tapping, right foot tapping, left foot tapping, right leg agility, left leg agility, right hand movement, left hand movement, right hand pronation / supination, left hand pronation / supination, walking, standing up from a chair, and posture. The information processing apparatus according to claim 1.

4. The aforementioned pre-configured tasks include five tasks, The five tasks mentioned above are tapping with the right leg, agility with the right leg, movement of the left hand, pronation and supination of the right hand, and walking. The information processing apparatus according to claim 1.

5. The coefficient of determination R^2 of the overall score estimation by the overall score estimation unit is 0.95 or greater. The information processing apparatus according to claim 4.

6. The aforementioned pre-configured tasks include 10 tasks, The ten tasks mentioned above are: left hand finger tapping, right foot tapping, right leg agility, left leg agility, right hand movement, left hand movement, right hand pronation / supination, walking, standing up from a chair, and posture. The information processing apparatus according to claim 1.

7. The coefficient of determination R^2 of the overall score estimation by the overall score estimation unit is 0.99 or greater. The information processing apparatus according to claim 6.

8. An information processing method performed by an information processing device that estimates the overall score of motor symptoms in patients with Parkinson's disease, The image acquisition step involves acquiring images of a Parkinson's disease patient performing some of several pre-set motor tasks, and A score extraction step of extracting individual scores corresponding to each of the aforementioned pre-set tasks from the aforementioned image, A total score estimation step in which the total score is estimated based on the individual scores of some of the pre-set tasks, Includes, Some of the aforementioned pre-configured tasks are The number of tasks is less than the total number of the aforementioned multiple motor tasks, and the tasks are selected by Recursive Feature Elimination (RFE). Information processing methods.

9. A computer that controls an information processing device that estimates the overall score of motor symptoms in Parkinson's disease patients, The image acquisition step involves acquiring images of a Parkinson's disease patient performing some of several pre-set motor tasks, and A score extraction step of extracting individual scores corresponding to each of the aforementioned pre-set tasks from the aforementioned image, A total score estimation step in which the total score is estimated based on the individual scores of some of the pre-set tasks, The control process including this is executed, Some of the aforementioned pre-configured tasks are The number of tasks is less than the total number of the aforementioned multiple motor tasks, and the tasks are selected by Recursive Feature Elimination (RFE). program.

Images