Rehabilitation training system for MCI patients in different scenarios based on deep learning

Abstract

The present invention discloses a rehabilitation training system for MCI patients in different scenarios based on deep learning, which belongs to the field of rehabilitation training technology and includes a data collection module, a data processing module, a model building module, a prediction analysis module and a rehabilitation training module. The present invention solves the problem that the existing rehabilitation training for MCI patients cannot perform rehabilitation training for MCI patients in different scenarios, and cannot adjust and optimize the rehabilitation training process in real time, resulting in poor rehabilitation training effect for MCI patients. The present invention performs rehabilitation training for MCI patients in different scenarios according to rehabilitation training schemes for different scenarios, adjusts and optimizes rehabilitation training schemes based on different scenarios for MCI patients based on feedback information from MCI patients, can perform rehabilitation training for MCI patients in different scenarios, can adjust and optimize the rehabilitation training process in real time according to the rehabilitation training situation of MCI patients, and can improve the rehabilitation training effect of MCI patients.

Description

Technical Field

[0001] The present invention relates to the field of rehabilitation training technology, and specifically to a rehabilitation training system for MCI patients in different scenarios based on deep learning. Background Art

[0002] MCI, or mild cognitive impairment (MCI), is an intermediate state between normal aging and dementia, and is a cognitive impairment syndrome. Compared with normal elderly people of the same age and education level, patients have mild cognitive impairment, but their daily abilities are not significantly affected. The core symptom of mild cognitive impairment is the decline of cognitive function. Depending on the cause or the location of brain damage, it can affect one or more of memory, executive function, language, application, and visual-spatial structural skills, leading to corresponding clinical symptoms.

[0003] The Chinese patent with publication number CN104248441A discloses a cognitive rehabilitation training system, which can train patients' reaction behavior, logical thinking ability, plane operation ability, search ability, detail memory, concentration, calculation ability and visual motor coordination ability, so that patients can maintain their existing cognitive functions and continue functional decline, enhance their self-care level and improve their quality of life; however, the patent has the following defects:

[0004] The existing rehabilitation training for MCI patients cannot provide rehabilitation training for MCI patients in different scenarios, and cannot adjust and optimize the rehabilitation training process in real time according to the rehabilitation training situation of MCI patients, resulting in poor rehabilitation training effects for MCI patients. Summary of the invention

[0005] The purpose of the present invention is to provide a rehabilitation training system for MCI patients in different scenarios based on deep learning, which can perform rehabilitation training for MCI patients in different scenarios, and can adjust and optimize the rehabilitation training process in real time according to the rehabilitation training situation of the MCI patients, thereby improving the rehabilitation training effect of MCI patients and solving the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] The deep learning-based rehabilitation training system for MCI patients uses different scenarios, including:

[0008] A data collection module is used to collect attribute information, medical history information, imaging information and genetic information of MCI patients, and determine multimodal real-time data based on MCI patients;

[0009] A data processing module is used to retrieve, convert and fuse multimodal real-time data based on MCI patients to determine multimodal feature data based on MCI patients;

[0010] A model building module is used to train the selected model architecture based on different scene rehabilitation training according to multimodal historical data to determine the deep learning model;

[0011] A prediction and analysis module, used to predict the multimodal feature data based on MCI patients based on the deep learning model, determine the model prediction results based on MCI patients, evaluate and analyze the model prediction results based on MCI patients, and determine the evaluation and analysis results based on MCI patients;

[0012] The rehabilitation training module is used to formulate rehabilitation training plans for MCI patients in different scenarios, and to conduct rehabilitation training for MCI patients in different scenarios according to the rehabilitation training plans for different scenarios.

[0013] Preferably, the data collection module includes:

[0014] An attribute collection unit, used for collecting attribute information of MCI patients in real time, wherein the attribute information of MCI patients includes name, gender and age;

[0015] A medical history collection unit, used to collect medical history information of MCI patients in real time, wherein the medical history information of MCI patients includes the occurrence and development of the disease, as well as the examination, diagnosis and treatment of MCI patients;

[0016] An image acquisition unit, used for real-time acquisition of image information of MCI patients;

[0017] A gene collection unit, used to collect genetic information of MCI patients in real time;

[0018] Among them, based on the attribute information, medical history information, imaging information and genetic information of MCI patients collected in real time, multimodal real-time data based on MCI patients is determined.

[0019] Preferably, the data processing module includes:

[0020] A data retrieval unit, used for retrieving multimodal real-time data based on MCI patients;

[0021] Obtain multimodal real-time data based on MCI patients;

[0022] Based on the sequential retrieval method, multimodal real-time data based on MCI patients are retrieved;

[0023] Remove duplicate data, missing data and outliers from multimodal real-time data of MCI patients that are not valuable for rehabilitation training of MCI patients in different scenarios;

[0024] Identify multimodal real-time data that is valuable for rehabilitation training in different scenarios for MCI patients;

[0025] A data conversion unit, used for converting the retrieved multimodal real-time data;

[0026] Obtain multimodal real-time data that is valuable for rehabilitation training of MCI patients in different scenarios after retrieval;

[0027] After retrieval, the multimodal real-time data that is valuable for rehabilitation training of MCI patients in different scenarios is converted to normalize or standardize the multimodal real-time data;

[0028] Identify multimodal real-time data with a unified format and units of measurement;

[0029] A data fusion unit, used for fusing the converted multi-modal real-time data;

[0030] Get multi-modal real-time data with unified formats and units of measurement;

[0031] And integrate multi-modal real-time data;

[0032] Determine the multimodal feature data based on MCI patients.

[0033] Preferably, the model building module includes:

[0034] A data storage unit, used for storing multimodal historical data of MCI patients;

[0035] Among them, the stored multimodal historical data of MCI patients are extracted and used as the model training set;

[0036] A model selection unit, used to select a model architecture based on rehabilitation training in different scenarios;

[0037] Based on the rehabilitation training needs of MCI patients in different scenarios;

[0038] According to multimodal historical data, select the model architecture based on different scenarios of rehabilitation training;

[0039] Model training unit, used to train the selected model architecture and build a deep learning model;

[0040] Obtain the selected model architecture and multimodal historical data based on different scenario rehabilitation training;

[0041] Based on the rehabilitation training needs of MCI patients in different scenarios, the selected model architecture based on rehabilitation training in different scenarios is trained according to multimodal historical data to determine the deep learning model.

[0042] Preferably, training the selected model architecture to construct a deep learning model includes:

[0043] Based on multimodal historical data, the selected model architecture based on rehabilitation training in different scenarios is trained to determine the deep learning model;

[0044] Among them, the deep learning model is verified;

[0045] Obtain a model validation set, validate the deep learning model based on the model validation set, and determine the model validation result;

[0046] Obtain model validation results, conduct in-depth mining and analysis on the model validation results, and determine the model adjustment and optimization plan;

[0047] The deep learning model is adjusted and optimized based on the model adjustment and optimization plan, and the best deep learning model is determined through iterative training and optimization process.

[0048] Preferably, the prediction analysis module includes:

[0049] A model prediction unit, used to predict multimodal feature data based on MCI patients;

[0050] Get the best deep learning model;

[0051] Obtain multimodal feature data based on MCI patients;

[0052] Input multimodal feature data from MCI patients into the best deep learning model;

[0053] Predict multimodal feature data of MCI patients based on deep learning models;

[0054] Determine the model's predictions based on MCI patients;

[0055] An evaluation and analysis unit, used for evaluating and analyzing the prediction results of the model based on MCI patients;

[0056] Obtain model prediction results based on MCI patients;

[0057] Evaluate and analyze the prediction results of the model based on MCI patients;

[0058] Results were determined based on the evaluation of patients with MCI.

[0059] Preferably, the rehabilitation training module includes:

[0060] Rehabilitation Development Unit, used to develop rehabilitation training plans for MCI patients in different scenarios;

[0061] Obtain assessment and analysis results based on MCI patients;

[0062] Conduct mining and analysis on the assessment and analysis results based on MCI patients;

[0063] Determine rehabilitation training programs based on different scenarios for MCI patients;

[0064] Rehabilitation training unit, used to provide rehabilitation training for MCI patients in different scenarios;

[0065] Obtain rehabilitation training plans based on different scenarios for MCI patients;

[0066] MCI patients are given rehabilitation training in different scenarios according to the rehabilitation training plan based on different scenarios of MCI patients.

[0067] Preferably, the rehabilitation training module further includes:

[0068] Patient feedback unit, used to record the feedback information of MCI patients on rehabilitation training in different scenarios;

[0069] When MCI patients are undergoing rehabilitation training in different scenarios, real-time feedback information from MCI patients on rehabilitation training in different scenarios is obtained;

[0070] Record the real-time feedback information of MCI patients on rehabilitation training in different scenarios;

[0071] A program adjustment unit is used to adjust and optimize the rehabilitation training programs based on different scenarios of MCI patients;

[0072] Obtain the feedback information of MCI patients on rehabilitation training in different scenarios;

[0073] Adjust and optimize rehabilitation training programs based on different scenarios for MCI patients based on feedback from MCI patients;

[0074] Determine the best rehabilitation training program based on different scenarios for MCI patients;

[0075] MCI patients are given the best rehabilitation training in different scenarios according to the best rehabilitation training plan based on different scenarios of MCI patients.

[0076] Preferably, the scheme adjustment unit includes:

[0077] A program temporary storage subunit is used to temporarily store the initial rehabilitation training program, and select the scene rehabilitation training program currently used by the MCI patient in the initial rehabilitation training program, and at the same time, use the scene rehabilitation training program as a prototype program;

[0078] The result prediction subunit is used to call the model of MCI patients, and evaluate and analyze the prototype solution based on the model of MCI patients to obtain prediction information of the prototype solution;

[0079] Contrast selection subunit, used for:

[0080] Read the feedback information, compare the feedback information with the predicted information, and set a comparison parameter range for the predicted information;

[0081] If the feedback information is within the control parameter range of the prediction information, the feedback information is imported into the normal feedback record set;

[0082] If the feedback information does not fall within the control parameter range of the prediction information, the feedback information is imported into the abnormal feedback record set;

[0083] Adjust the processing subunit to:

[0084] Retrieving feedback information in the abnormal feedback record set, and processing the feedback information in the abnormal feedback record set to obtain adjustment information;

[0085] Generate a replacement plan based on the adjustment information;

[0086] A replacement processing subunit is used to call the prototype scheme currently used by the MCI patient, and replace the prototype scheme based on the replacement scheme, and use the replaced prototype scheme as the scenario rehabilitation training scheme currently used by the MCI patient;

[0087] The information cache subunit is used to:

[0088] Obtaining behavior replacement records when replacing the prototype solution, and accumulating the records to obtain a first accumulation cycle;

[0089] Accumulate the normal feedback record set to obtain a second accumulation cycle;

[0090] A preset record accumulation period is obtained, and when the first accumulation period or the second accumulation period reaches the behavior replacement record or the normal feedback record set corresponding to the preset record accumulation period, the behavior replacement record or the normal feedback record set is removed.

[0091] Preferably, the rehabilitation training module further includes:

[0092] A data acquisition unit is used to obtain theoretical rehabilitation data corresponding to the rehabilitation training program for each scenario after formulating rehabilitation training programs for different scenarios for MCI patients, and at the same time, obtain actual rehabilitation data corresponding to the rehabilitation training programs for different scenarios;

[0093] A first calculation unit is used to calculate the comprehensive rehabilitation efficiency of rehabilitation training for MCI patients according to theoretical rehabilitation data and actual rehabilitation data corresponding to rehabilitation training programs in different scenarios;

[0094]

[0095] Where η represents the comprehensive rehabilitation efficiency of rehabilitation training for MCI patients; i represents the sequence number of the scene; n represents the total number of scenes; x i represents the actual rehabilitation data corresponding to the rehabilitation training program of the i-th scenario; X i represents the theoretical rehabilitation data corresponding to the rehabilitation training program of the i-th scenario, and x i ≤X i ; δ i represents the influence weight of the rehabilitation training program on the rehabilitation data in the i-th scenario; z represents the total actual rehabilitation data of MCI patients after receiving the rehabilitation training programs of all scenarios; Z represents the total theoretical rehabilitation data of MCI patients after receiving the rehabilitation training programs of all scenarios, and z≤Z;

[0096] The second calculation unit is used to retrieve the satisfaction survey data corresponding to the rehabilitation training program of each scene, and at the same time, calculate the total satisfaction of the MCI patient with rehabilitation training according to the satisfaction survey data corresponding to the rehabilitation training program of each scene;

[0097]

[0098] Where M represents the overall satisfaction of rehabilitation training for MCI patients; m i represents the satisfaction survey data corresponding to the rehabilitation training program of the i-th scenario; φ i represents the impact weight of the rehabilitation training program in the i-th scenario on the total satisfaction;

[0099] Scoring unit for:

[0100] Conduct a first rehabilitation effect evaluation based on the comprehensive rehabilitation efficiency of rehabilitation training for MCI patients and determine the first score for rehabilitation training;

[0101] Conduct a second rehabilitation effect evaluation based on the overall satisfaction of rehabilitation training for MCI patients and determine the second rehabilitation training score;

[0102] The comprehensive score of the rehabilitation effect of MCI patients during rehabilitation training is calculated based on the first score and the second score;

[0103] F = f1*σ1+f2*σ2;

[0104] Among them, F represents the comprehensive score of rehabilitation effect of MCI patients during rehabilitation training; f1 represents the first score; σ1 represents the influence weight of the comprehensive rehabilitation efficiency of rehabilitation training on the comprehensive score; f2 represents the second score; σ2 represents the influence weight of the total satisfaction of MCI patients with rehabilitation training on the comprehensive score;

[0105] A report generating unit, configured to generate a comprehensive evaluation report for the MCI patient after the rehabilitation training according to the rehabilitation training programs of different scenarios based on the first score, the second score and the comprehensive score;

[0106] The transmission unit is used to transmit the comprehensive evaluation report to the monitoring terminal.

[0107] Compared with the prior art, the present invention has the following beneficial effects:

[0108] 1. The present invention collects attribute information, medical history information, imaging information and genetic information of MCI patients, determines multimodal real-time data based on MCI patients, retrieves, converts and fuses multimodal real-time data based on MCI patients, determines multimodal feature data based on MCI patients, trains a selected model architecture based on rehabilitation training in different scenarios according to the multimodal historical data, determines a deep learning model, predicts the multimodal feature data based on MCI patients based on the deep learning model, determines model prediction results based on MCI patients, evaluates and analyzes the model prediction results based on MCI patients, and determines evaluation and analysis results based on MCI patients.

[0109] 2. The present invention develops rehabilitation training plans for MCI patients in different scenarios through evaluation and analysis results, conducts rehabilitation training for MCI patients in different scenarios according to the rehabilitation training plans for different scenarios, and adjusts and optimizes the rehabilitation training plans for MCI patients in different scenarios based on feedback information from MCI patients. Rehabilitation training for MCI patients in different scenarios can be conducted, and the rehabilitation training process can be adjusted and optimized in real time according to the rehabilitation training situation of the MCI patients, thereby improving the rehabilitation training effect of MCI patients.

[0110] 3. Obtain all rehabilitation training plans through the plan temporary storage subunit, and select the scenario rehabilitation training plan currently used by the MCI patient as the prototype plan. After the special selection is completed, call the model of the MCI patient in the evaluation and analysis unit through the result prediction subunit, substitute it into the prototype plan, and obtain the prediction information of the currently used rehabilitation training plan through the logical budget, so as to obtain the prediction information as a reference for comparing the feedback information. Then, compare the feedback information with the prediction information through the comparison selection subunit, and set the comparison parameter range. If the feedback information is within the prediction information comparison parameter range, import the feedback information into the normal feedback record set. If the feedback information is not within the prediction information comparison parameter range, import the feedback information into the abnormal feedback record set, so as to compare the feedback information with the prediction information. Then, call the abnormal feedback record set through the adjustment processing subunit. The feedback information in the feedback record set is obtained, and the feedback information in the abnormal feedback record set is processed to obtain adjustment information, and the adjustment information is imported into the rehabilitation designated unit to obtain a replacement plan, so as to obtain a replacement plan, call the prototype plan currently used by the MCI patient, and replace the corresponding behavior in the scene rehabilitation training plan with the obtained replacement plan, set the replaced scene rehabilitation training plan as the scene rehabilitation training plan used by the current MCI patient, and complete the replacement optimization of the plan content. Finally, through the information cache sub-unit, the corresponding behavior of the replacement is imported into the cache record to obtain the behavior replacement record, and the record is accumulated at a certain time period, and the feedback information in the normal feedback record set is obtained. The record is accumulated at a certain time period. When the cycle is completed, all the accumulated records are deleted to cache a certain amount of records and process them to save memory usage.

[0111] 4. After formulating rehabilitation training plans for MCI patients in different scenarios, the theoretical rehabilitation data of the rehabilitation training plans corresponding to each scenario and the actual rehabilitation data corresponding to the rehabilitation training plans for different scenarios are obtained, so as to effectively calculate the comprehensive rehabilitation efficiency of rehabilitation training for MCI patients. By retrieving the satisfaction survey data corresponding to the rehabilitation training plans for each scenario, the total satisfaction of rehabilitation training for MCI patients is effectively calculated. Then, a first score is determined according to the comprehensive rehabilitation efficiency of rehabilitation training for MCI patients and a second score is determined according to the total satisfaction of rehabilitation training for MCI patients. Based on the first score and the second score, a comprehensive score of the rehabilitation effect of rehabilitation training for MCI patients can be effectively calculated, thereby realizing the evaluation and monitoring efficiency of rehabilitation training for MCI patients in different scenarios, and then effectively realizing the reading of the comprehensive evaluation report based on the monitoring terminal, thereby improving the monitoring efficiency and evaluation objectivity of the rehabilitation training plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0112] Figure 1This is a module schematic diagram of the rehabilitation training system for MCI patients in different scenarios according to the present invention. DETAILED DESCRIPTION

[0113] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0114] In order to solve the problem that the existing rehabilitation training for MCI patients cannot provide rehabilitation training for MCI patients in different scenarios, and cannot adjust and optimize the rehabilitation training process in real time according to the rehabilitation training situation of MCI patients, resulting in poor rehabilitation training effect for MCI patients, please refer to Figure 1 , this embodiment provides the following technical solutions:

[0115] The deep learning-based rehabilitation training system for MCI patients uses different scenarios, including: data collection module, data processing module, model building module, prediction analysis module and rehabilitation training module.

[0116] It should be noted that through the interactive communication among the data collection module, the data processing module, the model building module, the prediction analysis module and the rehabilitation training module, rehabilitation training plans for different scenarios can be formulated for MCI patients, rehabilitation training for different scenarios can be performed on MCI patients according to the rehabilitation training plans for different scenarios, the rehabilitation training plans based on different scenarios for MCI patients can be adjusted and optimized based on the feedback information of MCI patients, rehabilitation training for different scenarios can be performed on MCI patients, the rehabilitation training process can be adjusted and optimized in real time according to the rehabilitation training situation of MCI patients, and the rehabilitation training effect of MCI patients can be improved.

[0117] Among them, the data collection module is used to collect attribute information, medical history information, imaging information and genetic information of MCI patients, and determine multimodal real-time data based on MCI patients;

[0118] In this embodiment, as a preferred technical solution of the present invention, the data collection module includes:

[0119] An attribute collection unit, used for collecting attribute information of MCI patients in real time, wherein the attribute information of MCI patients includes name, gender and age;

[0120] A medical history collection unit, used to collect medical history information of MCI patients in real time, wherein the medical history information of MCI patients includes the occurrence and development of the disease, as well as the examination, diagnosis and treatment of MCI patients;

[0121] An image acquisition unit, used for real-time acquisition of image information of MCI patients;

[0122] A gene collection unit, used to collect genetic information of MCI patients in real time;

[0123] Among them, based on the attribute information, medical history information, imaging information and genetic information of MCI patients collected in real time, multimodal real-time data based on MCI patients is determined.

[0124] Among them, the data processing module is used to retrieve, convert and fuse the multimodal real-time data based on MCI patients, and determine the multimodal feature data based on MCI patients;

[0125] In this embodiment, as a preferred technical solution of the present invention, the data processing module includes:

[0126] A data retrieval unit, used for retrieving multimodal real-time data based on MCI patients;

[0127] Obtain multimodal real-time data based on MCI patients;

[0128] Based on the sequential retrieval method, multimodal real-time data based on MCI patients are retrieved;

[0129] Remove duplicate data, missing data and outliers from multimodal real-time data of MCI patients that are not valuable for rehabilitation training of MCI patients in different scenarios;

[0130] Identify multimodal real-time data that is valuable for rehabilitation training in different scenarios for MCI patients;

[0131] A data conversion unit, used for converting the retrieved multimodal real-time data;

[0132] Obtain multimodal real-time data that is valuable for rehabilitation training of MCI patients in different scenarios after retrieval;

[0133] After retrieval, the multimodal real-time data that is valuable for rehabilitation training of MCI patients in different scenarios is converted to normalize or standardize the multimodal real-time data;

[0134] Identify multimodal real-time data with a unified format and units of measurement;

[0135] A data fusion unit, used for fusing the converted multi-modal real-time data;

[0136] Get multi-modal real-time data with unified formats and units of measurement;

[0137] And integrate multi-modal real-time data;

[0138] Determine the multimodal feature data based on MCI patients.

[0139] Among them, the model building module is used to train the selected model architecture based on different scene rehabilitation training according to multimodal historical data to determine the deep learning model;

[0140] In this embodiment, as a preferred technical solution of the present invention, the model building module includes:

[0141] A data storage unit, used for storing multimodal historical data of MCI patients;

[0142] Among them, the stored multimodal historical data of MCI patients are extracted and used as the model training set;

[0143] A model selection unit, used to select a model architecture based on rehabilitation training in different scenarios;

[0144] Based on the rehabilitation training needs of MCI patients in different scenarios;

[0145] According to multimodal historical data, select the model architecture based on different scenarios of rehabilitation training;

[0146] Model training unit, used to train the selected model architecture and build a deep learning model;

[0147] Obtain the selected model architecture and multimodal historical data based on different scenario rehabilitation training;

[0148] Based on the rehabilitation training needs of MCI patients in different scenarios, the selected model architecture based on rehabilitation training in different scenarios is trained according to multimodal historical data to determine the deep learning model.

[0149] In this embodiment, as a preferred technical solution of the present invention, the selected model architecture is trained to construct a deep learning model, including:

[0150] Based on multimodal historical data, the selected model architecture based on rehabilitation training in different scenarios is trained to determine the deep learning model;

[0151] Among them, the deep learning model is verified;

[0152] Obtain a model validation set, validate the deep learning model based on the model validation set, and determine the model validation result;

[0153] Obtain model validation results, conduct in-depth mining and analysis on the model validation results, and determine the model adjustment and optimization plan;

[0154] The deep learning model is adjusted and optimized based on the model adjustment and optimization plan, and the best deep learning model is determined through iterative training and optimization process.

[0155] Among them, the prediction and analysis module is used to predict the multimodal feature data based on MCI patients based on the deep learning model, determine the model prediction results based on MCI patients, evaluate and analyze the model prediction results based on MCI patients, and determine the evaluation and analysis results based on MCI patients;

[0156] In this embodiment, as a preferred technical solution of the present invention, the prediction analysis module includes:

[0157] A model prediction unit, used to predict multimodal feature data based on MCI patients;

[0158] Get the best deep learning model;

[0159] Obtain multimodal feature data based on MCI patients;

[0160] Input multimodal feature data from MCI patients into the best deep learning model;

[0161] Predict multimodal feature data of MCI patients based on deep learning models;

[0162] Determine the model's predictions based on MCI patients;

[0163] An evaluation and analysis unit, used for evaluating and analyzing the prediction results of the model based on MCI patients;

[0164] Obtain model prediction results based on MCI patients;

[0165] Evaluate and analyze the prediction results of the model based on MCI patients;

[0166] Results were determined based on the analysis of the assessment of patients with MCI.

[0167] Among them, the rehabilitation training module is used to formulate rehabilitation training plans for MCI patients in different scenarios, and conduct rehabilitation training for MCI patients in different scenarios according to the rehabilitation training plans for different scenarios.

[0168] In this embodiment, as a preferred technical solution of the present invention, the rehabilitation training module includes:

[0169] Rehabilitation Development Unit, used to develop rehabilitation training plans for MCI patients in different scenarios;

[0170] Obtain assessment and analysis results based on MCI patients;

[0171] Conduct mining and analysis on the assessment and analysis results based on MCI patients;

[0172] Determine rehabilitation training programs based on different scenarios for MCI patients;

[0173] Rehabilitation training unit, used to provide rehabilitation training for MCI patients in different scenarios;

[0174] Obtain rehabilitation training plans based on different scenarios for MCI patients;

[0175] MCI patients are given rehabilitation training in different scenarios according to the rehabilitation training plan based on different scenarios of MCI patients.

[0176] In this embodiment, as a preferred technical solution of the present invention, the rehabilitation training module further includes:

[0177] Patient feedback unit, used to record the feedback information of MCI patients on rehabilitation training in different scenarios;

[0178] When MCI patients are undergoing rehabilitation training in different scenarios, real-time feedback information from MCI patients on rehabilitation training in different scenarios is obtained;

[0179] Record the real-time feedback information of MCI patients on rehabilitation training in different scenarios;

[0180] A program adjustment unit is used to adjust and optimize the rehabilitation training programs based on different scenarios of MCI patients;

[0181] Obtain the feedback information of MCI patients on rehabilitation training in different scenarios;

[0182] Adjust and optimize rehabilitation training programs based on different scenarios for MCI patients based on feedback from MCI patients;

[0183] Determine the best rehabilitation training program based on different scenarios for MCI patients;

[0184] MCI patients are given the best rehabilitation training in different scenarios according to the best rehabilitation training plan based on different scenarios of MCI patients.

[0185] Therefore, rehabilitation training plans for different scenarios can be formulated for MCI patients, rehabilitation training for MCI patients in different scenarios can be carried out according to the rehabilitation training plans for different scenarios, the rehabilitation training plans based on different scenarios for MCI patients can be adjusted and optimized based on the feedback information of MCI patients, rehabilitation training for MCI patients in different scenarios can be carried out, the rehabilitation training process can be adjusted and optimized in real time according to the rehabilitation training situation of MCI patients, and the rehabilitation training effect of MCI patients can be improved.

[0186] In this embodiment, as a preferred technical solution of the present invention, the solution adjustment unit includes:

[0187] A program temporary storage subunit is used to temporarily store the initial rehabilitation training program, and select the scene rehabilitation training program currently used by the MCI patient in the initial rehabilitation training program, and at the same time, use the scene rehabilitation training program as a prototype program;

[0188] The result prediction subunit is used to call the model of MCI patients, and evaluate and analyze the prototype solution based on the model of MCI patients to obtain prediction information of the prototype solution;

[0189] Contrast selection subunit, used for:

[0190] Read the feedback information, compare the feedback information with the predicted information, and set a comparison parameter range for the predicted information;

[0191] If the feedback information is within the control parameter range of the prediction information, the feedback information is imported into the normal feedback record set;

[0192] If the feedback information does not fall within the control parameter range of the prediction information, the feedback information is imported into the abnormal feedback record set;

[0193] Adjust the processing subunit to:

[0194] Retrieving feedback information in the abnormal feedback record set, and processing the feedback information in the abnormal feedback record set to obtain adjustment information;

[0195] Generate a replacement plan based on the adjustment information;

[0196] A replacement processing subunit is used to call the prototype scheme currently used by the MCI patient, and replace the prototype scheme based on the replacement scheme, and use the replaced prototype scheme as the scenario rehabilitation training scheme currently used by the MCI patient;

[0197] The information cache subunit is used to:

[0198] Obtaining behavior replacement records when replacing the prototype solution, and accumulating the records to obtain a first accumulation cycle;

[0199] Accumulate the normal feedback record set to obtain a second accumulation cycle;

[0200] A preset record accumulation period is obtained, and when the first accumulation period or the second accumulation period reaches the behavior replacement record or the normal feedback record set corresponding to the preset record accumulation period, the behavior replacement record or the normal feedback record set is removed.

[0201] In this embodiment, the initial rehabilitation training program is a rehabilitation training program obtained when the feedback information is not processed, wherein there are multiple initial rehabilitation training programs.

[0202] In this embodiment, the prototype program is a rehabilitation training program currently being used by MCI patients.

[0203] In this embodiment, the evaluation and analysis unit information is a logical relationship used by the evaluation and analysis unit, which facilitates obtaining the feedback information that the rehabilitation training program should theoretically obtain.

[0204] In this embodiment, the prediction information is the feedback information that the rehabilitation training program should theoretically obtain, and is used as a reference for the feedback information.

[0205] In this embodiment, the normal feedback record set is a collection of feedback information within a certain period of time and within the prediction information comparison parameter range.

[0206] In this embodiment, the abnormal feedback record set is a collection of feedback information that is outside the prediction information comparison parameter range within a certain period of time.

[0207] In this embodiment, the feedback information in the abnormal feedback record set is processed to obtain the adjustment information, which is to analyze the feedback information in the abnormal feedback record set again through the formulation strategy of the rehabilitation training program to obtain the implementation conditions of the feedback information in the program.

[0208] In this embodiment, the replacement scheme is a final implementation scheme obtained by replacing the behavior corresponding to the adjustment information in the prototype scheme through the replacement condition.

[0209] In this embodiment, the behavior replacement record is a behavior record of the prototype solution implemented by the replacement condition.

[0210] The working principle and beneficial effects of the above technical solution are: all rehabilitation training solutions are obtained through the solution temporary storage subunit, and the scenario rehabilitation training solution currently used by the MCI patient is selected as the prototype solution. After the special selection is completed, the model of the MCI patient in the evaluation and analysis unit is called through the result prediction subunit, and the prototype solution is substituted into the prototype solution to obtain the prediction information of the currently used rehabilitation training solution through the logical budget, so as to obtain the prediction information as a reference for comparing the feedback information, and then the feedback information is compared with the prediction information through the comparison selection subunit, and the comparison parameter range is set. If the feedback information is within the prediction information comparison parameter range, the feedback information is imported into the normal feedback record set. If the feedback information is not within the prediction information comparison parameter range, the feedback information is imported into the abnormal feedback record set, so as to compare the feedback information with the prediction information, and then the adjustment processing subunit is used. Element, call the feedback information in the abnormal feedback record set, and process the feedback information in the abnormal feedback record set to obtain adjustment information, import the adjustment information into the rehabilitation designated unit to obtain a replacement plan, so as to obtain a replacement plan, call the prototype plan currently used by the MCI patient, replace the corresponding behavior in the scene rehabilitation training plan with the obtained replacement plan, set the replaced scene rehabilitation training plan as the scene rehabilitation training plan currently used by the MCI patient, complete the plan content replacement optimization, and finally import the replaced corresponding behavior into the cache record through the information cache sub-unit to obtain the behavior replacement record, accumulate the record in a certain period of time, obtain the feedback information in the normal feedback record set, accumulate the record in a certain period of time, and when the cycle is completed, delete all the accumulated records to cache a certain amount of records and process them to save memory usage.

[0211] In this embodiment, as a preferred technical solution of the present invention, the rehabilitation training module further includes:

[0212] A data acquisition unit is used to obtain theoretical rehabilitation data corresponding to the rehabilitation training program for each scenario after formulating rehabilitation training programs for different scenarios for MCI patients, and at the same time, obtain actual rehabilitation data corresponding to the rehabilitation training programs for different scenarios;

[0213] A first calculation unit is used to calculate the comprehensive rehabilitation efficiency of rehabilitation training for MCI patients according to theoretical rehabilitation data and actual rehabilitation data corresponding to rehabilitation training programs in different scenarios;

[0214]

[0215] Where η represents the comprehensive rehabilitation efficiency of rehabilitation training for MCI patients; i represents the sequence number of the scene; n represents the total number of scenes; x i represents the actual rehabilitation data corresponding to the rehabilitation training program of the i-th scenario; Xi represents the theoretical rehabilitation data corresponding to the rehabilitation training program of the i-th scenario, and x i ≤X i ; δ i represents the influence weight of the rehabilitation training program on the rehabilitation data in the i-th scenario; z represents the total actual rehabilitation data of MCI patients after receiving the rehabilitation training programs of all scenarios; Z represents the total theoretical rehabilitation data of MCI patients after receiving the rehabilitation training programs of all scenarios, and z≤Z;

[0216] The second calculation unit is used to retrieve the satisfaction survey data corresponding to the rehabilitation training program of each scene, and at the same time, calculate the total satisfaction of the MCI patient with rehabilitation training according to the satisfaction survey data corresponding to the rehabilitation training program of each scene;

[0217]

[0218] Where M represents the overall satisfaction of rehabilitation training for MCI patients; m i represents the satisfaction survey data corresponding to the rehabilitation training program of the i-th scenario; φ i represents the impact weight of the rehabilitation training program in the i-th scenario on the total satisfaction;

[0219] Scoring unit for:

[0220] Conduct a first rehabilitation effect evaluation based on the comprehensive rehabilitation efficiency of rehabilitation training for MCI patients and determine the first score for rehabilitation training;

[0221] Conduct a second rehabilitation effect evaluation based on the overall satisfaction of rehabilitation training for MCI patients and determine the second rehabilitation training score;

[0222] The comprehensive score of the rehabilitation effect of MCI patients during rehabilitation training is calculated based on the first score and the second score;

[0223] F = f1*σ1+f2*σ2;

[0224] Among them, F represents the comprehensive score of rehabilitation effect of MCI patients during rehabilitation training; f1 represents the first score; σ1 represents the influence weight of the comprehensive rehabilitation efficiency of rehabilitation training on the comprehensive score; f2 represents the second score; σ2 represents the influence weight of the total satisfaction of MCI patients with rehabilitation training on the comprehensive score;

[0225] A report generating unit, configured to generate a comprehensive evaluation report for the MCI patient after the rehabilitation training according to the rehabilitation training programs of different scenarios based on the first score, the second score and the comprehensive score;

[0226] The transmission unit is used to transmit the comprehensive evaluation report to the monitoring terminal.

[0227] In this embodiment, the first rehabilitation effect evaluation can be an evaluation of the rehabilitation effect of rehabilitation training for MCI patients based on the comprehensive rehabilitation efficiency of the rehabilitation training. For example, the maximum value of the comprehensive rehabilitation efficiency of the rehabilitation training is 100%, and the calculated value of the comprehensive rehabilitation efficiency of the rehabilitation training is 80%, then the first score is 80 points.

[0228] In this embodiment, the second rehabilitation effect evaluation can be an evaluation of the rehabilitation effect of the rehabilitation training for MCI patients based on the total satisfaction with the rehabilitation training. For example, the maximum value of the total satisfaction is 100, and the actual calculated value of the total satisfaction with the rehabilitation training for MCI patients is 85, then the second score is 85 points.

[0229] In this embodiment, the monitoring terminal may be a mobile phone, a computer or other device.

[0230] The working principle and beneficial effects of the above technical solution are: after formulating rehabilitation training plans for MCI patients in different scenarios, by obtaining the theoretical rehabilitation data of the rehabilitation training plan corresponding to each scenario and the actual rehabilitation data corresponding to the rehabilitation training plans for different scenarios, the comprehensive rehabilitation efficiency of the rehabilitation training for MCI patients is effectively calculated; by retrieving the satisfaction survey data corresponding to the rehabilitation training plan for each scenario, the total satisfaction of the rehabilitation training for MCI patients is effectively calculated; and then a first score is determined according to the comprehensive rehabilitation efficiency of the rehabilitation training for MCI patients and a second score is determined according to the total satisfaction of the rehabilitation training for MCI patients; and based on the first score and the second score, a comprehensive score of the rehabilitation effect of the rehabilitation training for MCI patients can be effectively calculated, thereby realizing the evaluation and monitoring efficiency of rehabilitation training for MCI patients in different scenarios, and then effectively realizing the reading of the comprehensive evaluation report based on the monitoring terminal, thereby improving the monitoring efficiency and evaluation objectivity of the rehabilitation training plan.

[0231] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0232] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A deep learning-based rehabilitation training system for MCI patients in different scenarios, characterized by: include: A data collection module is used to collect attribute information, medical history information, imaging information and genetic information of MCI patients, and determine multimodal real-time data based on MCI patients; A data processing module is used to retrieve, convert and fuse multimodal real-time data based on MCI patients to determine multimodal feature data based on MCI patients; A model building module is used to train the selected model architecture based on different scene rehabilitation training according to multimodal historical data to determine the deep learning model; A prediction and analysis module, used to predict the multimodal feature data based on MCI patients based on the deep learning model, determine the model prediction results based on MCI patients, evaluate and analyze the model prediction results based on MCI patients, and determine the evaluation and analysis results based on MCI patients; The rehabilitation training module is used to formulate rehabilitation training plans for MCI patients in different scenarios, and conduct rehabilitation training for MCI patients in different scenarios according to the rehabilitation training plans for different scenarios; The rehabilitation training module comprises: Patient feedback unit, used to record the feedback information of MCI patients on rehabilitation training in different scenarios; When MCI patients are undergoing rehabilitation training in different scenarios, real-time feedback information from MCI patients on rehabilitation training in different scenarios is obtained; Record the real-time feedback information of MCI patients on rehabilitation training in different scenarios; A program adjustment unit is used to adjust and optimize the rehabilitation training programs based on different scenarios of MCI patients; Obtain the feedback information of MCI patients on rehabilitation training in different scenarios; Adjust and optimize rehabilitation training programs based on different scenarios for MCI patients based on feedback from MCI patients; Determine the best rehabilitation training program based on different scenarios for MCI patients; Provide MCI patients with the best rehabilitation training based on different scenarios of MCI patients; The scheme adjustment unit comprises: A program temporary storage subunit is used to temporarily store the initial rehabilitation training program, and select the scene rehabilitation training program currently used by the MCI patient in the initial rehabilitation training program, and at the same time, use the scene rehabilitation training program as a prototype program; The result prediction subunit is used to call the model of MCI patients, and evaluate and analyze the prototype solution based on the model of MCI patients to obtain prediction information of the prototype solution; Contrast selection subunit, used for: Read the feedback information, compare the feedback information with the predicted information, and set a comparison parameter range for the predicted information; If the feedback information is within the control parameter range of the prediction information, the feedback information is imported into the normal feedback record set; If the feedback information does not fall within the control parameter range of the prediction information, the feedback information is imported into the abnormal feedback record set; Adjust the processing subunit to: Retrieving feedback information in the abnormal feedback record set, and processing the feedback information in the abnormal feedback record set to obtain adjustment information; Generate a replacement plan based on the adjustment information; A replacement processing subunit is used to call the prototype scheme currently used by the MCI patient, and replace the prototype scheme based on the replacement scheme, and use the replaced prototype scheme as the scenario rehabilitation training scheme currently used by the MCI patient; The information cache subunit is used to: Obtaining behavior replacement records when replacing the prototype solution, and accumulating the records to obtain a first accumulation cycle; Accumulate the normal feedback record set to obtain a second accumulation cycle; Obtaining a preset record accumulation period, and removing the behavior replacement record or normal feedback record set corresponding to the preset record accumulation period when the first accumulation period or the second accumulation period reaches the preset record accumulation period; The rehabilitation training module also includes: A data acquisition unit is used to obtain theoretical rehabilitation data corresponding to the rehabilitation training program for each scenario after formulating rehabilitation training programs for different scenarios for MCI patients, and at the same time, obtain actual rehabilitation data corresponding to the rehabilitation training programs for different scenarios; A first calculation unit is used to calculate the comprehensive rehabilitation efficiency of rehabilitation training for MCI patients according to theoretical rehabilitation data and actual rehabilitation data corresponding to rehabilitation training programs in different scenarios; ； in, It represents the comprehensive rehabilitation efficiency of rehabilitation training for MCI patients; Indicates the sequence number value of the scene; Indicates the total number of scenes; Indicates The actual rehabilitation data corresponding to the rehabilitation training program for each scenario; Indicates The theoretical rehabilitation data corresponding to the rehabilitation training program for each scenario, and ; Indicated in The weight of the impact of the rehabilitation training program on the rehabilitation data in each scenario; It represents the total actual rehabilitation data after MCI patients received rehabilitation training programs in all scenarios; represents the total theoretical rehabilitation data of MCI patients after receiving rehabilitation training programs in all scenarios, and ; The second calculation unit is used to retrieve the satisfaction survey data corresponding to the rehabilitation training program of each scene, and at the same time, calculate the total satisfaction of the MCI patient with rehabilitation training according to the satisfaction survey data corresponding to the rehabilitation training program of each scene; ； in, Indicates the overall satisfaction with rehabilitation training for MCI patients; Indicates Satisfaction survey data corresponding to the rehabilitation training programs in each scenario; Indicates The weight of the impact of the rehabilitation training program in each scenario on the overall satisfaction; Scoring unit for: Conduct a first rehabilitation effect evaluation based on the comprehensive rehabilitation efficiency of rehabilitation training for MCI patients and determine the first score for rehabilitation training; Conduct a second rehabilitation effect evaluation based on the overall satisfaction of rehabilitation training for MCI patients and determine the second rehabilitation training score; The comprehensive score of the rehabilitation effect of MCI patients during rehabilitation training is calculated based on the first score and the second score; ； in, It represents the comprehensive score of the rehabilitation effect of MCI patients during rehabilitation training; Indicates the first rating; It represents the weight of the comprehensive rehabilitation efficiency of rehabilitation training on the comprehensive score; Indicates the second rating; It represents the weight of the total satisfaction of MCI patients with rehabilitation training on the comprehensive score; A report generating unit, configured to generate a comprehensive evaluation report for the MCI patient after the rehabilitation training according to the rehabilitation training programs of different scenarios based on the first score, the second score and the comprehensive score; The transmission unit is used to transmit the comprehensive evaluation report to the monitoring terminal.

2. The deep learning-based rehabilitation training system for MCI patients in different scenarios according to claim 1, characterized in that: The data collection module comprises: An attribute collection unit, used for collecting attribute information of MCI patients in real time, wherein the attribute information of MCI patients includes name, gender and age; A medical history collection unit, used to collect medical history information of MCI patients in real time, wherein the medical history information of MCI patients includes the occurrence and development of the disease, as well as the examination, diagnosis and treatment of MCI patients; An image acquisition unit, used for real-time acquisition of image information of MCI patients; A gene collection unit, used to collect genetic information of MCI patients in real time; Among them, based on the attribute information, medical history information, imaging information and genetic information of MCI patients collected in real time, multimodal real-time data based on MCI patients is determined.

3. The deep learning-based rehabilitation training system for MCI patients in different scenarios according to claim 2, characterized in that: The data processing module comprises: A data retrieval unit, used for retrieving multimodal real-time data based on MCI patients; Obtain multimodal real-time data based on MCI patients; Based on the sequential retrieval method, multimodal real-time data based on MCI patients are retrieved; Remove duplicate data, missing data and outliers from multimodal real-time data of MCI patients that are not valuable for rehabilitation training of MCI patients in different scenarios; Identify multimodal real-time data that is valuable for rehabilitation training in different scenarios for MCI patients; A data conversion unit, used for converting the retrieved multimodal real-time data; Obtain multimodal real-time data that is valuable for rehabilitation training of MCI patients in different scenarios after retrieval; After retrieval, the multimodal real-time data that is valuable for rehabilitation training of MCI patients in different scenarios is converted to normalize or standardize the multimodal real-time data; Identify multimodal real-time data with a unified format and units of measurement; A data fusion unit, used for fusing the converted multi-modal real-time data; Get multi-modal real-time data with unified formats and units of measurement; And integrate multi-modal real-time data; Determine the multimodal feature data based on MCI patients.

4. The deep learning-based rehabilitation training system for MCI patients in different scenarios according to claim 3, characterized in that: The model building module includes: A data storage unit, used for storing multimodal historical data of MCI patients; Among them, the stored multimodal historical data of MCI patients are extracted and used as the model training set; A model selection unit, used to select a model architecture based on rehabilitation training in different scenarios; Based on the rehabilitation training needs of MCI patients in different scenarios; According to multimodal historical data, select the model architecture based on different scenarios of rehabilitation training; Model training unit, used to train the selected model architecture and build a deep learning model; Obtain the selected model architecture and multimodal historical data based on different scenario rehabilitation training; Based on the rehabilitation training needs of MCI patients in different scenarios, the selected model architecture based on rehabilitation training in different scenarios is trained according to multimodal historical data to determine the deep learning model.

5. The deep learning-based rehabilitation training system for MCI patients in different scenarios according to claim 4, characterized in that: Train the selected model architecture to build a deep learning model, including: Based on multimodal historical data, the selected model architecture based on rehabilitation training in different scenarios is trained to determine the deep learning model; Among them, the deep learning model is verified; Obtain a model validation set, validate the deep learning model based on the model validation set, and determine the model validation result; Obtain model validation results, conduct in-depth mining and analysis on the model validation results, and determine the model adjustment and optimization plan; The deep learning model is adjusted and optimized based on the model adjustment and optimization plan, and the best deep learning model is determined through iterative training and optimization process.

6. The deep learning-based rehabilitation training system for MCI patients in different scenarios according to claim 5, characterized in that: The prediction analysis module includes: A model prediction unit, used to predict multimodal feature data based on MCI patients; Get the best deep learning model; Obtain multimodal feature data based on MCI patients; Input multimodal feature data from MCI patients into the best deep learning model; Predict multimodal feature data of MCI patients based on deep learning models; Determine the model's predictions based on MCI patients; An evaluation and analysis unit, used for evaluating and analyzing the prediction results of the model based on MCI patients; Obtain model prediction results based on MCI patients; Evaluate and analyze the prediction results of the model based on MCI patients; Results were determined based on the analysis of the assessment of patients with MCI.

7. The deep learning-based rehabilitation training system for MCI patients in different scenarios according to claim 6, characterized in that: The rehabilitation training module includes: Rehabilitation Development Unit, used to develop rehabilitation training plans for MCI patients in different scenarios; Obtain assessment and analysis results based on MCI patients; Conduct mining and analysis on the assessment and analysis results based on MCI patients; Determine rehabilitation training programs based on different scenarios for MCI patients; Rehabilitation training unit, used to provide rehabilitation training for MCI patients in different scenarios; Obtain rehabilitation training plans based on different scenarios for MCI patients; MCI patients are given rehabilitation training in different scenarios according to the rehabilitation training plan based on different scenarios of MCI patients.

Images