Rehabilitation training assisting method and system based on artificial intelligence and terminal equipment

Abstract

The embodiment of the invention provides a rehabilitation training assisting method and system based on artificial intelligence and terminal equipment, and belongs to the technical field of artificial intelligence. The method comprises the following steps: acquiring preoperative body data corresponding to a target patient before an operation; acquiring corresponding operation record data of the target patient in the operation; obtaining fixed-item rehabilitation data of the target patient after the operation; analyzing and generating an initial postoperative rehabilitation plan according to the preoperative body data, the fixed rehabilitation data and the operation record data; dynamic rehabilitation monitoring data of the target patient in the initial rehabilitation training process according to the initial postoperative rehabilitation plan are obtained; based on a postoperative management scheme of a preset rehabilitation management mode, the rehabilitation progress of the target patient is predicted according to the dynamic rehabilitation monitoring data, and a rehabilitation progress prediction result is generated; and adjusting the initial postoperative rehabilitation plan according to the rehabilitation progress prediction result to obtain a target postoperative rehabilitation plan until the target patient meets the discharge condition of the preset rehabilitation management mode.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an artificial intelligence-based rehabilitation training auxiliary method, system and terminal device. Background Art

[0002] Against the backdrop of the advancement of modern medicine, perioperative management has gradually received more attention, especially the promotion of the concept of accelerated recovery surgery, which marks a major shift in the management of surgical patients. In addition, the concept of accelerated recovery surgery has further changed the concepts and practices of traditional surgery for perioperative management. These new concepts of preset rehabilitation management models significantly accelerate the recovery of patients and improve their quality of life after surgery by combining advanced minimally invasive surgical techniques with refined management measures. The preset rehabilitation model can effectively shorten the patient's recovery time, reduce the occurrence of complications, and improve the overall treatment effect. However, in the preset rehabilitation management model, medical staff cannot promptly judge and evaluate whether the postoperative rehabilitation plan truly meets the patient's actual needs. This lag may result in the inability to make timely adjustments to the rehabilitation strategy, thereby affecting the patient's recovery effect. Summary of the invention

[0003] The main purpose of the embodiments of the present invention is to provide an artificial intelligence-based rehabilitation training assistance method, system and terminal device, aiming to solve the problem that in the preset rehabilitation management mode, medical staff are unable to timely judge and evaluate whether the postoperative rehabilitation plan truly meets the actual needs of the patient. This lag may result in the inability to make timely adjustments to the rehabilitation strategy, thereby affecting the patient's recovery effect.

[0004] In a first aspect, an embodiment of the present invention provides a rehabilitation training auxiliary method based on artificial intelligence, comprising: Obtaining preoperative physical data corresponding to the target patient before surgery, wherein the preoperative physical data includes health data, medical history, and preoperative examination results; Obtaining surgical record data corresponding to the target patient during surgery, wherein the surgical record data is surgical process data obtained during the process of adjusting the target patient's physical condition to meet the inclusion conditions of the preset rehabilitation management model and excluding contraindications for surgery; Obtaining the targeted rehabilitation data of the target patient after surgery; wherein the targeted rehabilitation data includes movement status, pain score and functional recovery; Analyze the preoperative physical data, the fixed rehabilitation data and the surgical record data to generate an initial postoperative rehabilitation plan; the initial postoperative rehabilitation plan is used to guide the target patient to perform preliminary rehabilitation training; Acquiring dynamic rehabilitation monitoring data of the target patient during the initial rehabilitation training according to the initial postoperative rehabilitation plan; Based on the postoperative management plan of the preset rehabilitation management mode, predicting the rehabilitation progress of the target patient according to the dynamic rehabilitation monitoring data, and generating a rehabilitation progress prediction result; The initial postoperative rehabilitation plan is adjusted according to the rehabilitation progress prediction result to obtain a target postoperative rehabilitation plan, and the target postoperative rehabilitation plan is used to guide the target patient to perform target rehabilitation training until the target patient meets the discharge conditions of the preset rehabilitation management model.

[0005] In a second aspect, an embodiment of the present invention provides a rehabilitation training auxiliary system based on artificial intelligence, comprising: A first acquisition module is used to obtain the preoperative physical data of the target patient before the operation, wherein the preoperative physical data includes health data, medical history and preoperative examination results; The second acquisition module is used to obtain the surgical record data corresponding to the target patient during the operation, wherein the surgical record data is the surgical process data obtained during the operation of the target patient after adjusting the physical condition of the target patient to meet the inclusion conditions of the preset rehabilitation management mode and excluding contraindications; The third acquisition module is used to obtain the specific rehabilitation data of the target patient after surgery; wherein the specific rehabilitation data includes movement status, pain score and functional recovery; A plan generation module, used to generate an initial postoperative rehabilitation plan based on the preoperative physical data, the fixed rehabilitation data and the surgical record data; the initial postoperative rehabilitation plan is used to guide the target patient to perform preliminary rehabilitation training; A rehabilitation monitoring module, used to obtain dynamic rehabilitation monitoring data of the target patient during the initial rehabilitation training according to the initial postoperative rehabilitation plan; A progress prediction module, which is used to predict the rehabilitation progress of the target patient based on the postoperative management plan of the preset rehabilitation management mode and the dynamic rehabilitation monitoring data, and generate a rehabilitation progress prediction result; A plan adjustment module is used to adjust the initial postoperative rehabilitation plan according to the rehabilitation progress prediction result to obtain a target postoperative rehabilitation plan, and the target postoperative rehabilitation plan is used to guide the target patient to perform target rehabilitation training until the target patient meets the discharge conditions of the preset rehabilitation management model.

[0006] In a third aspect, an embodiment of the present invention further provides a terminal device, comprising a processor, a memory, a computer program stored in the memory and executable by the processor, and a data bus for realizing connection and communication between the processor and the memory, wherein when the computer program is executed by the processor, the steps of any one of the artificial intelligence-based rehabilitation training assistance methods provided in the specification of the present invention are implemented.

[0007] In a fourth aspect, an embodiment of the present invention further provides a storage medium for computer-readable storage, characterized in that the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any one of the artificial intelligence-based rehabilitation training assistance methods provided in the specification of the present invention.

[0008] The embodiment of the present invention provides a rehabilitation training auxiliary method, system and terminal device based on artificial intelligence, the method comprising: obtaining preoperative physical data corresponding to a target patient before surgery, wherein the preoperative physical data includes health data, medical history and preoperative examination results; obtaining surgical record data corresponding to the target patient during surgery, wherein the surgical record data is data obtained after the target patient adjusts his physical condition to meet the inclusion conditions of a preset rehabilitation management mode and excludes contraindications for surgery; obtaining fixed-item rehabilitation data of the target patient after surgery; wherein the fixed-item rehabilitation data includes movement conditions, pain scores and functional recovery; analyzing the preoperative physical data, fixed-item rehabilitation data and surgical record data to generate an initial postoperative rehabilitation plan; the initial postoperative rehabilitation plan is used to guide the target patient to perform preliminary rehabilitation training; obtaining dynamic rehabilitation monitoring data of the target patient during the preliminary rehabilitation training according to the initial postoperative rehabilitation plan; predicting the rehabilitation progress of the target patient according to the dynamic rehabilitation monitoring data based on a postoperative management plan of the preset rehabilitation management mode, and generating a rehabilitation progress prediction result; adjusting the initial postoperative rehabilitation plan according to the rehabilitation progress prediction result to obtain a target postoperative rehabilitation plan, and the target postoperative rehabilitation plan is used to guide the target patient to perform target rehabilitation training until the target patient meets the discharge conditions of the preset rehabilitation management mode. This method relies on dynamic rehabilitation monitoring data to automatically and timely predict the rehabilitation progress prediction results of the target patient. With accurate rehabilitation progress prediction results, medical staff can quickly adjust the postoperative rehabilitation plan of the target patient. This flexible management method enables the medical team to efficiently implement the rehabilitation plan, thereby optimizing the patient's recovery process, reducing the risk of re-hospitalization, and ultimately achieving more ideal clinical results. In addition, this method also solves the problem that in the preset rehabilitation management model, medical staff cannot timely judge and evaluate whether the postoperative rehabilitation plan truly meets the patient's actual needs. This lag may lead to the inability to make timely adjustments to the rehabilitation strategy, thereby affecting the patient's recovery effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A flowchart of an artificial intelligence-based rehabilitation training assistance method provided by an embodiment of the present invention; Figure 2 A schematic diagram of the module structure of a rehabilitation training auxiliary system based on artificial intelligence provided by an embodiment of the present invention; Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0011] 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 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.

[0012] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0013] It should be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0014] The embodiment of the present invention provides a method, system and terminal device for assisting rehabilitation training based on artificial intelligence. The method for assisting rehabilitation training based on artificial intelligence can be applied to a terminal device, which can be an electronic device such as a tablet computer, a laptop computer, a desktop computer, a personal digital assistant and a wearable device. The terminal device can be a server or a server cluster.

[0015] Some embodiments of the present invention are described in detail below in conjunction with the accompanying drawings. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0016] Please refer to Figure 1 , Figure 1 A flowchart of an artificial intelligence-based rehabilitation training assistance method provided in an embodiment of the present invention.

[0017] like Figure 1 As shown, the artificial intelligence-based rehabilitation training assistance method includes steps S101 to S107.

[0018] Step S101: Obtain preoperative physical data corresponding to the target patient before the operation, wherein the preoperative physical data includes health data, medical history, and preoperative examination results.

[0019] For example, the target patient is a patient who needs surgery, and the preoperative physical data is the target patient's corresponding health data, medical history, and preoperative examination results before the surgery, for example, the target patient's medical history data is obtained through a questionnaire, an interview, or an electronic medical record system. The target patient is then arranged to undergo necessary preoperative examinations, such as imaging examinations (such as X-rays, CT, MRI), laboratory tests (such as blood tests, urine tests), and physiological index measurements (such as blood pressure, heart rate), so as to obtain the preoperative examination results collected by the target patient under the preoperative examination.

[0020] Step S102, obtaining the surgical record data corresponding to the target patient during the operation, wherein the surgical record data is the surgical process data obtained during the operation of the target patient after adjusting his / her physical condition to meet the inclusion conditions of the preset rehabilitation management mode and excluding contraindications.

[0021] Exemplarily, after the target patient meets the preset rehabilitation management model inclusion criteria and undergoes surgery without contraindications, the corresponding surgical record data of the target patient during the surgery is obtained.

[0022] For example, surgical record data includes the following: basic information of the target patient, such as name, gender, department, ward, bed number, inpatient medical record number or medical record number; record the date and time when the operation starts and ends; clearly record the type of operation performed to facilitate the subsequent medical team to understand the details of the operation and perform corresponding treatment; include the names of the surgeon, first assistant and other personnel involved in the operation to ensure that the responsibility for the operation is clear; record the anesthesia method used, anesthetic drugs and dosage, anesthesia provider and other information to evaluate the effect and safety of anesthesia; describe the surgical process in detail, including patient position, skin disinfection method, incision site, exploration situation and main lesion site, surgical method and steps, and inventory of dressings and instruments after the operation; record any unexpected situations that occur during the operation (such as bleeding, injury, etc.) and corresponding treatment measures to ensure patient safety and surgical quality.

[0023] In some embodiments, before obtaining the surgical record data corresponding to the target patient during the operation, the method also includes: using a risk prediction model to perform risk prediction on the preoperative physical data to obtain a perioperative risk prediction result; generating preoperative intervention recommendation information based on the perioperative risk prediction result, and the preoperative intervention recommendation information is used to guide the target patient to adjust his physical condition to meet the inclusion conditions of a preset rehabilitation management model, wherein the preoperative intervention recommendation information includes lifestyle adjustment information, psychological support information, and a pre-rehabilitation plan.

[0024] Exemplarily, a suitable risk prediction model (such as a classification model, a machine learning model, etc.) is selected based on the type of surgery and data characteristics, and the model is trained using historical data and known surgical risks to ensure that the risk prediction model can accurately predict risks.

[0025] Exemplarily, the preoperative physical data is input into the relationship recognition layer of the risk prediction model, and the relationship recognition layer is used to analyze the correlation and potential relationship between various factors in the preoperative physical data as the target relationship output by the relationship recognition layer. For example, the connection between certain physiological indicators or the pattern related to surgical risk, so as to determine the key relationship between various factors in the preoperative physical data based on the target relationship, such as which indicators are most important for risk prediction.

[0026] For example, statistical methods (such as Pearson correlation coefficient and Spearman rank correlation coefficient) are used to calculate the correlation between any two physiological indicators in the preoperative physical data, determine which physiological indicators have significant linear or nonlinear relationships, and thus determine the correlation coefficients between the various physiological indicators, and then determine the target relationship between any two physiological indicators in the preoperative physical data based on the correlation coefficients.

[0027] Exemplarily, the preoperative body data is input into the feature representation layer of the risk prediction model. The feature representation layer converts the preoperative body data into an initial feature vector through a feature representation method (such as embedding, transformation, or mapping). The feature representation layer extracts key features from the data and encodes these features, thereby obtaining an initial feature vector corresponding to the preoperative body data based on the processing results of the feature representation layer.

[0028] Exemplarily, the initial feature vector and the identified target relationship are input into the relationship fusion layer of the risk prediction model, so that the relationship fusion layer combines the target relationship in the preoperative body data, further processes and fuses the initial feature vector, and obtains the target feature vector.

[0029] For example, a feature fusion strategy is formulated according to the target relationship to determine how to apply the target relationship to the initial feature vector. Feature fusion strategies include weighted fusion, feature interaction, etc. Therefore, according to the predefined feature fusion strategy, possible operations include weighting the features, generating new interactive features, or transforming using a relationship mapping function, so that the initial feature vector is combined with the feature fusion strategy corresponding to the target relationship to obtain the corresponding target feature vector. The target feature vector contains the feature information processed by the target relationship, which can better reflect the potential risk factors in the preoperative body data.

[0030] Exemplarily, the obtained target feature vector is input into the risk prediction layer of the risk prediction model for further risk assessment and prediction, thereby obtaining the perioperative risk prediction result corresponding to the target patient.

[0031] Specifically, the relationship recognition layer can provide a deeper understanding of the relationship between various physiological features in preoperative body data, which helps to improve the explanatory power and accuracy of the prediction model. The relationship fusion layer integrates the initial feature vector using the identified target relationship to generate a target feature vector that better reflects the complex interaction of preoperative body data, which can effectively improve the model performance and provide support for the subsequent accurate perioperative risk prediction results.

[0032] Exemplarily, an acceptable risk threshold is set according to the risk requirements of the target surgery. Preoperative intervention recommendation information is generated based on the predicted perioperative risk prediction results. The generated preoperative intervention recommendation information may include lifestyle adjustment information, psychological support information, and pre-rehabilitation plan. The risk threshold is compared with the perioperative risk prediction results, and then the preoperative intervention recommendation information corresponding to the target patient is obtained by searching in the preset rules according to the comparison results. That is, personalized preoperative medical intervention recommendations and preparation plans are provided to the target patient based on the perioperative risk prediction results, including but not limited to lifestyle adjustment information, psychological support information, and pre-rehabilitation plans.

[0033] For example, a detailed preset rule is compiled based on historical experience, which contains intervention strategies corresponding to different risk levels. For example, the risk level can be divided into three levels: low, medium, and high, and specific intervention measures for each level can be defined. The preset rules should include a detailed description of each risk level and the corresponding intervention strategies (such as lifestyle adjustments, drug treatment, further examinations, psychological interventions, etc.). In this way, the preset rules are classified according to the risk level to ensure that each risk level has corresponding preoperative intervention recommendation information. According to the comparison results, the preoperative intervention recommendation information corresponding to the perioperative risk prediction results of the target patient in the preset rules is found. For example, if the perioperative risk prediction result of the target patient is medium, the intervention measures corresponding to the "medium risk" defined in the preset rules are found, such as enhanced monitoring, adjustment of drug dosage, and increase of regular examinations.

[0034] Exemplarily, after the target patient executes the preoperative intervention recommendation information for a period of time, the third physical data corresponding to the target patient under the preoperative intervention recommendation information is obtained, and when the perioperative risk prediction result corresponding to the third physical data meets the corresponding risk requirements during the operation, the operation can be performed on the target patient; when the perioperative risk prediction result corresponding to the third physical data does not meet the corresponding risk requirements during the operation, the preoperative intervention recommendation information continues to be used to perform physical adjustments on the target patient until the perioperative risk prediction result corresponding to the third physical data meets the corresponding risk requirements during the operation.

[0035] Specifically, by analyzing preoperative physical data through risk prediction models, the surgical risk of each target patient can be accurately assessed, avoiding a one-size-fits-all risk assessment method. The preoperative intervention recommendation information formulated based on perioperative risk prediction results and preset rules is highly personalized and can provide appropriate preoperative intervention recommendation information based on the specific risk level of the target patient, so that each target patient can obtain the most suitable intervention method for him or her. Therefore, effective preoperative intervention recommendation information and improved health status can enhance the target patient's confidence in surgery and improve their overall medical experience.

[0036] Step S103, obtaining the targeted rehabilitation data of the target patient after surgery; wherein the targeted rehabilitation data includes movement status, pain score and functional recovery.

[0037] For example, the specific rehabilitation data that needs to be collected after the operation is determined, and the specific rehabilitation data includes movement status, pain score and functional recovery. After the operation, the target patient is arranged to undergo necessary postoperative examinations according to the doctor's advice and rehabilitation plan. For example, imaging examinations, laboratory tests and monitoring of physiological indicators are performed to obtain the specific rehabilitation data of the target patient after the operation, and the specific rehabilitation data includes movement status, pain score and functional recovery.

[0038] Step S104, analyzing the preoperative physical data, the fixed rehabilitation data and the surgical record data to generate an initial postoperative rehabilitation plan; the initial postoperative rehabilitation plan is used to guide the target patient to perform preliminary rehabilitation training.

[0039] For example, based on expert experience or the attending physician's analysis of preoperative physical data, specific rehabilitation data, and surgical record data, an initial postoperative rehabilitation plan corresponding to the target patient is generated; the initial postoperative rehabilitation plan is used to guide the target patient to conduct preliminary rehabilitation training.

[0040] Exemplarily, data with the same physiological characteristics, such as blood pressure, heart rate, body temperature, etc., are extracted from preoperative physical data, fixed rehabilitation data, and surgical record data, and the extracted common physiological characteristic data are compared at the same standard and time point to ensure data consistency and accuracy. Change detection methods, such as difference analysis or time series analysis, are used to identify changes in the same physiological characteristics in preoperative physical data, fixed rehabilitation data, and surgical record data. Change indicators, such as change rate, trend analysis, etc., are calculated based on the data changes to obtain target feature data. The target feature data reflects changes in physiological characteristics before and after surgery.

[0041] For example, the target feature data is input into the strategy classification model, and the trained strategy classification model is used to predict the input preoperative physical data, fixed rehabilitation data, and surgical record data. The model will output one or more rehabilitation strategy types, such as "physical therapy", "drug management" or "nutritional support".

[0042] Exemplarily, the obtained rehabilitation type is combined with the target feature data to analyze the relationship between the rehabilitation type and the target feature data. For example, certain rehabilitation types may require different rehabilitation frequencies or intensities, depending on the physiological state and needs of the target patient. Therefore, based on existing clinical standards or guidelines, combined with the target feature data, the rehabilitation frequency or intensity corresponding to each rehabilitation type is determined. For example, physical therapy may require a certain number of training sessions per day, while drug management may involve specific dosages and frequencies of medication, and then personalized adjustments are made based on the target feature data.

[0043] Exemplarily, the determined rehabilitation type is integrated with the corresponding rehabilitation frequency or intensity to form a comprehensive initial postoperative rehabilitation plan. For example, the rehabilitation type for a target patient is physical therapy, and the rehabilitation frequency is set to three times a week, 30 minutes each time. The initial postoperative rehabilitation plan includes the rehabilitation type, rehabilitation frequency or intensity.

[0044] In order to be able to formulate a more personalized rehabilitation plan based on the target patient's own situation, the target patient can be analyzed by the influencing factors that produce the event that requires rehabilitation. In this embodiment, the target patient's preoperative physical condition, intraoperative record data, and postoperative fixed rehabilitation data can be used as three dimensions for evaluating the rehabilitation plan for comprehensive consideration, so as to formulate a rehabilitation plan that is conducive to the target patient's rapid recovery and reaching the discharge standard. In this embodiment, preoperative intervention recommendation information can be determined based on preoperative physical data, and the preoperative intervention recommendation information is used to guide the target patient to adjust his physical condition to meet the inclusion conditions of the preset rehabilitation management model. The preoperative intervention recommendation information includes lifestyle adjustment information, psychological support information, and pre-rehabilitation plan. Among them, the pre-rehabilitation plan can be a preset rehabilitation plan matched by a doctor or expert based on the target patient's physical data. For example, it is possible to establish a corresponding table for a certain type of surgery corresponding to a pre-rehabilitation plan, and match the corresponding pre-rehabilitation plan according to the target patient's surgical type.

[0045] The dynamic data of the adjustment process and the adjustment result data of the target patient adjusting the physical condition based on the preoperative intervention recommendation information are obtained. The corresponding physical physiological indicators or survey indicators (for example, psychological questionnaire survey results, lifestyle survey results, etc.) recorded in real time during the process of adjusting the physical condition of the target patient, for example, weight data, weight change data, test data, diet data, sleep data, etc. recorded during the adjustment of malnutrition, are used to construct dynamic data of the adjustment process. After the target patient adjusts his physical condition to meet the inclusion conditions of the preset rehabilitation management mode, the corresponding adjustment result data is obtained. In this embodiment, the strength of the target patient's own physical function can be characterized by preoperative physical data, dynamic data of the adjustment process, and adjustment result data, which is of great significance for how to more personalizedly execute the corresponding rehabilitation plan during the postoperative rehabilitation process, such as diet, exercise intensity, physical rehabilitation tolerance, psychological acceptance, etc.

[0046] Based on the dynamic data of the adjustment process and the adjustment result data, the recovery ability of the target patient is evaluated to obtain a recovery ability evaluation result. In this embodiment, the recovery ability may include physiological recovery ability and psychological recovery willingness or psychological recovery ability. Different target patients may have similar preoperative physical data, but due to individual differences, their recovery abilities may be different. Therefore, in this embodiment, the recovery ability of the target patient can be evaluated based on the preoperative physical adjustment process and the results obtained after the adjustment. The dynamic data of the adjustment process can characterize the physical adjustment rate, and thus can more accurately understand the recovery ability of the individual target patient.

[0047] Extract the key surgical data from the surgical record data, and determine the difference information of the theoretical surgical data corresponding to the target patient; in this embodiment, a certain type of surgery often has theoretical surgical data. For the description of the surgical data, please refer to the description in the above embodiment, which will not be repeated in this embodiment. In actual surgery, different target patients often have different physical conditions and different surgeons, and the cooperation between doctors and anesthesiologists is different, resulting in differences between the actual key surgical data and the theoretical surgical data of each target patient during the operation. Therefore, in this embodiment, after obtaining the difference information, the difference information can be analyzed to evaluate the relevance of the current difference information to rehabilitation actions, rehabilitation projects and other rehabilitation plans or the importance of the rehabilitation plan. Based on the fixed rehabilitation data, the postoperative physical state of the target patient is evaluated to obtain a postoperative physical state evaluation result; based on the recovery ability evaluation result, the difference information and the postoperative physical state evaluation result, the initial postoperative rehabilitation plan is generated. Through the recovery ability evaluation result, the difference information and the postoperative physical state evaluation result, a comprehensive and personalized customized rehabilitation plan can be made based on the individual situation of the target patient, the surgical process, and the postoperative performance.

[0048] Step S105: obtaining dynamic rehabilitation monitoring data of the target patient during the initial rehabilitation training according to the initial postoperative rehabilitation plan.

[0049] For example, a high-resolution camera or video collector is selected to ensure that the rehabilitation behavior of the target patient can be clearly captured. A static camera or a dynamic camera can be selected as needed. The camera or video collector is arranged in a position that can fully cover the rehabilitation activity area to ensure that the dynamic rehabilitation monitoring data of the target patient during the initial rehabilitation training can be captured.

[0050] In some embodiments, after analyzing the preoperative physical data, the specific rehabilitation data, and the surgical record data to generate an initial postoperative rehabilitation plan, the method further includes: generating auxiliary rehabilitation training instructions, sending the auxiliary rehabilitation training instructions to a rehabilitation robot to instruct the rehabilitation robot to perform video and voice interaction actions according to the auxiliary rehabilitation training instructions to assist the target patient in conducting preliminary rehabilitation training.

[0051] Exemplarily, abnormal identification is performed based on the dynamic rehabilitation monitoring data of the target patient performing the initial postoperative rehabilitation plan to obtain the corresponding abnormal behavior of the target patient. The abnormal behavior is compared with the preset rules. The preset rules may include descriptions of various abnormal movements or postures, such as "incorrect posture" or "excessive exercise intensity". Based on the identified abnormal behaviors and preset rules, write a detailed abnormal description text. These descriptions should clearly and accurately describe the abnormal conditions of the target patients and include specific guidance. For example, "your range of motion is insufficient" or "your posture deviates from the standard."

[0052] For example, a suitable speech synthesis model is selected, which can convert text into speech. Modern speech synthesis technology can generate natural and fluent speech, supporting multiple languages ​​and intonations. The generated abnormal description text is then input into the speech synthesis model. The speech synthesis model will generate corresponding speech information based on the text. The intonation, speed and volume of the speech are adjusted as needed to ensure that the speech information can be clearly conveyed and is suitable for the hearing habits of the target patient.

[0053] Exemplarily, the generated target voice information is played to the target patient through an appropriate device (such as a smartphone, tablet or voice assistant). Ensure that the target patient can clearly hear and understand the voice guidance content. The voice information should include an explanation of abnormal behavior and specific rehabilitation suggestions. For example, the voice can guide the target patient on how to adjust his posture, how to perform the action correctly, or how to adjust the intensity of the exercise. The target patient thus adjusts his motor behavior according to the voice guidance and executes the corresponding rehabilitation plan, and then generates auxiliary rehabilitation training instructions based on the dynamic rehabilitation monitoring data, thereby sending the auxiliary rehabilitation training instructions to the rehabilitation robot to instruct the rehabilitation robot to perform video voice interaction actions according to the rehabilitation training instructions to assist the target patient in preliminary rehabilitation training.

[0054] Specifically, accurate abnormal description texts are generated by combining abnormal behaviors with preset rules, which provides an accurate diagnosis for the problems in the target patients' motor movements. Such diagnoses can accurately identify specific abnormalities such as substandard movement amplitude or incorrect posture. These description texts are not only detailed, but also provide clear feedback to the target patients about their motor defects, thereby greatly improving the accuracy and efficiency of problem solving. The speech synthesis model is then used to convert these abnormal description texts into voice information. This method not only makes the guidance content more natural and fluent, and easy to understand, but also the form of voice guidance is more helpful for the target patients to deeply understand the problem and the corresponding improvement methods. Compared with traditional text guidance, voice guidance is obviously more friendly and easier to be accepted and absorbed by the target patients. In addition, the voice information generated by the speech synthesis model can provide instant feedback to the target patients, so that the target patients can make real-time adjustments according to the actual situation during the rehabilitation training process, thereby avoiding the long-term accumulation of wrong movements. This instant voice guidance method can clearly and specifically guide the target patients on how to correct abnormal behaviors and optimize movements, thereby helping them to complete rehabilitation training more efficiently.

[0055] Step S106: Based on the postoperative management plan of the preset rehabilitation management mode, the rehabilitation progress of the target patient is predicted according to the dynamic rehabilitation monitoring data to generate a rehabilitation progress prediction result.

[0056] Exemplarily, data analysis technology and algorithms are used to identify the movement accuracy of the target patient in the dynamic rehabilitation monitoring data, so that the movement accuracy predicts the rehabilitation progress of the target patient, and then a rehabilitation progress prediction result corresponding to the target patient is generated according to the rehabilitation progress.

[0057] In some embodiments, predicting the rehabilitation progress of the target patient based on the dynamic rehabilitation monitoring data and generating a rehabilitation progress prediction result includes: obtaining first body data corresponding to the target patient before executing the initial postoperative rehabilitation plan; performing abnormality identification on the dynamic rehabilitation monitoring data to obtain abnormal behavior corresponding to the target patient; generating the auxiliary rehabilitation training instructions corresponding to the target patient based on the abnormal behavior; obtaining second body data corresponding to the target patient after performing rehabilitation training under the auxiliary rehabilitation training instructions; determining a matching degree between the initial postoperative rehabilitation plan and the target patient based on the first body data and the second body data; predicting the rehabilitation progress of the target patient based on the matching degree and generating the rehabilitation progress prediction result.

[0058] For example, the first physical data is the corresponding preoperative examination results of the target patient before the operation, for example, the medical history data of the target patient is obtained through a questionnaire survey, an interview or an electronic medical record system. The target patient is then arranged to undergo necessary preoperative examinations, such as imaging examinations (such as X-rays, CT, MRI), laboratory tests (such as blood tests, urine tests) and physiological index measurements (such as blood pressure, heart rate).

[0059] For example, data analysis techniques and algorithms are used to identify abnormal behaviors in dynamic rehabilitation monitoring data. These abnormal behaviors may include deviations in movement, incorrect posture, abnormal exercise intensity, etc. The identified abnormal behaviors are classified, such as irregular execution of movements, too high or too low exercise intensity, insufficient range of motion, etc. Different types of abnormalities may require different intervention measures.

[0060] Exemplarily, the causes of abnormal behavior are analyzed, including technical problems, physical condition problems, external environmental influences, etc. For example, if the target patient's range of motion is insufficient, it may be due to insufficient muscle strength or unclear movement guidance. Therefore, specific auxiliary rehabilitation training instructions are set according to the abnormal behavior and abnormal causes. For example, if it is found that the target patient has incorrect posture during exercise movements, a goal of improving posture can be set, and then corresponding auxiliary rehabilitation training instructions can be designed according to the type of abnormal behavior. Auxiliary rehabilitation training instructions include: movement correction: provide specific movement correction suggestions to help target patients adjust their movement postures, enhanced training: design targeted training programs to enhance the target patients' muscle strength or flexibility, improved technology: provide technical guidance and training to help target patients master correct exercise skills, and adjust intensity: adjust the intensity and frequency of exercise according to the actual situation of the target patient to avoid overtraining or insufficient training.

[0061] Exemplarily, during the implementation of the rehabilitation guidance or after the completion of the rehabilitation guidance, the second body data of the target patient corresponding to the auxiliary rehabilitation training instruction is retrieved.

[0062] Exemplarily, the second body data is used as a baseline, the second body data is compared with the first body data, and the change trend of each physiological index is observed, and then the change trend of each physiological index is integrated to determine the match between the initial postoperative rehabilitation plan and the target patient.

[0063] For example, the difference between the second body data and the first body data is calculated for each physiological indicator, that is, the amount of change before and after the initial postoperative rehabilitation plan is executed, so as to evaluate whether the initial postoperative rehabilitation plan matches the physical response of the target patient according to the change trend of each physiological indicator. When the change trend is positive and the change amplitude is large, the matching value should be set to a large positive value; when the change trend is positive but the change amplitude is small, the matching value should be set to a small positive value; when the change trend is negative and the change amplitude is large, the matching value should be set to a large negative value; when the change trend is negative and the change amplitude is small, the matching value should be set to a small negative value; thus, the matching values ​​of each physiological indicator are added together to obtain the matching degree between the initial postoperative rehabilitation plan and the target patient.

[0064] Exemplarily, based on the matching result, the deficiencies of the initial postoperative rehabilitation plan are identified. If the matching degree is low, it means that the initial postoperative rehabilitation plan does not completely match the physical condition of the target patient, and the rehabilitation progress prediction result is exceeded expectations. If the matching degree is high, it means that the initial postoperative rehabilitation plan matches the physical condition of the target patient, and the rehabilitation progress prediction result is in line with expectations.

[0065] In some embodiments, the performing abnormality identification on the dynamic rehabilitation monitoring data to obtain the abnormal behavior corresponding to the target patient includes: performing keyword extraction according to the initial postoperative rehabilitation plan to obtain the movement movements corresponding to the target patient, and obtaining the body parts associated with the movement movements; performing target identification on the dynamic rehabilitation monitoring data according to a target recognition model to obtain target associated data corresponding to the body parts in the dynamic rehabilitation monitoring data; obtaining the target contours corresponding to the body parts in the dynamic rehabilitation monitoring data, and determining the operation trajectory corresponding to the target patient according to the target contours and the movement movements; performing abnormality identification on the target patient according to the operation trajectory to obtain the abnormal behavior corresponding to the target patient.

[0066] Exemplarily, the named entity model is used to extract keywords from the initial postoperative rehabilitation plan to obtain the corresponding movement movements of the target patient, such as leg lifting, walking, grasping, squatting, shoulder rotation, etc. It should be noted that the named entity model can be trained based on the data corresponding to the annotated movement movements.

[0067] Exemplarily, information query is performed according to the mapping table to obtain the body parts associated with each movement, for example, "leg raising" mainly involves the thigh and knee, and "squatting" involves the whole body.

[0068] Exemplarily, a suitable target recognition model is selected according to the task requirements. For example, commonly used models include YOLO (You Only Look Once), SSD (Single Shot MultiBox Detector), Faster R-CNN, etc. The dynamic rehabilitation monitoring data is input into the target recognition model for inference, and the position of the target part in the dynamic rehabilitation monitoring data (such as the coordinates of the bounding box) is obtained, so that the position of the target part in the dynamic rehabilitation monitoring data is determined as the target association data corresponding to the body part.

[0069] Exemplarily, edge detection is performed on the target-related data using an edge detection algorithm to obtain the target contour corresponding to the body part target-related data. The target contour is matched with the posture corresponding to each stage of the motion action to obtain the target action corresponding to the target contour, and then the motion trajectory of the target patient on the time axis is reconstructed by using the continuous change of the target action and the timestamp information corresponding to the dynamic rehabilitation monitoring data.

[0070] For example, if the movement action is squatting, then splitting the movement action can obtain target patient standing, target patient knee bending (multiple bending degrees), target patient squatting, target patient standing up (multiple standing up degrees), target patient standing, etc., then match the target contour with the posture corresponding to each stage in squatting (target patient standing, target patient knee bending (multiple bending degrees), target patient squatting, target patient standing up (multiple standing up degrees), target patient standing), so as to obtain the different movement stages represented by the target contour, and then use the continuous change of the target contour, combined with the timestamp information corresponding to the dynamic rehabilitation monitoring data, to reconstruct the movement trajectory of the target patient on the timeline. When the target contour is in the target patient squatting stage, it is determined as the peak of the movement trajectory, and when the target contour is in the target patient standing stage, it is determined as the trough of the movement trajectory; when the target contour is in other stages, it is determined as the data between the peak and the trough.

[0071] Exemplarily, key features such as velocity peaks, acceleration changes, curvature of the trajectory, etc. are extracted from the reconstructed running trajectory. An anomaly detection model can be trained using machine learning or statistical learning methods such as support vector machines (SVM), random forests, or anomaly detection algorithms such as isolation forests. This model should be able to distinguish between normal and abnormal trajectories. The extracted trajectory features are input into the anomaly detection model, and the model will output a judgment on whether each trajectory point or action is abnormal. Based on the output of the model, determine which parts of the trajectory or which specific motion actions are identified as abnormal. This can include sudden speed changes, incoherent motion transitions, etc., so as to obtain the corresponding abnormal behavior of the target patient.

[0072] For example, in a motion trajectory, a peak represents the end point of the target patient's motor action. This refers to the highest point or key position in the trajectory during the movement, usually marking the completion or turning point of the movement. A valley represents the starting point of the motor action. That is, the lowest point or key position in the motion trajectory, usually marking the start or recovery phase of the movement. By measuring the time interval between adjacent peaks and valleys, the time it takes for the target patient to perform the motor action can be calculated. Specifically, the length of time between a peak and an adjacent valley represents the duration of a complete motor action. By counting the number of all peaks and adjacent valleys, the frequency of the target patient's motor action can be determined. That is, the total number of times the target patient performs the action within a specific time period. Thus, anomaly detection is performed based on the calculated movement time and frequency. For example: if the movement time or frequency deviates from the expected standard, it may indicate that the target patient's action does not meet the requirements. If the duration of a certain action is abnormally long, it may indicate that the target patient has difficulties or errors in performing the action. If the frequency of performing an action is lower than the expected standard, it may indicate that the target patient has problems in completing the motor task. Thus, by analyzing the anomaly detection results, the target patient's abnormal behavior can be identified. These abnormal behaviors may include substandard movement quality, excessive exercise time, or insufficient frequency.

[0073] For example, by analyzing the running trajectory and target contour, abnormal behaviors that occur during exercise can be accurately detected. Abnormal identification based on the running trajectory can provide an objective basis, avoid relying on subjective judgment, and improve the accuracy of abnormal detection. Abnormal identification can promptly detect risky behaviors or irregular movements that may occur in the target patient during exercise, thereby preventing potential sports injuries and improving the safety of the target patient. Furthermore, by continuously monitoring the target patient's movement trajectory and analyzing abnormal behaviors, the rehabilitation plan can be continuously optimized. Adjust the exercise intensity, frequency or movement details according to the actual performance of the target patient to make the rehabilitation plan more personalized and targeted.

[0074] Specifically, by quantifying the target profile into a running trajectory according to the movement, the actual performance of the target patient can be obtained more accurately, and the shortcomings of the target patient in performing the movement and the abnormalities of the movement intensity, frequency or movement details can be discovered in time. The quantified running trajectory can accurately record the path, speed and time of each movement of the target patient, thereby comprehensively reflecting the performance of the target patient in the movement, and then using objective standards to evaluate the movement quality of the target patient, rather than relying solely on subjective judgment. This helps to improve the accuracy and reliability of the evaluation. By analyzing the running trajectory, the shortcomings of the target patient in performing the movement can be discovered early. The running trajectory can reveal abnormalities in the intensity and frequency of movement. For example, if the trajectory shows that the intensity of the movement is too high or too low, or the frequency of the movement does not meet expectations, this information can help adjust the exercise plan to avoid overtraining or undertraining. The refined trajectory data helps to check the details of the movement, such as whether the movement amplitude meets the standard and whether the movement is coherent. Therefore, quantifying the target profile into a running trajectory can not only obtain the actual movement performance of the target patient more accurately, but also discover the shortcomings in the movement in time, optimize the rehabilitation plan, improve safety, and provide strong support for data-driven decision-making, which is of great significance for the subsequent improvement of the effect and efficiency of sports rehabilitation.

[0075] In some embodiments, determining the running trajectory corresponding to the target patient according to the target contour and the motion action includes: determining the motion direction corresponding to the target patient according to the motion action; when the motion direction is up and down motion, obtaining the user width and user height corresponding to the target contour, and determining the user ratio corresponding to the target patient according to the user width and the user height; obtaining the first time corresponding to the target contour from the target-associated data, and determining the running trajectory corresponding to the target patient according to the first time and the user ratio; when the motion direction is left and right motion, obtaining the area ratio corresponding to the target contour in the target-associated data; obtaining the second time corresponding to the target contour from the target-associated data, and determining the running trajectory corresponding to the target patient according to the second time and the area ratio.

[0076] Exemplarily, when the target patient performs various posture exercises in a fixed position, the corresponding movement direction of the target patient can be determined to be up and down movement; when the target patient exercises in different positions but with the same body posture, the corresponding movement direction of the target patient can be determined to be left and right movement. For example, when the movement action is squatting, the movement direction is up and down movement, and when the movement action is walking, the movement direction is left and right movement.

[0077] Exemplarily, when the motion direction is up and down motion, the user width and user height corresponding to the target patient in the target outline are obtained, so that the ratio between the user width and the user height is determined as the user ratio corresponding to the target patient. Then, the first time corresponding to the target outline is obtained from the target association data, so that the first time is determined as the horizontal axis and the user ratio is determined as the vertical axis, and the running trajectory corresponding to the target patient is determined according to the change trajectory of the user ratio at the first time.

[0078] First, the user width and user height of the target patient are extracted from the target contour. The user width usually refers to the horizontal size of the body parts of the target patient in the target associated data, while the user height usually refers to the vertical size of the body parts of the target patient in the target associated data.

[0079] Exemplarily, the user proportion is calculated based on the ratio between the user width and the user height. This ratio is the ratio of the body width to the body height of the target patient at a specific time point. The user proportion of the target patient is used to reflect the stage of the target patient's motion execution, and then the first time corresponding to the target contour is extracted from the target-related data.

[0080] Exemplarily, the first time is set as the horizontal axis. This represents the change in time, that is, the evolution of the action or posture. The calculated user ratio is set as the vertical axis. This represents the ratio of the width to the height of the user's body, that is, the stage of the target patient's motor action execution, so that by drawing a relationship graph between the user ratio and time, the running trajectory of the target patient during the execution of the motor action can be obtained. Specifically, this trajectory shows the change in the stage of the target patient's motor action execution at the first time.

[0081] For example, when the exercise action is squatting, the user ratio will change according to the target patient's various stages of squatting, so by drawing a relationship graph between the user ratio and the first time, the running trajectory of the target patient during the squatting process can be obtained.

[0082] Exemplarily, when the direction of motion is left-right motion, valuable information about the user's motion can be obtained by analyzing the area ratio of the target contour in the target-associated data. Since the direction of motion is left-right motion, the data size of the target patient in the target-associated data will gradually change as the motion progresses. For example, when the motion action is walking, as the target patient moves further and further forward during the motion, the area ratio of the target contour in the target-associated data will gradually decrease. This is because as the target patient moves, the relative position of the target contour and other elements in the data frame changes, causing the data area it occupies to gradually shrink.

[0083] Exemplarily, the second time corresponding to the target contour is obtained from the target association data, and the second time is set as the horizontal axis. This represents the change in time, that is, the evolution of the action or posture. The calculated area ratio is set as the vertical axis. Thus, by drawing a relationship graph between the area ratio and the second time, the running trajectory of the target patient during the execution of the movement action can be obtained. Specifically, this trajectory shows the changes in the execution stage of the movement action of the target patient at the second time.

[0084] For example, when the movement action is walking, the area proportion will change according to the target patient at each stage of walking, so by drawing a relationship graph between the area proportion and the second time, the running trajectory of the target patient during the walking process can be obtained.

[0085] Specifically, determining the specific movement direction (up and down or left and right) of the target patient helps to further analyze the user's physical reaction and dynamic performance when performing specific actions. In the case of up and down movement, by obtaining the user width and user height and calculating their ratio (i.e., user proportion), the stability of the user's body posture and movement can be understood. In the case of left and right movement, by obtaining the area proportion, the size of the space occupied by the user in the lateral movement can be tracked, which helps to analyze the user's lateral movement efficiency and body coordination. Combined with the first time and user proportion, the running trajectory of the user in the up and down movement can be drawn to intuitively display the user's movement changes in the vertical direction. Combined with the second time and area proportion, the running trajectory of the user in the left and right movement can be analyzed, and the dynamic performance of the user in the lateral movement can be observed. Therefore, by quantitatively analyzing the behavioral performance of the target patient at a specified time point, the accuracy and objectivity of the evaluation can be improved, which helps medical staff understand the target patient's movement situation and improvement strategy, and can also help the target patient understand his own movement status and effect. Therefore, through real-time monitoring and analysis of the running trajectory, immediate feedback can be provided for the target patient's movement, enabling the user to adjust the movement and exercise plan in time, avoid sports injuries and irregular movements, thereby improving exercise efficiency and rehabilitation effects.

[0086] In some embodiments, the first body data and the second body data are collected at the same time, and the determining of the matching degree between the initial postoperative rehabilitation plan and the target patient based on the first body data and the second body data includes: determining the body index corresponding to the target patient, obtaining the first data corresponding to the body index at the target time from the first body data and obtaining the second data corresponding to the body index at the target time from the second body data; determining the first state corresponding to the body index at the target time according to the first data and the preset data, and determining the second state corresponding to the body index at the target time according to the second data and the preset data; fusing the first state and the second state corresponding to the body index at the target time to obtain the state similarity between the first body data and the second body data; determining the first conformity between the initial postoperative rehabilitation plan and the target patient according to the state similarity; calculating the data gap between the first body data and the second body data, and determining the second conformity between the initial postoperative rehabilitation plan and the target patient according to the data gap; fusing the first conformity and the second conformity to determine the matching degree between the initial postoperative rehabilitation plan and the target patient; wherein the state similarity is obtained according to the following formula: ; in, represents the state similarity, N represents the total number of moments corresponding to the target moment, m represents the total number of indicators corresponding to the physical indicators, represents the value corresponding to the first state under the i-th physical indicator at the t-th target moment, represents the value corresponding to the second state of the i-th physical indicator at the t-th target moment, represents the value corresponding to the first state under the jth physical indicator at the tth target moment, Represents the value corresponding to the second state under the kth physical indicator at the tth target moment.

[0087] Exemplarily, the first body data and the second body data are collected at the same time, that is, the second body data and the first body data are collected at the same time on different dates. For example, the first body data is collected at 12:00 on July 1, and the second body data is collected at 12:00 on July 9.

[0088] Exemplarily, the physical indicators corresponding to the target patient are determined based on historical experience or expert experience, so as to obtain first data corresponding to the physical indicators at the target time from the first physical data and obtain second data corresponding to the physical indicators at the target time from the second physical data.

[0089] Exemplarily, the first data is compared with the preset data to determine the first state of the physical indicator at the target time. For example, if the preset data is set to a normal range, the first state is determined to be good if the first data is within the normal range; if the first data exceeds the normal range, the first state is determined to be poor.

[0090] Exemplarily, the second data is compared with the same preset data to determine the second state of the body index at the target time. Similarly, if the second data is within a normal range, the second state is determined to be better; if it is outside the normal range, the second state is determined to be worse.

[0091] Exemplarily, the first state and the second state corresponding to the physical index at the target moment are fused according to the following formula to obtain the state similarity corresponding to the first physical data and the second physical data: ; in, represents the state similarity, N represents the total number of moments corresponding to the target moment, and m represents the total number of indicators corresponding to the physical indicators. represents the value corresponding to the first state of the ith physical indicator at the tth target moment, represents the value corresponding to the second state of the ith physical indicator at the tth target moment, represents the value corresponding to the first state of the j-th physical indicator at the t-th target moment, Indicates the value corresponding to the second state of the kth body indicator at the tth target moment.

[0092] For example, if the preset data is set to a normal range, the first state is determined to be good if the first data is within the normal range, and the value corresponding to the first state is 1; if the first data exceeds the normal range, the first state is determined to be poor, and the value corresponding to the first state is 0. Similarly, if the second data is within the normal range, the second state is determined to be good, and the value corresponding to the second state is 1; if it exceeds the normal range, the second state is determined to be poor, and the value corresponding to the second state is 0.

[0093] For example, by fusing the first state and the second state of all physical indicators at the target time through the above formula, the physical state of the target patient can be comprehensively evaluated. This comprehensive analysis helps to more accurately understand the changes in physical indicators at different time points.

[0094] Exemplarily, when the state similarity is small, the matching degree between the initial postoperative rehabilitation plan and the target patient is greater, and when the state similarity is large, the matching degree between the initial postoperative rehabilitation plan and the target patient is smaller. Thus, the state similarity is normalized to between 0 and 1, and then 1 is subtracted from the normalized state similarity to determine the first conformity between the initial postoperative rehabilitation plan and the target patient.

[0095] Exemplarily, the data gap between the first body data and the second body data is calculated, and this step is intended to quantify the physical changes of the user after undergoing the rehabilitation strategy. The specific operation includes comparing the values ​​of the same physical indicators in the two sets of data and calculating the difference between them. This difference analysis helps to reveal the actual impact of the rehabilitation strategy on the user's body. After obtaining the data gap, the next step is to analyze the impact trend reflected by these data gaps. By comparing the first body data with the second body data, it can be determined whether the rehabilitation strategy has a positive impact on the user, that is, whether the physical indicators have improved in the expected direction, or whether a negative impact has occurred, that is, the physical indicators have not improved or even worsened. Based on the data gap and the impact trend revealed by it, the second conformity between the initial postoperative rehabilitation plan and the target patient can be further evaluated. If the data gap shows that the physical indicators have significantly improved and are consistent with the rehabilitation goals, then it can be considered that the rehabilitation strategy has a high degree of conformity with the user. On the contrary, if the data gap shows that the physical indicators have not improved significantly or are contrary to the rehabilitation goals, then the rehabilitation strategy has a low degree of conformity with the user.

[0096] Exemplarily, the first degree of conformity (based on state similarity) and the second degree of conformity (based on data gap) are fused to obtain the final degree of conformity. This can be done by taking into account the two degrees of conformity through weighted average, scoring system, etc., to obtain the degree of conformity between the initial postoperative rehabilitation plan and the target patient.

[0097] Specifically, fusing the first state and the second state and calculating the state similarity helps to comprehensively compare the health status of two different data sets at the same time point, thereby obtaining a more accurate assessment of the state of physical indicators. The effectiveness of the initial postoperative rehabilitation plan is assessed by state similarity. This helps to understand whether the rehabilitation strategy is adapted to the user's actual health status and whether it needs to be adjusted. By calculating the data gap, the actual impact of the rehabilitation strategy is evaluated. The first degree of conformity and the second degree of conformity are combined to form a comprehensive match. This comprehensive evaluation can provide more comprehensive feedback on the effectiveness of the rehabilitation strategy and ensure that the strategy can meet the user's health needs to the greatest extent.

[0098] Step S107, adjusting the initial postoperative rehabilitation plan according to the rehabilitation progress prediction result to obtain a target postoperative rehabilitation plan, wherein the target postoperative rehabilitation plan is used to guide the target patient to perform target rehabilitation training until the target patient meets the discharge conditions of the preset rehabilitation management model.

[0099] Exemplarily, when the predicted result of rehabilitation progress exceeds expectations, the parts of the initial postoperative rehabilitation plan that perform well are strengthened, such as increasing the frequency or increasing the intensity. Improve the parts of the initial postoperative rehabilitation plan that do not work well, such as adjusting the treatment methods or changing sports. Integrate the adjusted parts into a new postoperative rehabilitation plan, namely the target postoperative rehabilitation plan. The target postoperative rehabilitation plan is used to guide the target patient to conduct target rehabilitation training until the target patient meets the discharge conditions of the preset rehabilitation management model. Therefore, through this method, artificial intelligence can provide important support in the perioperative rehabilitation process, improve rehabilitation effects, shorten rehabilitation time, and at the same time reduce the workload of medical staff, and can promote and popularize remote, multi-center, and homogeneous preset rehabilitation technologies.

[0100] See also Figure 2 , Figure 2An artificial intelligence-based user rehabilitation training assistance system 200 is provided in an embodiment of the present application. The artificial intelligence-based user rehabilitation training assistance system 200 includes a first acquisition module 201, a second acquisition module 202, a third acquisition module 203, a plan generation module 204, a rehabilitation monitoring module 205, a progress prediction module 206, and a plan adjustment module 207. The first acquisition module 201 is used to obtain the preoperative physical data corresponding to the target patient before the operation, wherein the preoperative physical data includes health data, medical history and preoperative examination results; the second acquisition module 202 is used to obtain the surgical record data corresponding to the target patient during the operation, and the surgical record data is the data obtained after the target patient adjusts his physical condition to meet the inclusion conditions of the preset rehabilitation management mode and excludes contraindications for the operation; the third acquisition module 203 is used to obtain the targeted rehabilitation data of the target patient after the operation; wherein the targeted rehabilitation data includes Movement status, pain score and functional recovery; a plan generation module 204, used to analyze the preoperative physical data, the fixed rehabilitation data and the surgical record data to generate an initial postoperative rehabilitation plan; the initial postoperative rehabilitation plan is used to guide the target patient to perform preliminary rehabilitation training; a rehabilitation monitoring module 205, used to obtain the dynamic rehabilitation monitoring data of the target patient during the preliminary rehabilitation training according to the initial postoperative rehabilitation plan; a progress prediction module 206, used to predict the rehabilitation progress of the target patient based on the postoperative management plan of the preset rehabilitation management mode according to the dynamic rehabilitation monitoring data, and generate a rehabilitation progress prediction result; a plan adjustment module 207, used to adjust the initial postoperative rehabilitation plan according to the rehabilitation progress prediction result to obtain a target postoperative rehabilitation plan, and the target postoperative rehabilitation plan is used to guide the target patient to perform target rehabilitation training until the target patient meets the discharge conditions of the preset rehabilitation management mode.

[0101] In some embodiments, after analyzing the preoperative physical data, the specific rehabilitation data, and the surgical record data to generate the initial postoperative rehabilitation plan, the plan generation module 204 further performs: Generate auxiliary rehabilitation training instructions, and send the auxiliary rehabilitation training instructions to the rehabilitation robot to instruct the rehabilitation robot to perform video and voice interaction actions according to the auxiliary rehabilitation training instructions to assist the target patient in performing preliminary rehabilitation training.

[0102] In some implementations, before obtaining the surgical record data corresponding to the target patient during surgery, the second acquisition module 202 further performs: Using a risk prediction model to perform risk prediction on the preoperative physical data to obtain a perioperative risk prediction result; Preoperative intervention recommendation information is generated based on the perioperative risk prediction results, and the preoperative intervention recommendation information is used to guide the target patient to adjust his physical condition to meet the inclusion conditions of a preset rehabilitation management model, wherein the preoperative intervention recommendation information includes lifestyle adjustment information, psychological support information, and a pre-rehabilitation plan.

[0103] In some implementations, the progress prediction module 206 performs, during the process of predicting the rehabilitation progress of the target patient according to the dynamic rehabilitation monitoring data and generating a rehabilitation progress prediction result: obtaining first physical data corresponding to the target patient before the initial postoperative rehabilitation plan is performed; Performing abnormality identification on the dynamic rehabilitation monitoring data to obtain abnormal behavior corresponding to the target patient; generating the auxiliary rehabilitation training instruction corresponding to the target patient according to the abnormal behavior; Obtaining second body data corresponding to the target patient after the target patient performs rehabilitation training under the auxiliary rehabilitation training instruction; Determining a matching degree between the initial postoperative rehabilitation plan and the target patient according to the first body data and the second body data; The rehabilitation progress of the target patient is predicted according to the matching degree, and the rehabilitation progress prediction result is generated.

[0104] In some implementations, the progress prediction module 206 performs, during the process of identifying abnormalities in the dynamic rehabilitation monitoring data and obtaining abnormal behaviors corresponding to the target patient: Extract keywords according to the initial postoperative rehabilitation plan to obtain the motor movements corresponding to the target patient and obtain the body parts associated with the motor movements; Performing target recognition on the dynamic rehabilitation monitoring data according to a target recognition model to obtain target association data corresponding to the body part in the dynamic rehabilitation monitoring data; Obtaining a target contour corresponding to the body part in the dynamic rehabilitation monitoring data, and determining a running trajectory corresponding to the target patient according to the target contour and the movement action; The target patient is identified for abnormality according to the running trajectory to obtain the abnormal behavior corresponding to the target patient.

[0105] In some implementations, the progress prediction module 206, during the process of determining the running trajectory corresponding to the target patient according to the target profile and the motion action, performs: Determining the movement direction corresponding to the target patient according to the movement action; When the movement direction is up and down movement, the user width and user height corresponding to the target outline are obtained, and the user ratio corresponding to the target patient is determined according to the user width and the user height; Obtaining a first time corresponding to the target profile from the target association data, and determining the running trajectory corresponding to the target patient according to the first time and the user ratio; When the movement direction is left-right movement, the area proportion corresponding to the target contour in the target association data is obtained; A second time corresponding to the target contour is obtained from the target association data, and the running trajectory corresponding to the target patient is determined according to the second time and the area proportion.

[0106] In some embodiments, the first body data and the second body data are collected at the same time, and the progress prediction module 206 performs, in the process of determining the matching degree between the initial postoperative rehabilitation plan and the target patient according to the first body data and the second body data: Determine a physical index corresponding to the target patient, obtain first data corresponding to the physical index at a target time from the first physical data, and obtain second data corresponding to the physical index at the target time from the second physical data; Determine a first state corresponding to the body indicator at the target time according to the first data and the preset data, and determine a second state corresponding to the body indicator at the target time according to the second data and the preset data; fusing the first state and the second state corresponding to the physical indicator at the target moment to obtain a state similarity corresponding to the first physical data and the second physical data; Determining a first degree of compliance between the initial postoperative rehabilitation plan and the target patient according to the state similarity; calculating a data gap between the first body data and the second body data, and determining a second degree of compliance between the initial postoperative rehabilitation plan and the target patient according to the data gap; The first compliance degree and the second compliance degree are integrated to determine the matching degree between the initial postoperative rehabilitation plan and the target patient; The state similarity is obtained according to the following formula: ; in, represents the state similarity, N represents the total number of moments corresponding to the target moment, m represents the total number of indicators corresponding to the physical indicators, represents the value corresponding to the first state of the i-th physical indicator at the t-th target moment, represents the value corresponding to the second state of the i-th physical indicator at the t-th target moment, represents the value corresponding to the first state under the jth physical indicator at the tth target moment, Represents the value corresponding to the second state under the kth physical indicator at the tth target moment.

[0107] In some embodiments, the artificial intelligence-based user rehabilitation training assistance system 200 can be applied to a terminal device.

[0108] It should be noted that technical personnel in the relevant field can clearly understand that, for the convenience and conciseness of description, the specific working process of the user rehabilitation training assistance system 200 based on artificial intelligence described above can refer to the corresponding process in the aforementioned artificial intelligence-based rehabilitation training assistance method embodiment, and will not be repeated here.

[0109] See also Figure 3 , Figure 3 A schematic block diagram of the structure of a terminal device provided in an embodiment of the present invention.

[0110] like Figure 3 As shown, the terminal device 300 includes a processor 301 and a memory 302 , and the processor 301 and the memory 302 are connected via a bus 303 , such as an I2C (Inter-integrated Circuit) bus.

[0111] Specifically, the processor 301 is used to provide computing and control capabilities to support the operation of the entire terminal device. The processor 301 can be a central processing unit (CPU), and the processor 301 can also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0112] Specifically, the memory 302 may be a Flash chip, a read-only memory (ROM) disk, an optical disk, a USB flash drive, or a mobile hard disk.

[0113] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a partial structure related to the embodiment of the present invention, and does not constitute a limitation on the terminal device to which the embodiment of the present invention is applied. The specific server may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0114] The processor is used to run a computer program stored in the memory, and implement any one of the artificial intelligence-based rehabilitation training assistance methods provided in the embodiments of the present invention when executing the computer program.

[0115] In one embodiment, the processor is used to run a computer program stored in the memory, and implements the following steps when executing the computer program: Obtaining preoperative physical data corresponding to the target patient before surgery, wherein the preoperative physical data includes health data, medical history, and preoperative examination results; Obtaining surgical record data corresponding to the target patient during surgery, wherein the surgical record data is surgical process data obtained during the process of adjusting the target patient's physical condition to meet the inclusion conditions of the preset rehabilitation management model and excluding contraindications for surgery; Obtaining the targeted rehabilitation data of the target patient after surgery; wherein the targeted rehabilitation data includes movement status, pain score and functional recovery; Analyze the preoperative physical data, the fixed rehabilitation data and the surgical record data to generate an initial postoperative rehabilitation plan; the initial postoperative rehabilitation plan is used to guide the target patient to perform preliminary rehabilitation training; Acquiring dynamic rehabilitation monitoring data of the target patient during the initial rehabilitation training according to the initial postoperative rehabilitation plan; Based on the postoperative management plan of the preset rehabilitation management mode, predicting the rehabilitation progress of the target patient according to the dynamic rehabilitation monitoring data, and generating a rehabilitation progress prediction result; The initial postoperative rehabilitation plan is adjusted according to the rehabilitation progress prediction result to obtain a target postoperative rehabilitation plan, and the target postoperative rehabilitation plan is used to guide the target patient to perform target rehabilitation training until the target patient meets the discharge conditions of the preset rehabilitation management model.

[0116] In some embodiments, after the processor 301 analyzes the preoperative physical data, the specific rehabilitation data, and the surgical record data to generate an initial postoperative rehabilitation plan, it further executes: Determine preoperative intervention suggestion information based on preoperative physical data, wherein the preoperative intervention suggestion information is used to guide the target patient to adjust the physical condition to meet the inclusion conditions of the preset rehabilitation management model, and the preoperative intervention suggestion information includes lifestyle adjustment information, psychological support information, and pre-rehabilitation plan; Acquiring dynamic data of the adjustment process and adjustment result data of the target patient adjusting the physical condition based on the preoperative intervention recommendation information; Performing a recovery ability assessment on the target patient based on the adjustment process dynamic data and the adjustment result data to obtain a recovery ability assessment result; Extracting key surgical data from the surgical record data, and determining difference information of theoretical surgical data corresponding to the target patient; Evaluate the postoperative physical condition of the target patient based on the fixed-item rehabilitation data to obtain a postoperative physical condition evaluation result; The initial postoperative rehabilitation plan is generated based on the recovery ability assessment result, the difference information and the postoperative physical state assessment result.

[0117] In some implementations, the processor 301, before obtaining the surgical record data corresponding to the target patient during the surgery, further executes: The risk prediction model is used to perform risk prediction on the preoperative physical data to obtain a perioperative risk prediction result. The risk prediction model includes a relationship recognition layer, a feature representation layer, a relationship fusion layer and a risk prediction layer, wherein the relationship recognition layer is used to identify the target relationship between various factors in the preoperative physical data; the feature representation layer is used to convert the preoperative physical data into an initial feature vector; the relationship fusion layer is used to fuse the initial feature vector in combination with the target relationship to obtain a target feature vector; the risk prediction layer is used to perform risk assessment and prediction through the target feature vector to obtain a perioperative risk prediction result; Preoperative intervention recommendation information is generated based on the perioperative risk prediction results.

[0118] In some implementations, during the process of predicting the rehabilitation progress of the target patient according to the dynamic rehabilitation monitoring data and generating a rehabilitation progress prediction result, the processor 301 executes: obtaining first physical data corresponding to the target patient before executing the initial postoperative rehabilitation plan; Performing abnormality identification on the dynamic rehabilitation monitoring data to obtain abnormal behavior corresponding to the target patient; generating the auxiliary rehabilitation training instruction corresponding to the target patient according to the abnormal behavior; Obtaining second body data corresponding to the target patient after the target patient performs rehabilitation training under the auxiliary rehabilitation training instruction; Determining a matching degree between the initial postoperative rehabilitation plan and the target patient according to the first body data and the second body data; The rehabilitation progress of the target patient is predicted according to the matching degree, and the rehabilitation progress prediction result is generated.

[0119] In some implementations, the processor 301, during the process of identifying abnormalities in the dynamic rehabilitation monitoring data and obtaining abnormal behaviors corresponding to the target patient, executes: Extract keywords according to the initial postoperative rehabilitation plan to obtain the motor movements corresponding to the target patient and obtain the body parts associated with the motor movements; Performing target recognition on the dynamic rehabilitation monitoring data according to a target recognition model to obtain target association data corresponding to the body part in the dynamic rehabilitation monitoring data; Obtaining a target contour corresponding to the body part in the dynamic rehabilitation monitoring data, and determining a running trajectory corresponding to the target patient according to the target contour and the movement action; The target patient is identified for abnormality according to the running trajectory to obtain the abnormal behavior corresponding to the target patient.

[0120] In some implementations, during the process of determining the running trajectory corresponding to the target patient according to the target profile and the motion action, the processor 301 executes: Determining the movement direction corresponding to the target patient according to the movement action; When the movement direction is up and down movement, the user width and user height corresponding to the target outline are obtained, and the user ratio corresponding to the target patient is determined according to the user width and the user height; Obtaining a first time corresponding to the target profile from the target association data, and determining the running trajectory corresponding to the target patient according to the first time and the user ratio; When the movement direction is left-right movement, the area proportion corresponding to the target contour in the target association data is obtained; A second time corresponding to the target contour is obtained from the target association data, and the running trajectory corresponding to the target patient is determined according to the second time and the area proportion.

[0121] In some embodiments, the first body data and the second body data are collected at the same time, and the processor 301, in the process of determining the matching degree between the initial postoperative rehabilitation plan and the target patient according to the first body data and the second body data, executes: Determine a physical index corresponding to the target patient, obtain first data corresponding to the physical index at a target time from the first physical data, and obtain second data corresponding to the physical index at the target time from the second physical data; Determine a first state corresponding to the body indicator at the target time according to the first data and the preset data, and determine a second state corresponding to the body indicator at the target time according to the second data and the preset data; fusing the first state and the second state corresponding to the physical indicator at the target moment to obtain a state similarity corresponding to the first physical data and the second physical data; Determining a first degree of compliance between the initial postoperative rehabilitation plan and the target patient according to the state similarity; calculating a data gap between the first body data and the second body data, and determining a second degree of compliance between the initial postoperative rehabilitation plan and the target patient according to the data gap; The first compliance degree and the second compliance degree are integrated to determine the matching degree between the initial postoperative rehabilitation plan and the target patient; The state similarity is obtained according to the following formula: ; in, represents the state similarity, N represents the total number of moments corresponding to the target moment, m represents the total number of indicators corresponding to the physical indicators, represents the value corresponding to the first state under the i-th physical indicator at the t-th target moment, represents the value corresponding to the second state of the i-th physical indicator at the t-th target moment, represents the value corresponding to the first state under the jth physical indicator at the tth target moment, Represents the value corresponding to the second state under the kth physical indicator at the tth target moment.

[0122] It should be noted that technicians in the relevant field can clearly understand that for the convenience and brevity of description, the specific working process of the terminal device described above can refer to the corresponding process in the aforementioned artificial intelligence-based rehabilitation training assistance method embodiment, and will not be repeated here.

[0123] An embodiment of the present invention also provides a storage medium for computer-readable storage, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of any one of the artificial intelligence-based rehabilitation training assistance methods provided in the description of the embodiment of the present invention.

[0124] The storage medium may be an internal storage unit of the terminal device described in the foregoing embodiment, such as a hard disk or memory of the terminal device. The storage medium may also be an external storage device of the terminal device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc., equipped on the terminal device.

[0125] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware embodiment, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transient medium). As is known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0126] It should be understood that the term "and / or" used in the present specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations. It should be noted that, in this article, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "including a..." does not exclude the presence of other identical elements in the process, method, article or system including the element.

[0127] The serial numbers of the embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments. The above description is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A rehabilitation training auxiliary method based on artificial intelligence, characterized in that: The method comprises: Obtaining preoperative physical data corresponding to the target patient before surgery, wherein the preoperative physical data includes health data, medical history, and preoperative examination results; Obtaining surgical record data corresponding to the target patient during surgery, wherein the surgical record data is surgical process data obtained during the process of adjusting the target patient's physical condition to meet the inclusion conditions of the preset rehabilitation management model and excluding contraindications for surgery; Obtaining the targeted rehabilitation data of the target patient after surgery; wherein the targeted rehabilitation data includes movement status, pain score and functional recovery; Analyze the preoperative physical data, the fixed rehabilitation data and the surgical record data to generate an initial postoperative rehabilitation plan; the initial postoperative rehabilitation plan is used to guide the target patient to perform preliminary rehabilitation training; Acquiring dynamic rehabilitation monitoring data of the target patient during the initial rehabilitation training according to the initial postoperative rehabilitation plan; Based on the postoperative management plan of the preset rehabilitation management mode, predicting the rehabilitation progress of the target patient according to the dynamic rehabilitation monitoring data, and generating a rehabilitation progress prediction result; The initial postoperative rehabilitation plan is adjusted according to the rehabilitation progress prediction result to obtain a target postoperative rehabilitation plan, and the target postoperative rehabilitation plan is used to guide the target patient to perform target rehabilitation training until the target patient meets the discharge conditions of the preset rehabilitation management model.

2. The method according to claim 1, characterized in that: The generating of an initial postoperative rehabilitation plan by analyzing the preoperative physical data, the specific rehabilitation data and the surgical record data comprises: Determine preoperative intervention suggestion information based on preoperative physical data, wherein the preoperative intervention suggestion information is used to guide the target patient to adjust the physical condition to meet the inclusion conditions of the preset rehabilitation management model, and the preoperative intervention suggestion information includes lifestyle adjustment information, psychological support information, and pre-rehabilitation plan; Acquiring dynamic data of the adjustment process and adjustment result data of the target patient adjusting the physical condition based on the preoperative intervention recommendation information; Performing a recovery ability assessment on the target patient based on the adjustment process dynamic data and the adjustment result data to obtain a recovery ability assessment result; Extracting key surgical data from the surgical record data, and determining difference information of theoretical surgical data corresponding to the target patient; Evaluate the postoperative physical condition of the target patient based on the fixed-item rehabilitation data to obtain a postoperative physical condition evaluation result; The initial postoperative rehabilitation plan is generated based on the recovery ability assessment result, the difference information and the postoperative physical state assessment result.

3. The method according to claim 2, characterized in that The determining of preoperative intervention suggestion information based on preoperative physical data includes: The risk prediction model is used to perform risk prediction on the preoperative physical data to obtain a perioperative risk prediction result. The risk prediction model includes a relationship recognition layer, a feature representation layer, a relationship fusion layer and a risk prediction layer, wherein the relationship recognition layer is used to identify the target relationship between various factors in the preoperative physical data; the feature representation layer is used to convert the preoperative physical data into an initial feature vector; the relationship fusion layer is used to fuse the initial feature vector in combination with the target relationship to obtain a target feature vector; the risk prediction layer is used to perform risk assessment and prediction through the target feature vector to obtain a perioperative risk prediction result; Preoperative intervention recommendation information is generated based on the perioperative risk prediction results.

4. The method according to claim 2, characterized in that: The step of predicting the target patient's rehabilitation progress according to the dynamic rehabilitation monitoring data and generating a rehabilitation progress prediction result includes: obtaining first physical data corresponding to the target patient before executing the initial postoperative rehabilitation plan; Performing abnormality identification on the dynamic rehabilitation monitoring data to obtain abnormal behavior corresponding to the target patient; generating an auxiliary rehabilitation training instruction corresponding to the target patient according to the abnormal behavior; Obtaining second body data corresponding to the target patient after the target patient performs rehabilitation training under the auxiliary rehabilitation training instruction; Determining a matching degree between the initial postoperative rehabilitation plan and the target patient according to the first body data and the second body data; The rehabilitation progress of the target patient is predicted according to the matching degree, and the rehabilitation progress prediction result is generated.

5. The method according to claim 4, characterized in that The performing abnormal identification on the dynamic rehabilitation monitoring data to obtain abnormal behavior corresponding to the target patient includes: Extract keywords according to the initial postoperative rehabilitation plan to obtain the motor movements corresponding to the target patient and obtain the body parts associated with the motor movements; Performing target recognition on the dynamic rehabilitation monitoring data according to a target recognition model to obtain target association data corresponding to the body part in the dynamic rehabilitation monitoring data; Obtaining a target contour corresponding to the body part in the dynamic rehabilitation monitoring data, and determining a running trajectory corresponding to the target patient according to the target contour and the movement action; The target patient is identified for abnormality according to the running trajectory to obtain the abnormal behavior corresponding to the target patient.

6. The method according to claim 5, characterized in that The step of determining the movement trajectory corresponding to the target patient according to the target contour and the movement action includes: Determining the movement direction corresponding to the target patient according to the movement action; When the movement direction is up and down movement, the target patient width and the target patient height corresponding to the target contour are obtained, and the target patient proportion corresponding to the target patient is determined according to the target patient width and the target patient height; Obtaining a first time corresponding to the target profile from the target association data, and determining the running trajectory corresponding to the target patient according to the first time and the target patient ratio; When the movement direction is left-right movement, the area proportion corresponding to the target contour in the target association data is obtained; A second time corresponding to the target contour is obtained from the target association data, and the running trajectory corresponding to the target patient is determined according to the second time and the area proportion.

7. The method according to claim 4, characterized in that The first body data and the second body data are collected at the same time, and determining the matching degree between the initial postoperative rehabilitation plan and the target patient according to the first body data and the second body data includes: Determine a physical index corresponding to the target patient, obtain first data corresponding to the physical index at a target time from the first physical data, and obtain second data corresponding to the physical index at the target time from the second physical data; Determine a first state corresponding to the body indicator at the target time according to the first data and the preset data, and determine a second state corresponding to the body indicator at the target time according to the second data and the preset data; fusing the first state and the second state corresponding to the physical indicator at the target moment to obtain a state similarity corresponding to the first physical data and the second physical data; Determining a first degree of compliance between the initial postoperative rehabilitation plan and the target patient according to the state similarity; calculating a data gap between the first body data and the second body data, and determining a second degree of compliance between the initial postoperative rehabilitation plan and the target patient according to the data gap; The first compliance degree and the second compliance degree are integrated to determine the matching degree between the initial postoperative rehabilitation plan and the target patient; The state similarity is obtained according to the following formula: ； in, represents the state similarity, N represents the total number of moments corresponding to the target moment, m represents the total number of indicators corresponding to the physical indicators, represents the value corresponding to the first state under the i-th physical indicator at the t-th target moment, represents the value corresponding to the second state of the i-th physical indicator at the t-th target moment, represents the value corresponding to the first state under the jth physical indicator at the tth target moment, Represents the value corresponding to the second state under the kth physical indicator at the tth target moment.

8. A rehabilitation training auxiliary system based on artificial intelligence, characterized in that: include: A first acquisition module is used to obtain the preoperative physical data of the target patient before the operation, wherein the preoperative physical data includes health data, medical history and preoperative examination results; The second acquisition module is used to obtain the surgical record data corresponding to the target patient during the operation, wherein the surgical record data is the surgical process data obtained during the operation of the target patient after adjusting the physical condition of the target patient to meet the inclusion conditions of the FRAS management model and excluding contraindications; The third acquisition module is used to obtain the specific rehabilitation data of the target patient after surgery; wherein the specific rehabilitation data includes movement status, pain score and functional recovery; A plan generation module, used to generate an initial postoperative rehabilitation plan based on the preoperative physical data, the fixed rehabilitation data and the surgical record data; the initial postoperative rehabilitation plan is used to guide the target patient to perform preliminary rehabilitation training; A rehabilitation monitoring module, used to obtain dynamic rehabilitation monitoring data of the target patient during the initial rehabilitation training according to the initial postoperative rehabilitation plan; A progress prediction module, which is used to predict the rehabilitation progress of the target patient based on the postoperative management plan of the preset rehabilitation management mode and the dynamic rehabilitation monitoring data, and generate a rehabilitation progress prediction result; A plan adjustment module is used to adjust the initial postoperative rehabilitation plan according to the rehabilitation progress prediction result to obtain a target postoperative rehabilitation plan, and the target postoperative rehabilitation plan is used to guide the target patient to perform target rehabilitation training until the target patient meets the discharge conditions of the preset rehabilitation management model.

9. A terminal device, characterized in that: The terminal device includes a processor and a memory; The memory is used to store computer programs; The processor is used to execute the computer program and implement the artificial intelligence-based rehabilitation training assistance method as described in any one of claims 1 to 7 when executing the computer program.

10. A computer storage medium for computer storage, characterized in that: The computer storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the artificial intelligence-based rehabilitation training assistance method described in any one of claims 1 to 7.

Images