System for classifying pain severity and method for classifying pain severity using the same

Abstract

The present invention comprises: a data collection unit that collects data including a patient's brainwave signals and personal information; a data management unit that analyzes the collected data; an artificial intelligence model management unit that trains and manages an artificial intelligence model that classifies and predicts the severity of pain based on the analyzed data; a pain diagnosis unit that infers the severity of the patient's pain based on the artificial intelligence model transmitted from the artificial intelligence model management unit; and a central control unit that controls the operation of the data collection unit, the data management unit, the artificial intelligence model management unit, and the pain diagnosis unit. This is about a pain severity classification system.

Description

Technology Field

[0001] This invention relates to a pain severity classification system and a method for classifying pain severity using the same. Background Technology

[0003] Pain occurs when the affected area is touched during patient treatment in the medical field. In such cases, it can be utilized as an auxiliary tool for medical staff, allowing them to distinguish the severity and criticality of symptoms based on the pain's intensity and to recognize injuries the patient may not be aware of.

[0004] Various pain assessment scales have been developed in relation to the diagnosis of such pain, including the NRS (Numeral Rating Scale), VAS (Visual Analog Scale), and FLACC (face, legs, activity, cry, consolability scale). However, since pain is felt directly only by the patient, there is a problem in that it is difficult to objectify pain, such as different scores being obtained on the aforementioned pain assessment scales even if pain occurs in the same part due to the same cause.

[0005] Meanwhile, when analyzing human brainwave data, it was confirmed that brainwave abnormalities decreased as the severity of pain reported increased, and that the intensity of brainwaves varied depending on the frequency.

[0006] The paper serving as the background for this study (Jeffrey B Hargrove et al, Quantitative electroencephalographic abnormalities in fibromyalgia patients, 2010) showed that the correlation between pain severity and EEG indices was significant only in specific frequency bands and connectivity indices, indicating a complex pattern that makes it difficult to interpret as a simple linear relationship. The problem to be solved

[0007] The present invention aims to solve the problems of the aforementioned conventional technology by providing a pain severity classification system and a classification method for objectively evaluating the severity of pain in real time through artificial intelligence.

[0008] However, the technical problems that the embodiments of the present invention aim to solve are not limited to those described above, and other technical problems may exist. means of solving the problem

[0010] As a technical means for achieving the above-mentioned technical task, the first aspect of the present invention relates to a pain severity classification system comprising: a data collection unit for collecting data including a patient's brainwave signals and personal information; a data management unit for analyzing the collected data; an artificial intelligence model management unit for learning and managing an artificial intelligence model that classifies and predicts the severity of pain based on the analyzed data; a pain diagnosis unit for inferring the patient's pain severity based on an artificial intelligence model transmitted from the artificial intelligence model management unit; and a central control unit for controlling the operation of the data collection unit, the data management unit, the artificial intelligence model management unit, and the pain diagnosis unit; wherein the data management unit includes brainwave signals and personal information stored in advance.

[0011] According to one embodiment of the present invention, the data management unit may include, but is not limited to, a preprocessing unit for preprocessing the collected data and a database unit for storing the preprocessed data in a table.

[0012] According to one embodiment of the present invention, the preprocessing may include, but is not limited to, a method selected from the group consisting of noise removal, missing value removal, standardization, outlier removal, Fast Fourier transformation (FFT), power spectrum density (PSD) calculation, wavelet transformation, and combinations thereof.

[0013] According to one embodiment of the present invention, the artificial intelligence model management unit may include, but is not limited to: a clustering unit that clusters data stored in the database unit according to characteristics; an artificial intelligence model learning unit that trains the artificial intelligence model through data stored in the database unit; and a model transmission unit that transmits at least one of the trained artificial intelligence models to the pain diagnosis unit.

[0014] According to one embodiment of the present invention, the clustering unit may tabulate the patient's personal information according to an algorithm selected from a group consisting of K-means, hierarchical clustering, DBSCAN, and combinations thereof, but is not limited thereto.

[0015] According to one embodiment of the present invention, the artificial intelligence model may include an algorithm selected from the group consisting of CNN, RNN, NN, RF, XGboost, ensemble, and combinations thereof, but is not limited thereto.

[0016] According to one embodiment of the present invention, the artificial intelligence model learning unit may retrain the artificial intelligence model according to a predetermined period or the number of data collected by the data collection unit, but is not limited thereto.

[0017] According to one embodiment of the present invention, the pain diagnosis unit may include, but is not limited to: a cluster selection unit that selects a cluster to which the patient belongs among the clusters of the clustering unit based on the collected data; a model selection unit that selects an artificial intelligence model corresponding to the cluster to which the patient belongs among the artificial intelligence models; and a pain inference unit that diagnoses the severity of the patient's pain by inputting the patient's data into the selected artificial intelligence model.

[0018] According to one embodiment of the present invention, the pain severity classification system may additionally include an output unit that outputs the pain severity inferred by the pain diagnosis unit in the form of at least one of a visual medium and an auditory medium, but is not limited thereto.

[0019] According to one embodiment of the present invention, the personal information may include, but is not limited to, the information selected from the group consisting of the patient's physical information, social information, lifestyle information, disease information, and combinations thereof.

[0020] In addition, the second aspect of the present invention relates to a method for classifying pain severity using a pain severity classification system, comprising: (a) a step of collecting data including brainwave signals and personal information from a patient; (b) a step of analyzing the collected data; (c) a step of learning and managing an artificial intelligence model that classifies and predicts the severity of pain based on the analyzed data; and (d) a step of inferring the severity of the patient's pain based on the artificial intelligence model transmitted from the artificial intelligence model management unit; wherein step (c) involves learning and managing an artificial intelligence model based on the brainwave signals and personal information stored in advance, and the data analyzed in step (b).

[0021] According to one embodiment of the present invention, step (c) may include, but is not limited to, a step of evaluating the performance of the learned artificial intelligence model; and a step of replacing the existing artificial intelligence model with the learned artificial intelligence model if the evaluation result is greater than or equal to a threshold value.

[0022] In addition, the third aspect of the present invention relates to a recording medium in which a pain severity classification system according to the first aspect is implemented in the form of a program executed on a computer.

[0023] The means for solving the problem described above are merely exemplary and should not be interpreted as intended to limit the present invention. In addition to the exemplary embodiments described above, additional embodiments may exist in the drawings and the detailed description of the invention. Effects of the invention

[0024] According to the means for solving the problem of the present invention described above, the pain severity classification system and method according to the present invention overcome the limitations of conventional subjective pain evaluation methods and enable objective, real-time evaluation of a patient's pain.

[0025] In addition, the above pain severity classification system and method can minimize variability according to patient type by applying a customized artificial intelligence model optimized for characteristics clustered according to patient type (gender, age, etc.).

[0026] Furthermore, the aforementioned pain severity classification system and method provide medical staff with real-time pain information and severity classification results, thereby assisting in the constant monitoring of changes in the patient's condition and the flexible adjustment of treatment plans. Through this, diagnosis and treatment are integrated to enhance personalized medical services, and the physical burden on the patient is minimized through brainwave-based pain measurement.

[0027] In addition, since the above pain severity classification system and method periodically updates the learning model by utilizing continuously accumulated brainwave signal data and patient information, data updates are possible, and diagnostic accuracy and reliability improve as the amount of data collected increases.

[0028] That is, the pain severity classification system and method according to the present invention can maximize the work efficiency of medical staff and increase the efficacy of patient-tailored treatment by realizing the objectification of pain evaluation and real-time feedback.

[0029] However, the effects obtainable from this invention are not limited to those described above, and other effects may exist. Brief explanation of the drawing

[0030] FIG. 1 is a schematic diagram of a pain severity classification system according to one embodiment of the present invention. FIG. 2 is a schematic diagram of a pain severity classification system according to one embodiment of the present invention. FIG. 3 is a schematic diagram of a pain severity classification system according to one embodiment of the present invention. FIG. 4 is a schematic diagram of a pain severity classification system according to one embodiment of the present invention. FIG. 5 is a schematic diagram of a pain severity classification system according to one embodiment of the present invention. FIG. 6 is a schematic diagram of a pain severity classification system according to one embodiment of the present invention. FIG. 7 is a flowchart of a pain severity classification method according to one embodiment of the present invention. Specific details for implementing the invention

[0031] Below, embodiments of the present invention are described in detail with reference to the attached drawings so that those skilled in the art can easily implement them.

[0032] However, the present invention may be implemented in various different forms and is not limited to the embodiments described herein. Furthermore, in order to clearly illustrate the present invention in the drawings, parts unrelated to the description have been omitted, and similar parts throughout the specification have been given similar reference numerals.

[0033] Throughout this specification, when a part is described as being "connected" to another part, this includes not only cases where they are "directly connected," but also cases where they are "electrically connected" with other elements interposed between them.

[0034] Throughout the entire specification, when a component is described as being located "on," "on top," "on top," "under," "on bottom," or "on bottom" of another component, this includes not only cases where the component is in contact with the other component but also cases where another component exists between the two components.

[0035] Throughout this specification, when a part is described as "comprising" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0036] Hereinafter, a pain severity classification system and a pain severity classification method using the same according to one embodiment and example of the present invention will be described.

[0037] As a technical means for achieving the above-mentioned technical task, the first aspect of the present invention relates to a pain severity classification system comprising: a data collection unit (110) for collecting data including a patient's brainwave signal and personal information; a data management unit (130) for analyzing the collected data; an artificial intelligence model management unit (140) for learning and managing an artificial intelligence model that classifies and predicts the severity of pain based on the analyzed data; a pain diagnosis unit (150) for inferring the severity of the patient's pain based on the artificial intelligence model transmitted from the artificial intelligence model management unit (140); and a central control unit (120) for controlling the operation of the data collection unit (110), the data management unit (130), the artificial intelligence model management unit (140), and the pain diagnosis unit (150), wherein the data management unit (130) includes previously stored brainwave signal and personal information.

[0038] Generally, it is difficult to objectively assess pain severity in real-time, and while pain exhibits significant variability due to its subjective nature, it can be utilized as a diagnostic aid. To this end, conventional studies have attempted various approaches, such as analyzing differences between patients with chronic pain (e.g., fibromyalgia) and healthy controls to estimate pain severity within a certain range and develop pain assessment indices; however, these methods are limited in providing the immediate, real-time classification of pain intensity and personalized patient feedback required in actual clinical practice. Furthermore, electroencephalogram (EEG) signals collected from patients exhibit significant individual variability and can be heavily influenced by various factors, including age, gender, physical condition, sleep patterns, medication use, and psychological state. Moreover, while conventional studies have attempted to minimize this variability by limiting the sample size or controlling preconditions, this approach presents a practical problem in actual medical settings where patient restrictions cannot be imposed.

[0039] There is a problem in applying different conditions to each patient.

[0040] While conventional studies have often remained at the level of post-hoc analysis of data collected primarily in laboratory environments, it was confirmed that there is no brain-computer interface (BCI)-based pain severity classification system capable of enabling medical professionals to monitor patient severity in real time based on objective indicators and adjust treatment plans in a timely manner.

[0041] Meanwhile, to analyze the brainwave characteristics of chronic pain patients, Electroencephalography (EEG) was measured in the patients' resting state, and quantitative EEG (qEEG) indices were calculated by quantifying the power (delta, theta, alpha, beta, etc.) of the brainwaves by frequency band using signal processing techniques such as Fast Fourier Transform (FFT). Connectivity analysis was performed to derive brain network characteristics related to pain perception by measuring interactions between brain regions (coherence, correlation, impedance, etc.), and the differences between the chronic pain patient group and the healthy control group were statistically verified using the calculated qEEG indices and connectivity indices to compare the pain group and the control group.

[0042] The pain severity classification system according to the present invention relates to a system for objectively evaluating pain that is subjectively perceived by each individual. It aims to provide a system that objectively and accurately classifies pain severity based on real-time electroencephalogram (EEG) data from patients using Brain-Computer Interface (BCI) technology. Through this, the system overcomes the limitations of existing subjective pain assessment methods and enhances patient-tailored medical services by supporting immediate and reliable pain diagnosis and the establishment of treatment plans in clinical settings. The objective is to provide an integrated system that objectively evaluates pain severity in real-time, overcomes high variability caused by individual patient differences influenced by sleep patterns, medication use, psychological state, age, gender, and physical condition, and enables medical staff to monitor the patient's pain status in real-time and adjust treatment plans in a timely manner.

[0043] FIGS. 1 to 6 are schematic diagrams of a pain severity classification system according to one embodiment of the present invention. Specifically, FIG. 1 is a schematic diagram of a pain severity classification system, and FIG. 2 is a schematic diagram considering the operation of the system. In addition, FIG. 3 is a schematic diagram of the process of storing data collected from a patient in a data management unit (130), FIG. 4 is a schematic diagram of the operation of the data management unit (130) and the artificial intelligence model management unit (140), FIG. 5 is a schematic diagram of the operation of the artificial intelligence model management unit (140), and FIG. 6 is a schematic diagram of the operation of the pain diagnosis unit (150).

[0044] First, the data collection unit (110) can collect brainwave (EEG) signals in real time through a brainwave measuring device worn by the patient, and simultaneously collect personal information from the patient and transmit it to the central control unit (120). As will be described later, the patient's brainwave signals, physical information, social information, lifestyle information, and disease information are intended to evaluate the severity of the patient's pain.

[0045] According to one embodiment of the present invention, the personal information may include, but is not limited to, the information selected from the group consisting of the patient's physical information, social information, lifestyle information, disease information, and combinations thereof.

[0046] According to one embodiment of the present invention, the physical information may be selected from a group consisting of the patient's height, weight, gender, biosignals excluding brain waves (e.g., blood glucose), and combinations thereof; the social information may be selected from a group consisting of the patient's occupation, surrounding social environment, and combinations thereof; the lifestyle information may be selected from a group consisting of the patient's sleep pattern, smoking status, drinking status, and combinations thereof; and the disease information may be selected from a group consisting of the patient's underlying disease, medications taken, health status, and combinations thereof, but is not limited thereto.

[0047] Next, the central control unit (120) receives EEG data and patient personal information (physical information, social information, lifestyle, disease information) collected from the data collection unit (110), transmits the EEG data and personal information to the data management unit (130), the artificial intelligence model management unit (140), and the pain diagnosis unit (150), and controls the data flow between them. Additionally, although not shown in FIG. 2, the pain severity diagnosed by the pain diagnosis unit (150) can be transmitted to an external terminal, for example, a medical staff terminal, to help determine the patient's pain severity.

[0048] In FIG. 2, data is transmitted from the central control unit (120) to the data management unit (130), and the transmitted data is moved to the artificial intelligence model management unit (140) and the pain diagnosis unit (150), after which data exchange between the pain diagnosis unit (150) and the central control unit (120) is performed; however, as needed, data exchange may be performed between the central control unit (120) and the data management unit (130), between the central control unit (120) and the artificial intelligence model management unit (140), and between the central control unit (120) and the pain diagnosis unit (150), respectively.

[0049] According to one embodiment of the present invention, the data management unit (130) may include, but is not limited to, a preprocessing unit (131) for preprocessing the collected data and a database unit (132) for storing the preprocessed data in a table.

[0050] According to one embodiment of the present invention, the preprocessing may include, but is not limited to, a method selected from the group consisting of noise removal, missing value removal, standardization, outlier removal, Fast Fourier transformation (FFT), power spectrum density (PSD) calculation, wavelet transformation, and combinations thereof.

[0051] The data management unit (130) converts the EEG data received from the central control unit (120) into quantitative EEG (qEEG) data by undergoing a preprocessing process to remove noise or outliers. Additionally, the personal information received from the central control unit (120) undergoes basic preprocessing necessary for clustering, such as processing missing values, standardization, and removing outliers. The EEG data and personal information that have undergone the above preprocessing process are stored in the database unit (132), specifically divided into a raw EEG table, a qEEG table, a body information table, a social information table, a lifestyle information table, and a disease information table.

[0052] Specifically, the data collection unit (110) acquires time-series brainwave data in real time through an EEG measuring device. At this time, in order to minimize noise caused by poor electrode contact, electromyography noise, eye blinking, etc., the EEG data (raw EEG data) collected through the EEG measuring device is preprocessed using various techniques such as a band-pass filter, ICA (Independent Component Analysis), and an adaptive filter.

[0053] In addition, methods such as FFT (Fast Fourier Transform), PSD (Power Spectral Density) calculation, and Wavelet Transform are applied to the EEG signal to extract relative intensity and time-frequency characteristics for each frequency band (delta, theta, alpha, beta, gamma) from the EEG data.

[0054] According to one embodiment of the present invention, the brainwave may include delta waves of 0.5 Hz to 4 Hz, theta waves of 4 Hz to 8 Hz, alpha waves of 8 Hz to 13 Hz, beta waves of 13 Hz to 30 Hz, and gamma waves of 30 Hz or higher, but is not limited thereto.

[0055] Through the above process, qEEG data can be obtained, and specific features closely associated with pain perception can be obtained from the qEEG data.

[0056] Meanwhile, as will be described later, the data stored in the database unit (132) is clustered. To this end, preprocessing steps such as handling missing values ​​(e.g., average value, median replacement), removing outliers, or standardization or normalization for unifying the appropriate range are performed to prepare the clustering algorithm described later to operate stably. At this time, clustering and model training are performed smoothly even when there is significant variation among patients or when some data is missing.

[0057] The above database section (132) is configured by integrating preprocessed EEG data, qEEG data, personal information, lifestyle information, social information, and disease information, and by storing each of the data separately in the form of an EEG table, qEEG table, patient information table, and lifestyle table, it is possible to flexibly reference them later during the artificial intelligence model training and pain severity diagnosis stages.

[0058] According to one embodiment of the present invention, the artificial intelligence model management unit (140) may include: a clustering unit (141) that clusters data stored in the database unit (132) according to characteristics; an artificial intelligence model learning unit (142) that learns the artificial intelligence model through the data stored in the database unit (132); and a model transmission unit (143) that transmits at least one of the learned artificial intelligence models to the pain diagnosis unit (150), but is not limited thereto.

[0059] According to one embodiment of the present invention, the clustering unit (141) may table the patient's personal information according to an algorithm selected from a group consisting of K-means, hierarchical clustering, DBSCAN, and combinations thereof, but is not limited thereto.

[0060] According to one embodiment of the present invention, the artificial intelligence model may include an algorithm selected from the group consisting of CNN, RNN, NN, RF, XGboost, ensemble, and combinations thereof, but is not limited thereto.

[0061] The artificial intelligence model management unit (140) learns and manages an artificial intelligence model to classify and predict pain severity based on preprocessed data received from the data management unit (130) or data received from the central control unit (120). Specifically, by subdividing patient groups into groups A, B, C, etc., using a clustering algorithm, the data stored in the database unit (132) is learned to learn a pain model suitable for each group's characteristics and to secure customized diagnostic performance optimized for each individual patient. Subsequently, the learned models are managed through a version and distribution management system so that they can be utilized in real time by the pain diagnosis unit (150).

[0062] Specifically, the clustering unit (141) uses personal information stored in the database unit (132) as input variables and clusters by applying algorithms such as K-Means, Hierarchical Clustering, and DBSCAN. At this time, EEG and qEEG are not included in the clustering target, and patient groups are classified based only on the personal information of each individual patient. Through this process, patients can be classified according to various population group characteristics, such as elderly groups, young and middle-aged groups, and groups with similar lifestyle habits (e.g., smoker group / non-smoker group, drinking group / non-drinking group, exercise group / non-exercise group, etc.), and can be classified to include, for example, Group A (elderly), Group B (young, middle-aged, lifestyle), and Group C (specific vulnerability).

[0063] Next, the clustering results are analyzed to identify the characteristics of each cluster (e.g., average age, major diseases, medications taken, lifestyle habits, etc.), and the validity of the clusters is evaluated using statistical methods such as the Silhouette Score and the Elbow Method. At this time, the number of clusters or parameters are adjusted to secure an optimal cluster structure, and clinically meaningful clusters are determined through review by medical staff. By storing classification criteria (cluster centroids, hierarchical partitioning rules, etc.) based on the determined clusters in the model management unit (140), when a new patient is introduced and data is collected, the new patient can be automatically classified to which cluster they belong to.

[0064] In addition, to train an artificial intelligence model in the artificial intelligence model training unit (142), first, EEG signal characteristics (FFT, PSD, wavelet-based power, etc.) obtained through preprocessing and qEEG transformation processes, and clinically obtained pain severity labels (mild, moderate, severe, etc.) are configured as training data. At this time, cluster information to which each patient belongs can be used as an additional input feature, and the model can be trained to use an independent model for each cluster. Subsequently, the artificial intelligence model can be trained by inputting the above-mentioned EEG signal characteristics and pain severity labels into the model, and at this time, different data may be used for each model. For example, a Convolutional Neural Network (CNN) is used to automatically learn key patterns after converting time-series EEG data into a 2D (time × channel) format, and a Feedforward Neural Network (NN), which is a multi-layer perceptron structure, classifies pain severity by receiving qEEG indices (frequency band power, phase synchronization, etc.) and cluster information as input. Random Forest (RF) and XGBoost, among others, learn features extracted from EEG and patient personal information using tree-based ensemble techniques, thereby ensuring both interpretability and high generality. If necessary, different algorithms can be adopted for each cluster, a single cluster can use multiple algorithms, and accuracy can be maximized by combining the prediction results of multiple models through ensemble techniques.

[0065] In this regard, the above CNN, RNN, and NN may be deep learning algorithms, and RF and XGBoost may be machine learning algorithms.

[0066] Subsequently, once training is complete, the model is used to calculate artificial intelligence model performance metrics, including accuracy, precision, recall, and F1 score, for pain severity diagnosis using k-Fold Cross Validation or a separate test set, and the performance of each cluster model or the entire model is compared to finally select the model (or combination of models) with the best performance. At this time, since multiple models can be selected for a single cluster and a single model can be used for multiple clusters, if at least one of the components constituting the artificial intelligence model performance metrics falls below a certain level, it is necessary to increase the amount of data or perform retraining to reduce errors.

[0067] According to one embodiment of the present invention, the artificial intelligence model learning unit (142) may retrain the artificial intelligence model according to a predetermined period or the number of data collected from the data collection unit (110), but is not limited thereto.

[0068] According to one embodiment of the present invention, the above-mentioned predetermined period may be one week, one month, one quarter (three months), half a year (six months), or one year, but is not limited thereto.

[0069] The number of collected data refers to the number of data collected from the data collection unit (110) and newly input into the database unit (132), and since it is newly input, it is not used for training the artificial intelligence model. At this time, if the number of collected data is 50% to 200% of the number of existing data (the number of data stored in the existing database unit (132)), the artificial intelligence model can be retrained.

[0070] The aforementioned artificial intelligence model outputs the pain severity described below as a result after additional data is input during the pain diagnosis process; since the output result is reviewed by a human and corrected if incorrect, it is trained. However, because the artificial intelligence model can become more sophisticated over time and may not be optimized if left unattended after inputting data for patient pain diagnosis, data may accumulate indiscriminately, it is necessary to reorganize the model at regular time intervals or based on the amount of accumulated data. At this time, the reorganization of the artificial intelligence model can be performed automatically.

[0071] Specifically, the artificial intelligence model management unit (140) may include a model version management unit within the artificial intelligence model learning unit (142) or separately from the artificial intelligence model learning unit (142). The model version management unit stores the learned artificial intelligence models by version, continuously monitors the performance history (accuracy, recall, F1 score, etc.) of the corresponding models, and when a model showing optimal performance is determined, transmits it to the pain diagnosis unit (150) to be used for real-time inference of pain severity. At this time, the existing model is preserved as a backup for rollback, so that it can be switched immediately when necessary.

[0072] In addition, newly introduced patient EEG signals and personal information from the clinical site are transmitted to the data management unit (130), preprocessed, and then stored in the database unit (132). At this time, as data continues to accumulate in the database unit (132), the model version management unit may automatically determine the time for retraining the artificial intelligence model, or proceed semi-automatically according to the administrator's settings.

[0073] An automatic learning pipeline is executed according to conditions of a scheduler (e.g., weekly, monthly) or trigger (e.g., accumulation of data of a certain size or larger, need for cluster reclassification) within the model management unit (140), and the newly trained model is compared with the existing model's performance indicators through a test set or validation indicators. If the performance indicators of the retrained model are improved compared to the performance indicators of the existing model or if the diagnosis time is shortened, the version of the retrained model can be set as the new version.

[0074] If the model provided by the artificial intelligence model learning unit (142) continues to provide a pain severity diagnosis service, model improvement is performed without interruption of service by learning and verifying a new model different from the existing model in a parallel learning environment.

[0075] At this time, when it is decided to update the new model, the model transmission unit (143) automatically distributes the new model to the pain diagnosis unit (150) described later and transmits a signal to switch the existing model to a standby state. Users, such as medical staff, can always check the inference results of the latest model in real time without any separate operation, but if necessary, the latest model can be downloaded after the user's terminal requests a model update.

[0076] According to one embodiment of the present invention, the pain diagnosis unit (150) may include, but is not limited to: a cluster selection unit (151) that selects a cluster to which the patient belongs among the clusters of the clustering unit (141) based on the collected data; a model selection unit (152) that selects an artificial intelligence model corresponding to the cluster to which the patient belongs among the artificial intelligence models; and a pain inference unit (153) that diagnoses the severity of the patient's pain by inputting the patient's data into the selected artificial intelligence model.

[0077] The pain diagnosis unit (150) can infer the severity of a patient's pain in real time using a learning model transmitted from the model transmission unit (143) of the artificial intelligence model management unit (140). Specifically, upon receiving the patient's EEG signal and personal information, the unit includes a pain inference unit (153) that selects a cluster model or brainwave model with high similarity to the patient based on the input data from the cluster selection unit (151), selects an artificial intelligence model optimized for the selected model from the model selection unit (152), and diagnoses the severity of pain by inputting the patient's data into the selected artificial intelligence model. The pain severity information calculated by the pain inference unit (153) can be transmitted to an external terminal via the central control unit (120) to assist the medical staff in making a judgment. Through this process, the medical staff can intuitively and reliably verify the patient's pain information and establish a rapid and appropriate treatment plan to provide customized medical services to the patient.

[0078] First, the data collection unit (110) collects the patient's real-time brainwave signals and the patient's latest personal information and transmits them to the central control unit (120). Subsequently, the central control unit (120) transmits the brainwave signals and the personal information to the pain diagnosis unit (150), and the pain diagnosis unit (150) identifies which cluster the patient belongs to among the clusters defined in the clustering unit (141) based on the transmitted brainwave signals and personal information (cluster selection unit (151)). Subsequently, among the multiple models transmitted from the model transmission unit (143), a model optimized for the cluster to which the patient belongs is selected (model selection unit (152)), and the patient's data is input into the selected model to infer the severity of pain (mild, moderate, severe, etc.) (pain inference unit (153)). The result value calculated by the pain inference unit (153) is transmitted to an external terminal via the central control unit (120).

[0079] However, the process of selecting the above cluster may be performed by the central control unit (120).

[0080] The aforementioned external terminal is used by medical staff, who can monitor the patient's pain status displayed on the terminal in real time. Through this, appropriate treatments, such as prescribing medication or physical therapy, can be administered to the patient as needed. Additionally, the patient's real-time pain trends (markers, graphs, numerical indicators, etc.) and past history are visualized together, allowing for easy identification of changes in the condition. Furthermore, if the degree of pain changes rapidly or results differ from expectations, the medical staff can conduct immediate examinations or re-evaluations.

[0081] According to one embodiment of the present invention, the pain severity classification system may additionally include an output unit that outputs the pain severity inferred from the pain diagnosis unit (150) in the form of at least one of a visual medium and an auditory medium, but is not limited thereto.

[0082] The above output unit may be built into the terminal.

[0083] In addition, the second aspect of the present invention relates to a method for classifying pain severity using a pain severity classification system, comprising: (a) collecting data including brainwave signals and personal information from a patient; (b) analyzing the collected data; (c) learning and managing an artificial intelligence model that classifies and predicts the severity of pain based on the analyzed data; and (d) inferring the severity of the patient's pain based on the artificial intelligence model transmitted from the artificial intelligence model management unit (140); wherein step (c) involves learning and managing an artificial intelligence model based on the brainwave signals and personal information stored in advance, and the data analyzed in step (b).

[0084] At this time, the method for classifying pain severity according to the second aspect utilizes the pain severity classification system according to the first aspect.

[0085] FIG. 7 is a flowchart of a pain severity classification method according to one embodiment of the present invention.

[0086] First, (a) collect data including brainwave signals and personal information from the patient.

[0087] Next, (b) analyze the collected data above.

[0088] Next, (c) an artificial intelligence model is trained and managed to classify and predict the severity of pain based on the analyzed data above.

[0089] In this regard, the artificial intelligence model training in step (c) above is first performed based on data collected and analyzed from multiple patients. While performing step (d) with the trained artificial intelligence model, data from new patients is collected and analyzed. At this time, if the number of analyzed data accumulates to a certain level or if a certain amount of time has elapsed since the model training, the model may be retrained.

[0090] According to one embodiment of the present invention, step (c) may include, but is not limited to, a step of evaluating the performance of the learned artificial intelligence model; and a step of replacing the existing artificial intelligence model with the learned artificial intelligence model if the evaluation result is greater than or equal to a threshold value.

[0091] If the performance indicators of the above model are below a threshold, data input and analysis can be performed under human supervision to improve the model's performance indicators to be above the threshold. However, if, while using the artificial intelligence model, the performance indicators of a newly trained model are above the threshold, the existing artificial intelligence model may be replaced with the newly trained model.

[0092] Next, (d) the severity of the patient's pain is inferred based on the artificial intelligence model transmitted from the artificial intelligence model management unit (140).

[0093] Inference of the patient's pain severity is performed using new patient data that was not used for model training. At this time, the newly input data is stored in the database unit (132) via the data management unit (130), and pain severity inference is performed simultaneously.

[0094] In addition, the third aspect of the present invention relates to a recording medium in which a pain severity classification system according to the first aspect is implemented in the form of a program executed on a computer.

[0095] The foregoing description of the present invention is for illustrative purposes only, and those skilled in the art will understand that other specific forms can be easily modified without altering the technical concept or essential features of the present invention. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single unit may be implemented in a distributed manner, and components described as distributed may likewise be implemented in a combined form.

[0096] The scope of the present invention is defined by the claims set forth below rather than by the detailed description above, and all modifications or variations derived from the meaning and scope of the claims and the concept of equivalents thereof should be interpreted as being included within the scope of the present invention. Explanation of the symbols

[0097] 110: Data Collection Unit 120 : Central Control Unit 130 : Data Management Department 131 : Preprocessing section 132 : Database Section 140 : Artificial Intelligence Model Management Department 141 : Clustering section 142 : Artificial Intelligence Model Training Unit 143 : Model Transmission Section 150 : Pain Diagnosis Section 151 : Cluster selection section 152 : Model Selection Section 153 : Pain Inference Section

Claims

Claim 1 A pain severity classification system comprising: a data collection unit for collecting data including a patient's brainwave signals and personal information; a data management unit for analyzing the collected data; an artificial intelligence model management unit for learning and managing an artificial intelligence model that classifies and predicts the severity of pain based on the analyzed data; a pain diagnosis unit for inferring the patient's pain severity based on an artificial intelligence model transmitted from the artificial intelligence model management unit; and a central control unit for controlling the operation of the data collection unit, the data management unit, the artificial intelligence model management unit, and the pain diagnosis unit, wherein the data management unit includes previously stored brainwave signals and personal information. Claim 2 A pain severity classification system according to claim 1, wherein the data management unit comprises a preprocessing unit that preprocesses the collected data and a database unit that stores the preprocessed data in a table. Claim 3 A pain severity classification system according to paragraph 2, wherein the preprocessing comprises a method selected from the group consisting of noise removal, missing value removal, standardization, outlier removal, Fast Fourier transformation (FFT), power spectrum density (PSD) calculation, wavelet transformation, and combinations thereof. Claim 4 A pain severity classification system according to claim 1, wherein the artificial intelligence model management unit comprises: a clustering unit that clusters data stored in the database unit according to characteristics; an artificial intelligence model learning unit that trains the artificial intelligence model through data stored in the database unit; and a model transmission unit that transmits at least one of the trained artificial intelligence models to the pain diagnosis unit. Claim 5 In paragraph 4, the above-mentioned clustering unit tabulates the patient's personal information according to an algorithm including those selected from a group consisting of K-means, hierarchical clustering, DBSCAN, and combinations thereof, in a pain severity classification system. Claim 6 A pain severity classification system according to claim 4, wherein the artificial intelligence model comprises an algorithm selected from the group consisting of CNN, RNN, NN, RF, XGBoost, ensemble, and combinations thereof. Claim 7 A pain severity classification system according to claim 4, wherein the artificial intelligence model learning unit retrains the artificial intelligence model according to a predetermined period or the number of data collected by the data collection unit. Claim 8 In claim 4, the pain diagnosis unit comprises: a cluster selection unit that selects a cluster to which the patient belongs among the clusters of the clustering unit based on the collected data; a model selection unit that selects an artificial intelligence model corresponding to the cluster to which the patient belongs among the artificial intelligence models; and a pain inference unit that diagnoses the pain severity of the patient by inputting the patient's data into the selected artificial intelligence model; a pain severity classification system. Claim 9 A pain severity classification system according to claim 1, wherein the pain severity classification system further includes an output unit that outputs the pain severity inferred from the pain diagnosis unit in the form of at least one of a visual medium and an auditory medium. Claim 10 A pain severity classification system according to claim 1, wherein the personal information comprises selected from the group consisting of the patient's physical information, social information, lifestyle information, disease information, and combinations thereof. Claim 11 A method for classifying pain severity using a pain severity classification system, comprising: (a) a step of collecting data including brainwave signals and personal information from a patient; (b) a step of analyzing the collected data; (c) a step of training and managing an artificial intelligence model that classifies and predicts the severity of pain based on the analyzed data; and (d) a step of inferring the severity of the patient's pain based on the artificial intelligence model transmitted from the artificial intelligence model management unit; wherein step (c) involves training and managing an artificial intelligence model based on the stored brainwave signals and personal information, and the data analyzed in step (b). Claim 12 A method for classifying pain severity according to claim 11, wherein step (c) comprises: a step of evaluating the performance of the learned artificial intelligence model; and a step of replacing the existing artificial intelligence model with the learned artificial intelligence model if the evaluation result is greater than or equal to a threshold value. Claim 13 A recording medium in which a pain severity classification system according to paragraph 1 is implemented and recorded in the form of a program executed on a computer.

Images