CN119949795B - An intelligent automatic pain analysis system - Google Patents
An intelligent automatic pain analysis systemInfo
- Publication number
- CN119949795B CN119949795B CN202510291539.5A CN202510291539A CN119949795B CN 119949795 B CN119949795 B CN 119949795B CN 202510291539 A CN202510291539 A CN 202510291539A CN 119949795 B CN119949795 B CN 119949795B
- Authority
- CN
- China
- Prior art keywords
- pain
- wearable device
- hand
- touch pressure
- remote server
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
本发明提供了一种智能疼痛自动分析系统，其特征在于，包括手、腕部、足三处可穿戴装置，分别用于手触压、心率与血压、足触压随时间变化函数的数据采集，手指EDA皮电信号测量装置，用于与腕部可穿戴装置连接而采集手皮电信号，以及远程服务器，其中，所述手、足两处可穿戴装置产生的信号无线传给腕部可穿戴装置，并由腕部可穿戴装置连带皮电信号测量装置测量的手指皮电信号一并发送给所述远程服务器，由远程服务器利用预训练的基于异种信号时域注意力机制的智能模型预测出疼痛级别。实现了在心率、血压、手足触压力、皮电的数据监测基础上，利用远程服务器进行基于异种注意力信号时域注意力机制的智能模型得到疼痛级别的自动分析，使得疼痛的人体检测能够在院外不脱产地用于药物不良临床反应研究。
The present invention provides an intelligent automatic pain analysis system, characterized in that it includes three wearable devices at the hands, wrists, and feet, which are respectively used for data collection of hand touch pressure, heart rate and blood pressure, and foot touch pressure over time function, a finger EDA skin electrode signal measuring device, which is used to connect with the wrist wearable device to collect hand skin electrode signals, and a remote server, wherein the signals generated by the two wearable devices at the hands and feet are wirelessly transmitted to the wrist wearable device, and the wrist wearable device sends the finger skin electrode signals measured by the skin electrode signal measuring device to the remote server together, and the remote server predicts the pain level using a pre-trained intelligent model based on the time domain attention mechanism of heterogeneous signals. Based on the data monitoring of heart rate, blood pressure, hand and foot touch pressure, and skin electrode, the remote server is used to perform automatic analysis of the pain level using an intelligent model based on the time domain attention mechanism of heterogeneous attention signals, so that the human body detection of pain can be used for the study of adverse clinical reactions of drugs without leaving the production site outside the hospital.
Description
Technical Field
The invention relates to an automatic pain analysis system, in particular to an intelligent automatic pain analysis system, and belongs to the field of intelligent medical systems.
Background
The existing medicine adopts mice for experiments, so that the research of drug application to pain is carried out. However, research results from animal to human are often not linear concepts. Therefore, how to directly collect pain data of a human body is the most accurate to study the stimulation condition of a clinical test of a drug. Thus, analysis of the data base of pain signals in combination with models is required to obtain accurate drug assessments.
On the other hand, the acquisition data for the life outside the hospital is also extremely important. The volunteer can thus perform a large-scale test without be temporarily released from one's regular work, thereby improving the efficiency of the test. Therefore, how to adopt a flexible and convenient acquisition device is also a key of the technology.
Finally, since pain is a subjective perception of a person, it can be described in terms of language, the mechanism behind which is abnormality of physiological parameters, and thus can be described as parameter abnormality-pain description, whereas parameter abnormality is caused by factors such as medication administration, and thus pain can be described as a process of medication-parameter-pain description over time, essentially regarded as a process of natural language generation. Thus, if attention mechanisms are considered, a description of the level of pain may be automatically analyzed in real time instead of a person.
Disclosure of Invention
Based on the problems in the prior art, the invention has the following technical points that the first device for acquiring heart rate, blood pressure and finger and foot pressure data is considered, and the second device adopts an attention mechanism intelligent model based on the time interval of heterogeneous signals to analyze pain data so as to train the level of pain according to the time occurrence relation among four heterogeneous signals of hands, wrists, feet and skin electricity.
Based on the above consideration, the present invention provides an intelligent pain automatic analysis system, comprising three wearable devices of hand, wrist and foot, which are respectively used for data acquisition of hand touch pressure, heart rate and blood pressure, foot touch pressure time-varying function, a finger EDA skin electric signal measuring device, which is used for connecting with the wrist wearable device to acquire hand skin electric signals, and a remote server, wherein,
The signals generated by the hand wearable device and the foot wearable device are transmitted to the wrist wearable device in a wireless mode, the wrist wearable device is connected with the finger skin electric signal measured by the skin electric signal measuring device, the finger skin electric signal measured by the finger skin electric signal measuring device is transmitted to the remote server, and the remote server predicts the pain level by utilizing a pre-trained intelligent model based on heterogeneous signal time domain attention mechanisms.
Optionally, the hand wearable device includes a plurality of hand touch pressure sensors that set up on the flexible substrate to and the first integrated circuit that connects all hand touch pressure sensors that sets up on the flexible substrate, the foot wearable device includes flexible podotheca, flexible podotheca includes the base pad that is provided with a plurality of foot touch pressure sensors and is provided with the second integrated circuit that is connected with all foot touch pressure sensors, the wrist wearable device includes host computer and the wrist strap of being connected with the host computer, EDA skin electric signal measurement device includes a plurality of skin electric acquisition pieces, through signal line with the interface connection that can dismantle electric connection on the host computer, wherein, first integrated circuit and second integrated circuit all include data processor, and data wireless transmitter for with the host computer communication or directly to remote server carries out data transmission.
The intelligent model pre-training method based on the heterogeneous signal time domain attention mechanism comprises the following steps:
s1, constructing a foot touch pressure, a hand touch pressure, SCL (skin electric signal level) and heart rate time change function, and acquiring all wave crests of the SCL;
S2 for each SCL wave crest, dividing the time span of the wave crest into n time elements ,,。。。,Constructing a long-short range memory model by taking each time element as a unitWill beCorrespondence of infinitesimalDetected heart rateInput the first cell, input the next cell through the intermediate transport layer, and thenDetected heart rate corresponding to the infinitesimalInputting the second unit, inputting the next unit via the intermediate transmission layer, and so on to obtainA plurality of output values input into the firstIn the function, a plurality of outputs are obtainedForming an output matrixSimultaneously measuring blood pressure in each time element,,。。。,The matrix formedAnd the output matrixPerforming matrix multiplication to obtain attention valueWhereinA heart rate matrix formed for the heart rates of the n time bins;
s3 for each SCL peak, the same is followed by the step of S2 based on the same n time bins ,,。。。,Establishing a hand touch pressure short range memory modelAlso using a secondThe function obtains multiple outputs to form a matrixMatrix formed by foot contact pressureMatrix multiplication to form motion values,The matrix is touched by hands;
s4 calculating pain class Repeating steps S2-S3 to obtain pain levels of all SCL wave crests, thereby completing the construction of the model, whenThe pain level cannot be judged in exercise, otherwiseFor the purpose of mood swings,For a slight pain, the pain, for example,For moderate pain, the pain is not limited to,The pain of the patient is extremely high,A perceived pain threshold for somatosensory evaluation of all volunteers;
S5, recruiting volunteers, wearing three wearable devices, namely a hand, a wrist and a foot, carrying out medicine intake by using an EDA skin electric signal measuring device, and collecting data by a remote server for training a model according to the steps of S2-S4.
It will be appreciated that by introducing training of hand-foot contact pressure, the resulting exercise value excludes the skin-electricity fluctuations due to exercise, mood, etc., and filters out the relatively pure contributions belonging to pain.
Optionally, the hand wearable device is not attached to the palm face by hydrogel, and the wrist wearable device is a smart watch.
Advantageous effects
The data monitoring of heart rate, blood pressure, hand-foot touch pressure and skin electricity is realized through the sensing of hands, feet and wrists and the intelligent device, and the data monitoring is transmitted to a remote server to perform the automatic analysis of pain level based on the intelligent model of the attention mechanism of the heterogeneous attention signal time domain. So that the human body detection of pain can be used for the research of adverse clinical reaction of medicines without be temporarily released from one's regular work at the hospital.
Drawings
Figure 1 is a schematic diagram of the components of a wearable intelligent pain automatic analysis system according to an embodiment of the present invention,
Figure 2 is a graph of foot pressure, hand pressure, SCL, heart rate signal as a function of time,
FIG. 3 is a logic diagram of a pre-training method algorithm for an intelligent model based on a heterogeneous signal time domain attention mechanism.
Detailed Description
Fig. 1 shows a wearable intelligent pain automatic analysis system, which comprises a flexible substrate with hydrogel attached to the palm surface, a smart watch with a main body with a touch screen, wrist bands connected to two sides of the main body, a foot sleeve, a finger EDA skin electric signal measuring device, a remote server, and a remote server, wherein the data acquisition is respectively used for acquiring hand touch pressure, heart rate and blood pressure, and foot touch pressure time-varying functions (shown in fig. 2).
In fig. 1, a flexible substrate with hydrogel attached to the palm surface is provided with a right thumb, an index finger and a middle finger, and two hand touch pressure sensors a of the palm center, the group comprises a base pad and a top surface, the base pad is provided with a heel one and a foot front end one, the foot touch pressure sensor B of the thumb one, each hand touch pressure sensor a and each foot touch pressure sensor are respectively connected with a first integrated circuit of the heel and a second integrated circuit arranged on the top surface through flexible wires, and the two integrated circuits are respectively provided with a data processing and wireless transmitter for processing and transmitting acquired hand touch pressure data and foot touch pressure data to a remote server.
The finger EDA skin electric signal measuring device is provided with three skin electric acquisition sheets which are arranged on the thumb, the index finger and the palm surface of the left hand and are connected with an interface which is detachably and electrically connected with the host through signal wires, and the host uploads skin electric signals to a remote server. The remote server uses a pre-trained intelligent model of the attention mechanism based on the time intervals of the xenogenic signals to input the received data to predict pain levels.
Fig. 2 shows a plot of foot pressure, hand pressure, SCL, heart rate signal as a function of time, including two phases after taking the subject. For convenience, the time period from the taking of the medicine to the pain is cut off for convenience of explanation. In the figure, after pain is generated, the heart rate starts to accelerate, which is shown by the shortened period, and three corresponding obvious main peaks (indicated by downward arrows) appear on the SCL. The hand-touch function graph shows a plurality of peaks generated by disassembling the package when taking medicine and a stable pressure signal after holding the cup.
The pressure reduction caused by the lifting of the legs when the two legs leave the ground due to pain is generated on the foot touch pressure function chart, which indicates that the experimenter generates body movement due to the pain. However, as the wave peaks of the hand touch and the foot touch are less, the factors of physical exercise can be eliminated. Therefore, from the whole analysis of the heterogeneous signals, the intelligent automatic analysis and perception of the pain sense is realized, and the analysis is needed to be carried out aiming at the attention mechanism of each signal.
The intelligent model pre-training method based on the heterogeneous signal time domain attention mechanism comprises the following steps:
S1, constructing a foot touch, hand touch, SCL skin electric signal level and heart rate time change function, and acquiring all wave crests of the SCL (for example, figure 2);
S2 for each SCL wave crest, dividing the time span of the wave crest into n time elements ,,。。。,Constructing a long-short range memory model by taking each time element as a unitWill beCorrespondence of infinitesimalDetected heart rateInput the first cell, input the next cell through the intermediate transport layer, and thenDetected heart rate corresponding to the infinitesimalInputting the second unit, inputting the next unit via the intermediate transmission layer, and so on to obtainA plurality of output values input into the firstIn the function, a plurality of outputs are obtainedForming an output matrixSimultaneously measuring blood pressure in each time element,,。。。,The matrix formedAnd the output matrixPerforming matrix multiplication to obtain attention valueWhereinA heart rate matrix formed for the heart rates of the n time bins;
s3 for each SCL peak, the same is followed by the step of S2 based on the same n time bins ,,。。。,Establishing a hand touch pressure short range memory modelAlso using a secondThe function results in a plurality of outputsForming a matrixMatrix formed by foot contact pressureMatrix multiplication to form motion values,The matrix is touched by hands;
s4 calculating pain class Repeating steps S2-S3 to obtain pain levels of all SCL wave crests, thereby completing the construction of the model, whenThe pain level cannot be judged in exercise, otherwiseFor the purpose of mood swings,For a slight pain, the pain, for example,For moderate pain, the pain is not limited to,The pain of the patient is extremely high,A perceived pain threshold for somatosensory evaluation of all volunteers;
S5, recruiting volunteers, wearing three wearable devices, namely a hand, a wrist and a foot, carrying out medicine intake by using an EDA skin electric signal measuring device, and collecting data by a remote server for training a model according to the steps of S2-S4.
Claims (1)
1. An intelligent pain automatic analysis system is characterized by comprising three wearable devices, namely a hand wearable device, a wrist wearable device and a foot wearable device, which are respectively used for acquiring data of hand touch pressure, heart rate and blood pressure and foot touch pressure time-dependent functions, a finger EDA skin electric signal measuring device which is used for acquiring hand skin electric signals by connecting with the wrist wearable device, and a remote server,
The signals generated by the hand wearable device and the foot wearable device are transmitted to the wrist wearable device in a wireless mode, the wrist wearable device is connected with the finger skin electric signal measured by the skin electric signal measuring device, the finger skin electric signal is transmitted to the remote server, and the remote server predicts the pain level by utilizing a pre-trained intelligent model based on a heterogeneous signal time domain attention mechanism;
The hand wearable device comprises a plurality of hand touch pressure sensors arranged on a flexible substrate and a first integrated circuit which is arranged on the flexible substrate and is connected with all the hand touch pressure sensors; the wrist wearable device comprises a host and a wrist strap connected with the host, the EDA skin electric signal measuring device comprises a plurality of skin electric acquisition sheets and is connected with an interface detachably and electrically connected with the host through a signal wire, wherein the first integrated circuit and the second integrated circuit both comprise a data processor and a data wireless transmitter, and the data wireless transmitter is used for communicating with the host or directly transmitting data to the remote server;
The intelligent model pre-training method based on the heterogeneous signal time domain attention mechanism comprises the following steps:
s1, constructing a foot touch pressure, a hand touch pressure, a skin electric signal level (SCL) and a heart rate time change function, and acquiring all wave crests of the SCL;
S2 for each SCL wave crest, dividing the time span of the wave crest into n time elements Constructing a long-short range memory model by taking each time element as a unitWill beCorrespondence of infinitesimalDetected heart rateInput the first cell, input the next cell through the intermediate transport layer, and thenDetected heart rate corresponding to the infinitesimalInputting the second unit, inputting the next unit via the intermediate transmission layer, and so on to obtainA plurality of output values input into the firstIn the function, a plurality of outputs are obtainedForming an output matrixSimultaneously measuring blood pressure in each time elementThe matrix formedAnd the output matrixPerforming matrix multiplication to obtain attention valueWhereinA heart rate matrix formed for the heart rates of the n time bins;
s3 for each SCL peak, the same is followed by the step of S2 based on the same n time bins Establishing a hand touch pressure short range memory modelAlso using a secondThe function results in a plurality of outputsForming a matrixMatrix formed by foot contact pressureMatrix multiplication to form motion values,The matrix is touched by hands;
s4 calculating pain class Repeating steps S2-S3 to obtain pain levels of all SCL wave crests, thereby completing the construction of the model, whenThe pain level cannot be judged in exercise, otherwiseFor the purpose of mood swings,For a slight pain, the pain, for example,For moderate pain, the pain is not limited to,The pain of the patient is extremely high,The pain threshold was felt for all volunteer somatosensory evaluation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510291539.5A CN119949795B (en) | 2025-03-12 | 2025-03-12 | An intelligent automatic pain analysis system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510291539.5A CN119949795B (en) | 2025-03-12 | 2025-03-12 | An intelligent automatic pain analysis system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN119949795A CN119949795A (en) | 2025-05-09 |
| CN119949795B true CN119949795B (en) | 2025-07-18 |
Family
ID=95598415
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510291539.5A Active CN119949795B (en) | 2025-03-12 | 2025-03-12 | An intelligent automatic pain analysis system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119949795B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016124482A1 (en) * | 2015-02-02 | 2016-08-11 | Koninklijke Philips N.V. | Pain management wearable device |
Family Cites Families (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11631280B2 (en) * | 2015-06-30 | 2023-04-18 | University Of South Florida | System and method for multimodal spatiotemporal pain assessment |
| US12507948B2 (en) * | 2015-06-30 | 2025-12-30 | University Of South Florida | System and method for attentional multimodal pain estimation |
| AU2017335497B2 (en) * | 2016-09-27 | 2020-01-16 | Boston Scientific Neuromodulation Corporation | System for pain management using objective pain measure |
| EP3532155B1 (en) * | 2016-10-25 | 2025-08-27 | Boston Scientific Neuromodulation Corporation | System for pain management using baroreflex sensitivity |
| US20190343457A1 (en) * | 2018-05-08 | 2019-11-14 | The Regents Of The University Of California | Pain assessment method and apparatus for patients unable to self-report pain |
| CN112741593B (en) * | 2019-10-29 | 2022-09-02 | 重庆医科大学附属儿童医院 | Pain assessment method, system, device and storage medium |
| CN111210907A (en) * | 2020-01-14 | 2020-05-29 | 西北工业大学 | A Pain Intensity Estimation Method Based on Spatial-Temporal Attention Mechanism |
| WO2021188079A2 (en) * | 2020-03-16 | 2021-09-23 | Eski̇şehi̇r Osmangazi̇ Üni̇versi̇tesi̇ | An artificial intelligence method to ensure the assessment the level of comfort and pain of babies |
| KR102433828B1 (en) * | 2020-08-20 | 2022-08-18 | 부산대학교 산학협력단 | Pain Behavior Monitoring System and Method for Managing Pain using the same |
| AU2022299056B2 (en) * | 2021-06-21 | 2025-02-13 | Boston Scientific Neuromodulation Corporation | Cloud-based patient monitoring system |
| IL310637A (en) * | 2021-08-05 | 2024-04-01 | Sweetch Health Ltd | Systems and methods for prediction, prevention and management of chronic and autoimmune diseases |
| US20240194343A1 (en) * | 2022-12-09 | 2024-06-13 | Hero Medical Technologies Inc. | Pain detection via machine learning applications |
| CN118370521B (en) * | 2024-06-20 | 2024-09-06 | 南方医科大学珠江医院 | Auxiliary recognition system for neuropathic pain |
| CN119339950A (en) * | 2024-10-21 | 2025-01-21 | 上海环医智能科技有限公司 | System, method, processor and computer-readable storage medium for implementing ICU patient pain scoring |
-
2025
- 2025-03-12 CN CN202510291539.5A patent/CN119949795B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016124482A1 (en) * | 2015-02-02 | 2016-08-11 | Koninklijke Philips N.V. | Pain management wearable device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN119949795A (en) | 2025-05-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Mao et al. | A health monitoring system based on flexible triboelectric sensors for intelligence medical internet of things and its applications in virtual reality | |
| US20170258336A1 (en) | Automated diagnosis based at least in part on pulse waveforms | |
| US20130281796A1 (en) | Biosensor with exercise amount measuring function and remote medical system thereof | |
| CN204654887U (en) | Based on the Chinese medicine pulse assistant diagnosis system in high in the clouds | |
| CN202477682U (en) | Non-contact intelligent electrocardiogram monitoring system | |
| CN103462596A (en) | Intelligent health mouse for human sign information and real-time monitoring system | |
| WO2017123725A1 (en) | System and method for diagnosis and notification regarding the onset of a stroke | |
| CN102551676A (en) | Personal health monitoring system | |
| Fallahzadeh et al. | Context-aware system design for remote health monitoring: An application to continuous edema assessment | |
| US20230298760A1 (en) | Systems, devices, and methods for determining movement variability, illness and injury prediction and recovery readiness | |
| Hoang et al. | Metrological evaluation of contactless sleep position recognition using an accelerometric smart bed and machine learning | |
| CN115700556A (en) | A gesture recognition method, device and system | |
| CN105708441B (en) | Wearable fingerstall and electrocardiogram and pulse-tracing collection terminal | |
| CN119949795B (en) | An intelligent automatic pain analysis system | |
| Kumar et al. | FPGA interfaced IOT system for smart medical robot monitoring system | |
| CN115054252A (en) | Psychological state evaluating system based on physiological signal complexity asymmetry | |
| Rehman et al. | IoT-based emergency cardiac death risk rescue alert system | |
| Sharma et al. | IoT-based monitoring system for diagnosing adverse drug reactions and enhancing drug compliance in TB patients | |
| CN102949175A (en) | Instant whole body meridian detection system and detection method thereof | |
| Jia et al. | A sensing chair design for home based physiological signs monitoring | |
| Rahman et al. | A field study to capture events of interest (EoI) from living labs using wearables for spatiotemporal monitoring towards a framework of smart health (sHealth) | |
| Syed | Portable smart mini patient monitor for improved healthcare outcomes: patient-based care device | |
| Umamageswaran et al. | Intelligent technique for stress level management for it employees in the work from home environment using iot technology | |
| US12488888B2 (en) | Edge-intelligent IoT-based wearable device for detection of cravings in individuals | |
| TWM468291U (en) | Cloud health management system incorporated with heart rate variability device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |