CN115024707B - A multi-channel pulse feature optimization method based on traditional Chinese medicine principles - Google Patents
A multi-channel pulse feature optimization method based on traditional Chinese medicine principles Download PDFInfo
- Publication number
- CN115024707B CN115024707B CN202210659174.3A CN202210659174A CN115024707B CN 115024707 B CN115024707 B CN 115024707B CN 202210659174 A CN202210659174 A CN 202210659174A CN 115024707 B CN115024707 B CN 115024707B
- Authority
- CN
- China
- Prior art keywords
- pulse
- channel
- feature
- chinese medicine
- traditional chinese
- 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
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/02—Detecting, measuring or recording for evaluating the cardiovascular system, e.g. pulse, heart rate, blood pressure or blood flow
- A61B5/024—Measuring pulse rate or heart rate
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/48—Other medical applications
- A61B5/4854—Diagnosis based on concepts of alternative medicine, e.g. homeopathy or non-orthodox
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Medical Informatics (AREA)
- Physics & Mathematics (AREA)
- Surgery (AREA)
- Veterinary Medicine (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Pathology (AREA)
- Molecular Biology (AREA)
- Public Health (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Physiology (AREA)
- Signal Processing (AREA)
- Artificial Intelligence (AREA)
- Cardiology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Psychiatry (AREA)
- Alternative & Traditional Medicine (AREA)
- Evolutionary Computation (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
本发明提供了一种基于中医原理的脉搏通道优选方法，包括以下步骤：S1、采集左右手寸、关和尺部的脉搏数据，并进行滤波处理和周期划分；S2、对每个通道的特征进行重要性计算，构件多个特征组合；S3、选取各个位置分类性能最高的特征组合，从而构建左右三通道、右手三通道和双手六通道的特征子集；S4、对特征子集进行分类训练。本发明对中医脉诊的通道进行优选，以提高各类疾病的脉搏诊断正确率。
The present invention provides a pulse channel optimization method based on the principle of traditional Chinese medicine, comprising the following steps: S1, collecting pulse data of the cun, guan and chi parts of the left and right hands, and performing filtering processing and period division; S2, calculating the importance of the features of each channel, and constructing multiple feature combinations; S3, selecting the feature combination with the highest classification performance at each position, thereby constructing feature subsets of left and right three channels, right hand three channels and both hands six channels; S4, classifying and training the feature subsets. The present invention optimizes the channels of traditional Chinese medicine pulse diagnosis to improve the pulse diagnosis accuracy of various diseases.
Description
Technical Field
The invention belongs to the technical field of traditional Chinese medicine diagnosis and machine learning, and particularly relates to a multichannel pulse characteristic optimization method based on the traditional Chinese medicine principle.
Background
The pulse diagnosis of traditional Chinese medicine is one of the most representative diagnosis modes in the aspects of 'inspection, smelling, asking and cutting'. The traditional Chinese medicine pulse diagnosis is widely applied to clinic, accurately utilizes left and right hand pulses, and is particularly important to judging physiological and pathological information of patients. The medical practitioner of traditional Chinese medicine usually uses a mode of 'three parts and nine days' to diagnose the states and disease information of different organs of the human body, wherein excellent doctors can judge the life habits, body constitution, past medical history and the like of patients through mastered pulse information. However, the diagnosis method has great influence on the diagnosis result due to different hosts and objects, and the diagnosis of the pulse condition has no quantitative data and lacks objectivity. With the development of sensors and artificial intelligence technology, computer-aided medical treatment has been widely used.
The traditional pulse diagnosis auxiliary diagnosis comprises the steps of feature optimization, namely feature optimization is not performed, the planned features are directly extracted for research and analysis, the feature selection flow in other fields is directly used for selection, the theory that pulses at different positions of traditional Chinese medicine are related to different organs of a human body is not combined, the traditional Chinese medicine considers that the left-hand position is related to heart and heart envelope, the left-hand position is related to liver and gall bladder, the left-hand ruler position is related to kidney and bladder, the right-hand position is related to lung and chest, the right-hand position is related to spleen and stomach, and the right-hand ruler position is related to kidney and large intestine.
The traditional Chinese medicine is a mature medical system, and the objectified pulse diagnosis of the traditional Chinese medicine is also analyzed based on the theoretical method of the traditional Chinese medicine system, so that the development efficiency of the objectified pulse diagnosis system and the diagnosis performance of related diseases can be improved. Therefore, an objective, accurate and rapid diagnosis mode is required to be provided by combining the traditional Chinese medicine pulse diagnosis and the machine learning algorithm.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides the multichannel pulse characteristic optimization method based on the traditional Chinese medicine principle, which has scientific and reasonable design, strong practicability and simple operation, can rapidly and accurately judge the disease type of a patient, and can be popularized and used.
In order to solve the technical problems, the technical scheme adopted by the invention is that the multichannel pulse characteristic optimization method based on the traditional Chinese medicine principle is characterized by comprising the following operation steps:
S1, collecting six channel pulse signals of left and right hand parts, a closing part and a ruler part of a sample person, preprocessing the collected six channel pulse signal data, firstly removing noise and removing baseline drift of the collected pulse signals through a Butterworth filter to obtain high-quality pulse signals, and then searching a single-period pulse signal starting point by using a period segmentation algorithm, and identifying and segmenting single-period pulses according to the starting point;
S2, respectively extracting pulse characteristics of six channel pulse signal data after pretreatment of sample personnel;
s3, data feature primary selection, namely carrying out importance calculation on pulse features of six channels of the sample personnel extracted in the S2 by using a neighbor component analysis method, carrying out primary selection on the pulse features of the six channels according to an importance ranking result, and selecting the pulse feature with high importance of each channel as a basic feature of the channel;
S4, feature optimization, namely constructing a plurality of different feature combinations according to the set basic feature quantity required by constructing feature combinations based on the basic features of the same channel selected in the S3, sending the feature combinations into different machine learning algorithms, comprehensively selecting the feature combinations with good classification effect and the machine learning algorithm, and taking the feature combinations with good classification effect as the obvious pulse features of the channel;
S5, model training, namely selecting the obvious pulse characteristics of a plurality of channels to be combined based on different association degrees between organs corresponding to the specific diseases and the pulse diagnosis positions of the traditional Chinese medicine, constructing a plurality of different characteristic subsets, sending the characteristic subsets into a machine learning algorithm with good classification effect, which is preferably obtained in S4, for classification model training, carrying out classification effect test on the trained classification model, and selecting the characteristic subsets with good comprehensive classification effect, wherein the channels corresponding to the characteristic subsets with good classification effect are used as the characteristic subsets for diagnosing the diseases of the specific diseases.
In the view of traditional Chinese medicine, hypertension is classified into four types, namely liver-yang hyperactivity type, yin deficiency and yang hyperactivity type, phlegm-dampness middle-resistance type and qi stagnation and blood stasis type. Hypertension in most people in life is caused by liver-yang hyperactivity and qi stagnation and blood stasis. After hypertension, it may damage the heart, kidneys and cerebral vessels of humans, with a very high incidence. Therefore, when the pulse of the patient suffering from hypertension is studied, the pulse of the left-hand cun guan chi part contains the information of heart, liver and kidney channels, the difference from the normal person should be larger, and the difference from the normal person should be smaller. When the pulse characteristics of the left and right cun guan chi parts are optimized, the classification accuracy is improved and the data processing complexity is reduced by considering the above. When the specific disease patient is a hypertension patient, based on the close association of the hypertension disease and the heart, focusing on the left dimensional position pulse, constructing feature subsets corresponding to the left three-channel, the right three-channel and the two six-channel, and considering the feature subset matched with the hypertension disease as the feature subset corresponding to the left three-channel after model training.
Compared with the prior art, the invention has the following advantages:
1. The invention has scientific and reasonable design and strong practicability, can effectively improve the speed and accuracy of traditional Chinese medicine diagnosis, and improves the development efficiency of an objective pulse diagnosis system and the diagnosis performance of related diseases.
2. The invention has simple operation, reasonably applies machine learning and model training in the field of traditional Chinese medicine diagnosis, rapidly collects the original pulse diagnosis data, rapidly obtains the diagnosis conclusion, reduces the working intensity and difficulty of the traditional Chinese medicine doctor, and improves the accuracy.
The invention is described in further detail below with reference to the drawings and examples.
Drawings
Fig. 1 is a functional block diagram of the present invention.
FIG. 2 is a schematic diagram of the operational steps of the present invention.
Fig. 3 is a schematic diagram showing selection of importance of basic features by neighbor component analysis in the present invention.
FIG. 4 is a schematic diagram of the present invention for feeding feature combinations into a machine learning algorithm for evaluation ranking.
Fig. 5 is a schematic view of viscera corresponding to different pulse positions of the left and right hands in the present invention.
Detailed Description
As shown in fig. 1 and fig. 2, the multichannel pulse feature optimization method based on the principle of traditional Chinese medicine according to the embodiment of the invention is applied to diagnosis of hypertension, and comprises the following specific operation steps:
S1, collecting six channel pulse signals of a left hand inch part, a right hand inch part, a closing part and a ruler part of a sample person by using multi-channel piezoelectric sensor pulse collecting equipment, wherein the sample person comprises healthy personnel and specific disease patients, filtering the collected six channel pulse signal data base Yu Bate Wolth filter, eliminating noise, removing baseline drift and obtaining high-quality pulse signals. Then, a cycle segmentation algorithm is used to determine the starting point of the monocycle pulse, and the monocycle pulse is identified and extracted according to the starting point.
As shown in fig. 5, according to the principle of traditional Chinese medicine, the left-hand cun portion, the close portion and the ruler portion correspond to the heart, the liver and the kidney in sequence, the right-hand cun portion, the close portion and the ruler portion correspond to the lung, the spleen and the life in sequence, the patient with the specific disease is a patient with hypertension, and the six parts are important to pay attention to the pulse diagnosis of the patient with hypertension in traditional Chinese medicine.
S2, respectively extracting pulse characteristics of six channel pulse signal data after pretreatment of sample personnel;
And S3, data feature initial selection, namely, as shown in FIG 3, carrying out importance calculation on pulse features of six channels of the sample personnel extracted in the S2 by using a neighbor component analysis method, and extracting the first six features according to an importance ranking result. Wherein the left hand position is characterized by mean, main amplitude, counterpulsation amplitude, main time, counterpulsation wave time, main to counterpulsation wave time, the left hand position is characterized by variance, main amplitude, counterpulsation wave time, main to counterpulsation wave time, the left hand ruler is characterized by main amplitude, counterpulsation wave amplitude, main time, counterpulsation wave time, main to counterpulsation wave time, counterpulsation to counterpulsation wave time, right hand position is characterized by wavelet packet energy characteristic, variance, counterpulsation wave amplitude, counterpulsation wave time, right hand position is characterized by variance, main amplitude, counterpulsation wave amplitude, wavelet packet energy characteristic, counterpulsation wave time, main to counterpulsation wave time, right hand ruler is characterized by variance, main wave amplitude, counterpulsation time, counterpulsation wave time. The pulse characteristics of the six channels are initially selected and combined, and the pulse characteristics with high importance of each channel are selected as basic characteristics of the channel;
And S4, preferably, constructing a plurality of different feature combinations according to the number of 3 basic features required for constructing each channel feature combination and the hypertension disease setting based on the basic features of the same channel selected in the step S3, and sending the feature combinations to different machine learning algorithms as shown in the figure 4. The method comprises the steps of randomly extracting 20% of data from a data set, training and testing the data by using a five-fold cross-validation method, analyzing and selecting corresponding feature combinations according to classification effects, preferably using KNN, SVM, DT three machine learning algorithms, comprehensively selecting the feature combinations with good classification effects on healthy people/hypertensive patients and the machine learning algorithms, taking the feature combinations with good classification effects as obvious pulse features of a channel, and when DT is used, using the feature combinations of main wave time, counterpulsation wave time and main wave to counterpulsation wave time at a left-hand position, using the feature combinations of variance, counterpulsation wave amplitude and main wave to counterpulsation wave time at a left-hand position, and using the feature combinations of main wave time, counterpulsation wave time and main wave to counterpulsation wave time at a left-hand position. When using an SVM, the left-hand position uses a feature combination of main amplitude, and main to counterpulsation wave time, the left-hand position uses a feature combination of variance, main amplitude, and counterpulsation wave amplitude, and the left-hand position uses a feature combination of variance, main amplitude, and counterpulsation wave amplitude. When KNN is used, the left-hand position uses a feature combination of main wave amplitude, dicrotic wave amplitude, main to dicrotic wave time, the left-hand off position uses a feature combination of variance, dicrotic wave amplitude, dicrotic wave time, and the left-hand ruler position uses a feature combination of variance, main wave amplitude, main to dicrotic wave. According to the result, constructing a characteristic subset of a left-hand three-channel, adopting a healthy person/hypertension patient data set, training and testing a model by using five-fold cross validation, and performing such as
As shown in table 1, the classification effect of the KNN algorithm on healthy people\hypertension patients is found to be better;
Table 1 model Performance comparison table for feature subset training for different algorithms
S5, model training, namely, based on different association degrees between organs corresponding to the specific diseases and the pulse diagnosis positions of the traditional Chinese medicine, selecting the obvious pulse characteristics of a plurality of channels to be combined, constructing a plurality of different characteristic subsets, sending the characteristic subsets into a machine learning algorithm with good classification effect, which is preferably obtained in S4, training and testing a model by adopting a five-fold cross-validation mode by adopting data of a data set constructed by 35 healthy people and 11 hypertensive patients, and selecting the characteristic subsets with good comprehensive classification effect of the healthy people/the hypertensive patients according to the result, wherein the channels corresponding to the characteristic subsets with good classification effect are used as the characteristic subsets for disease diagnosis of the specific diseases.
S5, specifically, a classification model constructed by the characteristics in the six channels of the two hands, the three channels of the left hand and the three channels of the right hand is obtained, and the three classification models are compared and optimized, as shown in the table 2:
Table 2 model performance lookup table for feature subset training for three channels
Aiming at the hypertension, according to the evaluation result, the classification model of the feature subset constructed by adopting the six-channel features and the classification model of the left-hand three-channel are better than the right-hand three-channel. The number of the three channels of the left hand is less, and the trained model also has good classification effect, so that the method is suitable for actual diagnosis scenes. It was concluded that the feature subset adapted to hypertension disease is the feature subset corresponding to the left hand three channels.
The above description is only of the preferred embodiments of the present invention, and is not intended to limit the present invention. Any simple modification, variation and equivalent variation of the above embodiments according to the technical substance of the invention still fall within the scope of the technical solution of the invention.
Claims (4)
1. The multichannel pulse characteristic optimization method based on the traditional Chinese medicine principle is characterized by comprising the following operation steps of:
S1, collecting six channel pulse signals of left and right hand parts, a closing part and a ruler part of a sample person, preprocessing the collected six channel pulse signal data, firstly removing noise and removing baseline drift of the collected pulse signals through a Butterworth filter to obtain high-quality pulse signals, and then searching a single-period pulse signal starting point by using a period segmentation algorithm, and identifying and segmenting single-period pulses according to the starting point;
S2, respectively extracting pulse characteristics of six channel pulse signal data after pretreatment of sample personnel;
S3, data feature primary selection, namely carrying out importance calculation on the pulse features of the six channels of the sample personnel extracted in the S2, carrying out primary selection on the pulse features of the six channels according to an importance ranking result, and selecting the pulse feature with high importance of each channel as a basic feature of the channel;
S4, feature optimization, namely constructing a plurality of different feature combinations according to the set basic feature quantity required by constructing feature combinations based on the basic features of the same channel selected in the S3, sending the feature combinations into different machine learning algorithms, comprehensively selecting the feature combinations with good classification effect and the machine learning algorithm, and taking the feature combinations with good classification effect as the obvious pulse features of the channel;
S5, model training, namely selecting the obvious pulse characteristics of a plurality of channels to be combined based on different association degrees between organs corresponding to the specific diseases and the pulse diagnosis positions of the traditional Chinese medicine, constructing a plurality of different characteristic subsets, sending the characteristic subsets into a machine learning algorithm with good classification effect, which is preferably obtained in S4, for classification model training, carrying out classification effect test on the trained classification model, and selecting the characteristic subsets with good comprehensive classification effect, wherein the channels corresponding to the characteristic subsets with good classification effect are used as the characteristic subsets for diagnosing the diseases of the specific diseases.
2. The method for optimizing pulse characteristics based on the principle of traditional Chinese medicine according to claim 1, wherein the pulse signals of healthy people and patients with specific diseases are acquired by using the multi-channel piezoelectric sensor pulse acquisition equipment in the step S1.
3. The multi-channel pulse characteristic optimization method based on the principle of traditional Chinese medicine according to claim 1, wherein the importance calculation is performed on pulse signal characteristics of six channels by using a neighbor component analysis method in S3.
4. The method according to claim 1, wherein when the patient with the specific disease is a hypertensive patient, based on the close association of the hypertensive disease with the heart, focusing on the left-hand position pulse, a feature subset corresponding to the left-hand three-channel, the right-hand three-channel and the two-hand six-channel is constructed, and the feature subset which is considered to be matched with the hypertensive disease after model training is the feature subset corresponding to the left-hand three-channel.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210659174.3A CN115024707B (en) | 2022-06-13 | 2022-06-13 | A multi-channel pulse feature optimization method based on traditional Chinese medicine principles |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210659174.3A CN115024707B (en) | 2022-06-13 | 2022-06-13 | A multi-channel pulse feature optimization method based on traditional Chinese medicine principles |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115024707A CN115024707A (en) | 2022-09-09 |
| CN115024707B true CN115024707B (en) | 2025-03-14 |
Family
ID=83125387
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210659174.3A Active CN115024707B (en) | 2022-06-13 | 2022-06-13 | A multi-channel pulse feature optimization method based on traditional Chinese medicine principles |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115024707B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115778343B (en) * | 2022-12-12 | 2025-03-18 | 西安邮电大学 | A method, device and equipment for identifying hypertension based on pulse cycle characteristics |
| CN115736850B (en) * | 2023-01-05 | 2023-04-21 | 南京大经中医药信息技术有限公司 | Pulse data classification system and classification method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110752032A (en) * | 2019-12-02 | 2020-02-04 | 山东浪潮人工智能研究院有限公司 | A traditional Chinese medicine diagnosis method based on convolutional neural network and laser vibration measurement |
| CN110897616A (en) * | 2019-12-05 | 2020-03-24 | 中国科学院微电子研究所 | Customia pulse position detection device and method |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3170449B1 (en) * | 2015-11-20 | 2024-05-22 | Tata Consultancy Services Limited | Device to detect diabetes in a person using pulse palpation signal |
| US20170245767A1 (en) * | 2016-02-25 | 2017-08-31 | Echo Labs, Inc. | Systems and methods for modified pulse transit time measurement |
| US20170303805A1 (en) * | 2016-04-21 | 2017-10-26 | Mona Boudreaux | Method and Apparatus for Simulating the Wrist Pulse Patterns for Pulse Diagnosis |
| EP3785174B1 (en) * | 2018-04-27 | 2022-12-14 | Medtronic Ardian Luxembourg S.à.r.l. | Identifying patients suited for renal denervation therapy |
| KR20190129342A (en) * | 2018-05-10 | 2019-11-20 | (주)에이치피케이 | Pulse measurement apparatus based on artificial intelligence and method of measuring pulse of thereof |
| WO2020095326A1 (en) * | 2018-11-11 | 2020-05-14 | Houronearth Creative Solutions Pvt Ltd | System and method for classifying a pulse morphology of a user |
| CN111938597A (en) * | 2020-08-17 | 2020-11-17 | 西安邮电大学 | A multi-modal traditional Chinese medicine pulse collection system |
| CN112274126B (en) * | 2020-10-28 | 2022-11-29 | 河北工业大学 | A non-invasive continuous blood pressure detection method and device based on multiple pulse waves |
| CN114145722B (en) * | 2021-12-07 | 2023-08-01 | 西安邮电大学 | Pulse pathological feature mining method for pancreatitis patients |
-
2022
- 2022-06-13 CN CN202210659174.3A patent/CN115024707B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110752032A (en) * | 2019-12-02 | 2020-02-04 | 山东浪潮人工智能研究院有限公司 | A traditional Chinese medicine diagnosis method based on convolutional neural network and laser vibration measurement |
| CN110897616A (en) * | 2019-12-05 | 2020-03-24 | 中国科学院微电子研究所 | Customia pulse position detection device and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115024707A (en) | 2022-09-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Talukder et al. | An efficient deep learning model to categorize brain tumor using reconstruction and fine-tuning | |
| CN110368300B (en) | Intelligent acupuncture diagnosis and treatment system | |
| CN111759345B (en) | Heart valve abnormality analysis method, system and device based on convolutional neural network | |
| CN115024707B (en) | A multi-channel pulse feature optimization method based on traditional Chinese medicine principles | |
| CN113362944B (en) | Assessment method of functional dyspepsia and needling curative effect prediction model based on machine learning | |
| Ji et al. | An intelligent diagnostic method of ECG signal based on Markov transition field and a ResNet | |
| Jiang et al. | Features fusion of multichannel wrist pulse signal based on KL-MGDCCA and decision level combination | |
| Zhang et al. | Application of deep neural network for congestive heart failure detection using ECG signals | |
| Wang et al. | A new deep learning model for assisted diagnosis on electrocardiogram | |
| Chen et al. | Artificial intelligence for heart sound classification: A review | |
| CN118053192B (en) | Adenoids hypertrophy recognition system based on multi-angle face images | |
| Tung et al. | Multi-lead ECG classification via an information-based attention convolutional neural network | |
| CN119014878A (en) | A CNN-GRU ECG signal classification method based on attention mechanism | |
| Zhang et al. | Missing-view completion for fatty liver disease detection | |
| CN117562583A (en) | Artificial intelligence-assisted cardiac function detection system and method | |
| CN112767374A (en) | Alzheimer disease focus region semantic segmentation algorithm based on MRI | |
| Khatar et al. | GAF-GradCAM: Guided dynamic weighted fusion of temporal and frequency GAF 2D matrices for ECG-based arrhythmia detection using deep learning | |
| CN120147266B (en) | Method and device for analyzing spinal arthritis based on learning model | |
| Ranta et al. | Principal component analysis and interpretation of bowel sounds | |
| Urooj et al. | Computer-aided system for pneumothorax detection through chest X-ray images using convolutional neural network | |
| CN113628174A (en) | Subjective and objective ultrasonic medical image quality evaluation method and system | |
| CN111265234A (en) | Method and system for judging properties of lung mediastinal lymph nodes | |
| Kusuma et al. | Bidlnet: An integrated deep learning model for ecg-based heart disease diagnosis | |
| Krupadanam et al. | Analysis of convolutional neural network algorithm for cardiac diagnosis compared with accuracy of artificial neural network algorithm | |
| CN117649590B (en) | Myocardial infarction identification method based on position backtracking deep learning network |
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 |