CN115500822A - Parkinson detection system and equipment based on wrist tremor signal - Google Patents
Parkinson detection system and equipment based on wrist tremor signal Download PDFInfo
- Publication number
- CN115500822A CN115500822A CN202211294802.9A CN202211294802A CN115500822A CN 115500822 A CN115500822 A CN 115500822A CN 202211294802 A CN202211294802 A CN 202211294802A CN 115500822 A CN115500822 A CN 115500822A
- Authority
- CN
- China
- Prior art keywords
- data
- signal
- value
- wrist
- module
- 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.)
- Pending
Links
- 210000000707 wrist Anatomy 0.000 title claims abstract description 97
- 206010044565 Tremor Diseases 0.000 title claims abstract description 88
- 238000001514 detection method Methods 0.000 title claims abstract description 49
- 238000000034 method Methods 0.000 claims abstract description 40
- 230000002159 abnormal effect Effects 0.000 claims abstract description 29
- 238000012549 training Methods 0.000 claims abstract description 29
- 238000012937 correction Methods 0.000 claims abstract description 18
- 238000007781 pre-processing Methods 0.000 claims abstract description 14
- 208000018737 Parkinson disease Diseases 0.000 claims abstract description 13
- 238000012360 testing method Methods 0.000 claims abstract description 13
- 238000010801 machine learning Methods 0.000 claims abstract description 11
- 238000011897 real-time detection Methods 0.000 claims abstract description 6
- 241000254158 Lampyridae Species 0.000 claims description 52
- 238000004422 calculation algorithm Methods 0.000 claims description 28
- 238000000354 decomposition reaction Methods 0.000 claims description 27
- 201000010099 disease Diseases 0.000 claims description 27
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims description 27
- 239000003814 drug Substances 0.000 claims description 24
- 230000008569 process Effects 0.000 claims description 19
- 229940079593 drug Drugs 0.000 claims description 16
- 230000001133 acceleration Effects 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 14
- 230000003595 spectral effect Effects 0.000 claims description 14
- 238000013499 data model Methods 0.000 claims description 11
- 238000012417 linear regression Methods 0.000 claims description 11
- 230000002194 synthesizing effect Effects 0.000 claims description 9
- 230000009467 reduction Effects 0.000 claims description 8
- 208000027089 Parkinsonian disease Diseases 0.000 claims description 7
- 206010034010 Parkinsonism Diseases 0.000 claims description 7
- 230000006870 function Effects 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 7
- 238000000605 extraction Methods 0.000 claims description 6
- 238000001228 spectrum Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 5
- 230000003068 static effect Effects 0.000 claims description 4
- 238000005314 correlation function Methods 0.000 claims description 3
- 238000000513 principal component analysis Methods 0.000 claims description 3
- 208000025174 PANDAS Diseases 0.000 claims description 2
- 208000021155 Paediatric autoimmune neuropsychiatric disorders associated with streptococcal infection Diseases 0.000 claims description 2
- 240000000220 Panda oleosa Species 0.000 claims description 2
- 235000016496 Panda oleosa Nutrition 0.000 claims description 2
- 238000010521 absorption reaction Methods 0.000 claims description 2
- 238000004364 calculation method Methods 0.000 claims description 2
- GNBHRKFJIUUOQI-UHFFFAOYSA-N fluorescein Chemical compound O1C(=O)C2=CC=CC=C2C21C1=CC=C(O)C=C1OC1=CC(O)=CC=C21 GNBHRKFJIUUOQI-UHFFFAOYSA-N 0.000 claims description 2
- 230000031700 light absorption Effects 0.000 claims description 2
- 238000012847 principal component analysis method Methods 0.000 claims description 2
- 238000012800 visualization Methods 0.000 claims description 2
- 238000013480 data collection Methods 0.000 claims 1
- 238000012216 screening Methods 0.000 claims 1
- 238000003745 diagnosis Methods 0.000 abstract description 4
- 230000000694 effects Effects 0.000 description 5
- 238000011156 evaluation Methods 0.000 description 2
- 230000000284 resting effect Effects 0.000 description 2
- 208000024891 symptom Diseases 0.000 description 2
- 206010006100 Bradykinesia Diseases 0.000 description 1
- 208000006083 Hypokinesia Diseases 0.000 description 1
- WTDRDQBEARUVNC-LURJTMIESA-N L-DOPA Chemical compound OC(=O)[C@@H](N)CC1=CC=C(O)C(O)=C1 WTDRDQBEARUVNC-LURJTMIESA-N 0.000 description 1
- WTDRDQBEARUVNC-UHFFFAOYSA-N L-Dopa Natural products OC(=O)C(N)CC1=CC=C(O)C(O)=C1 WTDRDQBEARUVNC-UHFFFAOYSA-N 0.000 description 1
- 208000002740 Muscle Rigidity Diseases 0.000 description 1
- 208000012902 Nervous system disease Diseases 0.000 description 1
- 206010071390 Resting tremor Diseases 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000003759 clinical diagnosis Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 230000005021 gait Effects 0.000 description 1
- 229960004502 levodopa Drugs 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
- A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
- A61B5/1101—Detecting tremor
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/40—Detecting, measuring or recording for evaluating the nervous system
- A61B5/4076—Diagnosing or monitoring particular conditions of the nervous system
- A61B5/4082—Diagnosing or monitoring movement diseases, e.g. Parkinson, Huntington or Tourette
-
- 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
- A61B5/7225—Details of analog processing, e.g. isolation amplifier, gain or sensitivity adjustment, filtering, baseline or drift compensation
-
- 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
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/74—Details of notification to user or communication with user or patient ; user input means
- A61B5/746—Alarms related to a physiological condition, e.g. details of setting alarm thresholds or avoiding false alarms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Animal Behavior & Ethology (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Biophysics (AREA)
- Physiology (AREA)
- Surgery (AREA)
- Pathology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Artificial Intelligence (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Psychiatry (AREA)
- Neurology (AREA)
- Neurosurgery (AREA)
- Evolutionary Computation (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Fuzzy Systems (AREA)
- Developmental Disabilities (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Dentistry (AREA)
- Power Engineering (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a Parkinson detection system and device based on a wrist tremor signal, and relates to a Parkinson detection system and device based on a wrist tremor signal. The invention aims to solve the problem that the existing method has low diagnosis accuracy on the Parkinson's disease. The system comprises: the device comprises a data acquisition module, a data division module, a data detection and correction module, a data preprocessing module, a characteristic data set acquisition module, a training set and test set acquisition module, a machine learning module and a detection module; the data acquisition module is used for acquiring a tremor signal of the wrist when the patient is attacked and a normal signal of the wrist when the patient is not attacked; the data dividing module is used for dividing the data of the wrist tremor signal acquired by the data acquisition module in case of illness and the normal signal of the wrist in case of no illness; the data detection and correction module is used for carrying out data real-time detection on the divided data and realizing correction on missing values and abnormal values. The invention is used for the technical field of wrist tremor signal identification.
Description
Technical Field
The invention relates to a Parkinson detection system and device based on a wrist tremor signal. The invention relates to the technical field of wrist tremor signal identification.
Background
Parkinson is a very common nervous system disease, the number of people with Parkinson in China is about 300 to ten thousand, the Parkinson prevalence rate of the old people over 65 years old is about 1.7 to 5 percent, and the Parkinson is the first disease killer of the middle-aged and old people in the world. Common symptoms of Parkinson include resting tremor, muscular rigidity, gait abnormality, bradykinesia and the like, but early onset is more hidden, diagnosis is mainly based on clinical manifestations and treatment effect on levogyration levodopa, biological indexes are lacked, energy consumption is high, and clinical diagnosis efficiency is low. Therefore, the Parkinson robust and efficient identification based on the wrist tremor signal has great practical significance for the medical diagnosis and the later treatment effect evaluation of the Parkinson patients.
Disclosure of Invention
The invention aims to solve the problem that the existing method is low in accuracy rate of diagnosing Parkinson's diseases, and provides a Parkinson detection system and equipment based on wrist tremor signals.
A Parkinson detection system based on wrist tremor signals comprises:
the device comprises a data acquisition module, a data division module, a data detection and correction module, a data preprocessing module, a characteristic data set acquisition module, a training set and test set acquisition module, a machine learning module and a detection module;
the data acquisition module is used for acquiring a tremor signal of the wrist when the patient is attacked and a normal signal of the wrist when the patient is not attacked;
the data dividing module is used for dividing data of the wrist tremor signal acquired by the data acquisition module in case of illness and the normal wrist signal acquired by the data acquisition module in case of no illness;
the data detection and correction module is used for carrying out data real-time detection on the divided data to realize correction on missing values and abnormal values;
the data preprocessing module is used for carrying out filtering and synthesizing acceleration data preprocessing operations on the data after the missing values and the abnormal values are corrected;
the characteristic data set acquisition module is used for carrying out characteristic extraction and characteristic dimension reduction on the preprocessed signals to obtain a characteristic data set;
the training set and test set acquisition module is used for dividing the characteristic data set into a training set and a test set;
the machine learning module is used for obtaining a training data model, performing machine learning training on a training set by using a K nearest neighbor algorithm and verifying by using a test set to finally obtain a trained training data model;
the detection module is used for inputting the wrist tremor signal to be detected into the trained training data model, classifying the wrist tremor signal to be detected, and if the wrist tremor signal to be detected is a normal signal, continuously acquiring the wrist tremor signal; and if the signal is an abnormal signal, constructing a multi-dimensional linear regression equation by using the frequency characteristics, and evaluating the Parkinson disease degree.
A Parkinson detection device based on wrist tremor signals comprises a processor and a memory, wherein at least one instruction is stored in the memory, and the at least one instruction is loaded and executed by the processor to realize the Parkinson detection device based on the wrist tremor signals.
The beneficial effects of the invention are as follows:
1. the patient is monitored in real time through an efficient identification algorithm, so that the disease condition can be identified in time and early warning is performed.
2. Through recording relevant physiological indexes during disease attack, doctors can diagnose patients and monitor treatment effects conveniently.
3. The tremor signal is analyzed during the onset of the disease, so that the severity of the onset of the disease of the patient can be accurately judged, and the dosage of the medicine can be determined.
Drawings
FIG. 1 is a flow chart of a Parkinson's detection method according to the invention;
FIG. 2 is a flow chart of a firefly algorithm optimized variational modal decomposition denoising method.
Detailed Description
The first embodiment is as follows: this embodiment a parkinson detecting system based on wrist tremor signal includes:
the system comprises a data acquisition module, a data division module, a data detection and correction module, a data preprocessing module, a characteristic data set acquisition module, a training set and test set acquisition module, a machine learning module and a detection module;
the data acquisition module is used for acquiring a wrist tremor signal when a disease occurs and a wrist normal signal when the disease does not occur;
the data dividing module is used for dividing data of the wrist tremor signal acquired by the data acquisition module in case of illness and the normal wrist signal acquired by the data acquisition module in case of no illness;
the data detection and correction module is used for carrying out data real-time detection on the divided data, and constructing a correlation function based on a multiple filling method and the basic principle of the Chebyshev inequality so as to realize correction on missing values and abnormal values;
the data preprocessing module is used for performing data preprocessing operations such as filtering and synthesizing acceleration on the data after the missing values and the abnormal values are corrected;
the characteristic data set acquisition module is used for carrying out characteristic extraction and characteristic dimension reduction on the preprocessed signals, determining a characteristic effect and obtaining a characteristic data set;
the training set and test set acquisition module is used for dividing the characteristic data set into a training set and a test set to obtain training learning data in a uniform format;
the machine learning module is used for obtaining a training data model, performing machine learning training on a training set by using a K nearest neighbor algorithm and verifying by using a test set to finally obtain a trained training data model;
the detection module is used for inputting the to-be-detected wrist tremor signal into the trained training data model, classifying the to-be-detected wrist tremor signal, and if the to-be-detected wrist tremor signal is a normal signal, continuously acquiring the wrist tremor signal; and if the signal is an abnormal signal, constructing a multi-dimensional linear regression equation by using the frequency characteristics, and evaluating the Parkinson disease degree.
The second embodiment is as follows: the difference between the embodiment and the first embodiment is that the data acquisition module is used for acquiring a tremor signal of the wrist when the patient is attacked and a normal signal of the wrist when the patient is not attacked; wrist tremor signal and normal signal of wrist when not morbidity are all obtained by triaxial acceleration sensor during morbidity, wrist tremor signal and normal signal of wrist when not morbidity that triaxial acceleration sensor obtained are all triaxial acceleration signal during morbidity.
Other steps and parameters are the same as those in the first embodiment.
The third concrete implementation mode: the difference between the first embodiment and the second embodiment is that the data dividing module is used for dividing the data of the wrist tremor signal acquired by the data acquisition module in case of illness and the normal wrist signal acquired by the data acquisition module in case of no illness; the specific process is as follows:
calculating the time domain characteristics of the collected wrist tremor signals during the onset of the disease, wherein the time domain characteristics comprise the characteristics of the peak value of the signals during the onset of the Parkinson's disease, the zero crossing times of the signals during the onset of the Parkinson's disease and the like;
calculating the frequency domain characteristics of the acquired wrist tremor signals during onset, wherein the frequency domain characteristics comprise main frequency of the signals during onset of Parkinson, power spectral density and energy spectral density of the signals during onset of Parkinson, weighted power spectral density obtained from the power spectral density, maximum power spectral value obtained from the power spectral density and the like;
the normal signal data of the wrist when the wrist is not attacked is divided into two states of a motion state interference segment and a static state stable segment, wherein the motion state interference segment and the static state stable segment respectively account for 50%.
The tremor signals are collected when a person suffers from diseases, and the normal signals are collected when the person does not suffer from diseases (including the motion state signals when the person moves when the person does not suffer from diseases and the resting state signals when the person is resting when the person does not suffer from diseases).
Other steps and parameters are the same as those in the first or second embodiment.
The fourth concrete implementation mode: the difference between the first embodiment and the third embodiment is that the data detection and correction module is used for carrying out data real-time detection on the divided data, and constructing a correlation function based on a multiple filling method and the basic principle of Chebyshev inequality so as to realize correction on missing values and abnormal values;
the specific process is as follows:
step 301, a df.isnull function in a pandas library is called in python, a missing value NAN in the divided data is searched, the missing value NAN is the ith data in the divided data, data before and after the missing value NAN, namely the (i-1) th data and the (i + 1) th data are selected, a section formed by the (i-1) th data to the (i + 1) th data is divided into m parts to generate m data, the generated m data are respectively subjected to Bayesian estimation to obtain the probability corresponding to each data, and the maximum value in the m probability values is taken as the estimated value of the missing value NAN;
step 302, detecting abnormal values contained in the divided data by a 3 sigma detection method, removing the abnormal values, and correcting the abnormal values according to the step 301;
the 3 sigma detection method comprises the following steps:
p(|X-E(X)|≥ε)≤D(X)/ε 2 ≈0.11
wherein: epsilon is the length of the interval around the mean value, epsilon =3 sigma, sigma is the standard deviation, D (X) = sigma is the standard deviation, X is the sample value, E (X) is the mean value of the sample, p () is the probability of the occurrence of an abnormal value;
when E (X) -epsilon is less than or equal to X and less than or equal to E (X) + epsilon, X is a normal value;
when X > E (X) + ε or X < E (X) - ε, X is an abnormal value.
Other steps and parameters are the same as those in the first or second embodiment.
The fifth concrete implementation mode is as follows: the present embodiment is different from the first to the fourth embodiments in that the data preprocessing module is configured to perform data preprocessing operations such as filtering and synthesizing acceleration on the data after the missing value and the abnormal value are corrected; the specific process is as follows:
step 401, performing FA-VMD-WTD denoising processing on the data after the abnormal value and the missing value are corrected to obtain denoised signal data;
the FA-VMD-WTD method combines the variation modal decomposition denoising and the wavelet hard threshold processing after the firefly optimization algorithm is optimized;
and the FA is a firefly algorithm, and determines an important parameter decomposition layer number K value and a penalty factor alpha value for the VMD through the FA. Obtaining the optimal parameter K value and the optimal parameter alpha value according to the judgment basis of the signal-to-noise ratio and the root-mean-square error;
the VMD is a variation modal decomposition, which can be divided into two aspects of constructing a variation problem and solving the variation problem, and the optimal solution is iteratively searched to determine the component center frequency and the bandwidth of each component, so that the non-stationarity of a complex time sequence can be reduced;
the WTD is wavelet hard threshold denoising; the wavelet hard threshold denoising is one of wavelet threshold denoising, wherein the wavelet threshold denoising is to select a proper threshold after wavelet transformation is carried out on an original signal, so that a wavelet coefficient which is larger than the threshold is considered to be reserved, and a wavelet coefficient which is smaller than the threshold is made to be zero to achieve the purpose of denoising;
according to the parkinson detection method provided by the invention, in the step 401, the FA-VMD-WTD method is a denoising method combining variational modal decomposition denoising after firefly optimization algorithm optimization with wavelet hard threshold processing, and the optimization flow is shown in fig. 2.
Step 402, synthesizing the denoised signal data into acceleration (independent of 401):
wherein: a. The x 、A y 、A z The de-noised acceleration signals of the x axis, the y axis and the z axis are respectively.
Other steps and parameters are the same as in one of the first to fourth embodiments.
The sixth specific implementation mode: the difference between this embodiment and the first to fifth embodiments is that, in step 401, the FA-VMD-WTD denoising processing is performed on the data after the correction of the abnormal value and the missing value, so as to obtain denoised signal data;
the FA-VMD-WTD method combines the variable mode decomposition denoising and the wavelet hard threshold processing after the firefly optimization algorithm is optimized;
the FA is a firefly algorithm; the VMD is a variational modal decomposition; the WTD is wavelet hard threshold denoising;
the specific process is as follows:
step 4011, solving optimal variational modal decomposition method parameters K and α based on a firefly algorithm FA; the specific process is as follows:
s1, setting value ranges of K and alpha in a variational modal decomposition VMD algorithm according to key parameters required by the variational modal decomposition;
k is the modal number of the decomposition target, and alpha is a secondary penalty factor;
k is more than or equal to 3 and less than or equal to 20;
alpha is more than or equal to 500 and less than or equal to 2500;
s2, initializing basic parameters of a firefly algorithm FA, namely problem dimension, group size and maximum attraction degree beta 0 The light absorption coefficient (0.0001), the step factor (0.97), the iteration times and the value ranges of K and alpha;
k is the modal number of the decomposition target, and alpha is a secondary penalty factor;
k is more than or equal to 3 and less than or equal to 20;
alpha is more than or equal to 500 and less than or equal to 2500;
s3, randomly initializing the K value and the alpha value of the ith firefly as initial positions in space, wherein i =1,2, \ 8230;, 25;
s4, different fireflies have different K values and alpha values, and the fitness function of the ith firefly is calculated to serve as the fluorescein value l of the ith firefly at the time t i (t) as the maximum fluorescence intensity I of the ith firefly 0 ;
S5, calculating the relative brightness I of the ith firefly in the population r And attraction degree beta, determining the maximum I r Value, I-th firefly to maximum I r (determining firefly) value moving direction;
r represents the distance between two fireflies, gamma represents the light intensity absorption coefficient, I 0 Denotes the maximum fluorescence intensity, I r Denotes the relative brightness of firefly, beta denotes the attraction degree, beta 0 The maximum attraction is represented and the value is 1.
S6, based on beta and I in S5 r (each firefly has moved to a new location) the spatial location of the ith firefly, i.e., the new K value and α value, is updated by the formula:
said X i (t + 1) represents the updated spatial position of the ith firefly, X i (t) represents the spatial position of the ith firefly, X j (t) represents the spatial position of the jth firefly,denotes the step size factor, rand denotes [0,1 ]]Uniformly distributed random factors are subjected to;
based on the new K value and the alpha value, re-executing S4 to obtain the maximum fluorescence brightness;
randomly moving to the position of the firefly (at the optimal position) with stronger fluorescence brightness, and recalculating the brightness of the firefly according to the updated firefly position, wherein the updated brightness of the firefly is enhanced;
s7, judging whether the set maximum iteration number is reached, if so, outputting the brightness and the position of the optimal firefly, namely the optimal K value and the optimal alpha value (outputting the optimal firefly luminous intensity); if not, based on the new K value and the new alpha value, executing S4-S6 again until the maximum iteration times is reached, and outputting the optimal firefly position, namely the optimal K value and the optimal alpha value.
Preferably, in the process of optimizing the denoising method by using the firefly algorithm in step 401, the signal-to-noise ratio is used as a fitness function to solve the optimal parameters K and α in the variational modal decomposition:
wherein: n is the signal length, f (t) i ) In order to obtain a signal containing noise,f is a fitness function for the noise-reduced signal;
step 4012, performing variational modal decomposition on the collected wrist tremor signal data according to the K value and the alpha value obtained in step 4011, and determining a dominant signal component, an interference signal component and a noise signal component; the specific process is as follows:
introducing a variational modal decomposition module vmdpy into pyson, and introducing a K value and an alpha value obtained by an FA algorithm and the collected wrist tremor signal into the variational modal decomposition module to obtain K decomposed signals;
calculating the correlation between each decomposed signal and the collected wrist tremor signal (original signal), wherein a component with the correlation of 0.5-1.0 is taken as a dominant signal component, a component with the correlation of 0.2-0.5 is taken as an interference signal component, and a component with the correlation of 0-0.2 is taken as a noise signal component;
step 4013, performing wavelet denoising processing on the dominant signal component, the interference signal component and the noise signal component determined in step 4012; the specific process is as follows:
for noise signal components, directly discarding;
for the dominant signal component, directly preserving;
for interference signal components, denoising by adopting a wavelet hard threshold;
step 4014, directly synthesizing the components subjected to wavelet denoising processing in step 4013 to obtain denoised signal data.
Other steps and parameters are the same as in one of the first to fifth embodiments.
The seventh concrete implementation mode: the difference between the present embodiment and one of the first to sixth specific embodiments is that the feature data set acquisition module is configured to perform feature extraction and feature dimension reduction on the signal after noise removal by the FA-VMD-WTD method, and determine a feature effect to obtain a feature data set; the specific process is as follows:
extracting characteristics (such as time domain characteristics and frequency domain characteristics) of the signals after noise is removed by the FA-VMD-WTD method;
the method comprises the steps that signals have more than ten frequency domain features, redundant features can influence algorithm efficiency and precision, a principal component analysis algorithm is a data dimension reduction algorithm which is most widely used, the importance of the principal component analysis algorithm on extracted features is calculated, the importance of the features is screened out by using a principal component analysis method, feature dimension reduction is achieved on the features with the calculated importance through TSNE visualization, and a feature data set is obtained.
Other steps and parameters are the same as those in one of the first to sixth embodiments.
The specific implementation mode is eight: the difference between the first embodiment and the seventh embodiment is that the detection module is used for inputting the to-be-detected wrist tremor signal into the trained training data model, classifying the to-be-detected wrist tremor signal, and if the to-be-detected wrist tremor signal is a normal signal, continuously acquiring the wrist tremor signal; if the signal is an abnormal signal, a multi-dimensional linear regression equation is constructed by using the frequency characteristics, and the Parkinson disease degree is evaluated; the specific process is as follows:
selecting weighted power spectrum density, main frequency, maximum value of power spectrum and peak value to construct a four-dimensional linear regression equation, wherein the expression is as follows:
Rank=a 1 ×PSD Mean +a 2 ×F 0 +a 3 ×PSD Max +a 4 ×Peak+b
wherein Rank is the score, a 1 、a 2 、a 3 、a 4 B is a four-dimensional linear regression equation coefficient, PSD Mean To weight the power spectral density, F 0 Dominant frequency, PSD Max Peak is the Peak value for the maximum of the power spectrum.
The severity of the parkinson's disease diagnosis was assessed quantitatively using the Rank score, with higher scores the more severe parkinson symptoms.
Other steps and parameters are the same as those in one of the first to seventh embodiments.
The specific implementation method nine: the difference between the embodiment and one of the first to eighth embodiments is that the parkinson detection system based on the tremor signal of the wrist further comprises a drug dosage determination module, the drug dosage determination module is used for determining the drug dosage, and the drug dosage calculation mode is as follows:
the severity of the wrist tremor is divided into five grades of no, mild, small, medium and large amplitude;
rank values are 0,1, 2, 3, 4;
0 corresponds to none in severity of wrist tremor;
1 mild to moderate severity of carpal tremor;
2 small amplitude in the severity of the corresponding wrist tremor;
3 moderate amplitude in the severity of carpal tremor;
4 large amplitude in severity of tremor of the wrist;
the dosage is determined based on different Rank values and tremor onset times:
MT(t+1)=MT(t)+AT-OT×ω
wherein Do sin gN (t) represents the dosage of the t-th medicine taking in the absence of disease,indicates the dosage of the i' th drug, and the size of the drug is only equal to the scoreThe higher the correlation, the higher the score,the larger, a i′ The medicine type is shown, and the medicine is needed to be taken according to different disease conditions;
wherein Do sin gY (t) represents the dosage of the t-th medicine taking at the onset of disease,indicates that the dosage of the drug of the i < th > type is only related to the score, the higher the score,the larger; beta is a beta i″ Medicines which are required to be taken according to different disease conditions for the types of the medicines;
MT (t + 1) indicates the dosing time of the t +1 th time, MT (t) indicates the dosing time of the t th time, AT indicates the dosing interval, OT indicates the onset time, ω indicates the influence of onset on the dosing time, ω =1 AT onset, and ω =0 AT no onset.
The above formula is compiled as a function in python, setting different parameters α, β, ω for different conditions.
Other steps and parameters are the same as those in one to eight of the embodiments.
The detailed implementation mode is ten: the embodiment is a parkinson detection device based on a wrist tremor signal, the device comprises a processor and a memory, it should be understood that any device described in the present invention, which comprises a processor and a memory, may also comprise other units and modules for displaying, interacting, processing, controlling and the like through signals or instructions;
the memory has stored therein at least one instruction that is loaded and executed by the processor to implement a parkinson's detection system based on wrist tremor signals.
Simulation experiment
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without inventive effort based on the embodiments of the present invention, are within the scope of the present invention.
And solving the regression equation coefficients according to the unified Parkinson scale scores and the four frequency characteristics. For the obtained four-dimensional linear regression equation of the predicted parkinsonism tremor score, if the predicted score is a negative number, the value is 0 (normal state).
TABLE 1 detailed List of raw feature sets
The embodiment provides a Parkinson detection system and equipment based on a wrist tremor signal, obtains a related quick and effective machine learning training library and a four-dimensional linear regression equation for evaluation, can be applied to a bottom chip, and can quickly judge and ensure a certain accuracy.
A Parkinson detection system based on a wrist tremor signal can be used for a mobile phone terminal, a PC terminal, a reminding device of a rehabilitation hospital and the like.
Those of ordinary skill in the art will understand that: the figures are merely schematic representations of one embodiment, and the blocks or flow diagrams in the figures are not necessarily required to practice the present invention.
The present invention is capable of other embodiments and its several details are capable of modifications in various obvious respects, all without departing from the spirit and scope of the present invention.
Claims (10)
1. A Parkinson's detection system based on wrist tremor signal, which is characterized in that: the system comprises:
the system comprises a data acquisition module, a data division module, a data detection and correction module, a data preprocessing module, a characteristic data set acquisition module, a training set and test set acquisition module, a machine learning module and a detection module;
the data acquisition module is used for acquiring a wrist tremor signal when a disease occurs and a wrist normal signal when the disease does not occur;
the data dividing module is used for dividing data of the wrist tremor signal acquired by the data acquisition module in case of illness and the normal wrist signal acquired by the data acquisition module in case of no illness;
the data detection and correction module is used for carrying out data real-time detection on the divided data to realize correction on missing values and abnormal values;
the data preprocessing module is used for carrying out filtering and synthesizing acceleration data preprocessing operations on the data after the missing values and the abnormal values are corrected;
the characteristic data set acquisition module is used for carrying out characteristic extraction and characteristic dimension reduction on the preprocessed signals to obtain a characteristic data set;
the training set and test set acquisition module is used for dividing the characteristic data set into a training set and a test set;
the machine learning module is used for obtaining a training data model, performing machine learning training on a training set by using a K nearest neighbor algorithm and verifying by using a test set to finally obtain a trained training data model;
the detection module is used for inputting the wrist tremor signal to be detected into the trained training data model, classifying the wrist tremor signal to be detected, and if the wrist tremor signal to be detected is a normal signal, continuously acquiring the wrist tremor signal; and if the signal is an abnormal signal, constructing a multi-dimensional linear regression equation by using the frequency characteristics, and evaluating the Parkinson disease degree.
2. The parkinsonism detection system according to claim 1, wherein: the data acquisition module is used for acquiring a wrist tremor signal when a disease occurs and a wrist normal signal when the disease does not occur; the wrist tremor signal when the disease occurs and the wrist normal signal when the disease does not occur are both obtained by the three-axis acceleration sensor, and the wrist tremor signal when the disease occurs and the wrist normal signal when the disease does not occur which are obtained by the three-axis acceleration sensor are both three-axis acceleration signals.
3. The parkinsonism detection system based on a wrist tremor signal of claim 2, wherein: the data dividing module is used for dividing data of the wrist tremor signal acquired by the data acquisition module in case of illness and the normal wrist signal acquired by the data acquisition module in case of no illness; the specific process is as follows:
calculating the time domain characteristics of the collected wrist tremor signals during the onset of the disease, wherein the time domain characteristics comprise the characteristics of the peak value of the signals during the onset of the Parkinson's disease, the zero crossing times of the signals during the onset of the Parkinson's disease and the like;
calculating the frequency domain characteristics of the acquired wrist tremor signals during onset, wherein the frequency domain characteristics comprise main frequency of the signals during onset of Parkinson, power spectral density and energy spectral density of the signals during onset of Parkinson, weighted power spectral density obtained from the power spectral density, maximum power spectral value obtained from the power spectral density and the like;
the wrist normal signal data collected by the data collection module during non-morbidity is divided into two states of a motion state interference segment and a static state stable segment, wherein the motion state interference segment and the static state stable segment respectively account for 50%.
4. The parkinsonism detection system according to claim 3, wherein: the data detection and correction module is used for carrying out data real-time detection on the divided data, and constructing a correlation function based on a multiple filling method and the basic principle of the Chebyshev inequality so as to realize correction on missing values and abnormal values;
the specific process is as follows:
step 301, a df.isnull function in a pandas library is called in python, a missing value NAN in the divided data is searched, the missing value NAN is the ith data in the divided data, the data before and after the missing value NAN, namely the (i-1) th data and the (i + 1) th data, are selected, a section formed by the (i-1) th data to the (i + 1) th data is divided into m parts to generate m data, the generated m data are respectively estimated by Bayesian to obtain the probability corresponding to each data, and the maximum value in the m probability values is taken as the estimated value of the missing value NAN;
step 302, detecting abnormal values contained in the divided data by a 3 sigma detection method, removing the abnormal values, and correcting the abnormal values according to the step 301;
the 3 sigma detection method comprises the following steps:
p(|X-E(X)|≥ε)≤D(X)/ε 2 ≈0.11
wherein: epsilon is the length of the interval around the mean, epsilon =3 sigma, sigma is the standard deviation, D (X) = sigma is the standard deviation, X is the sample value, E (X) is the mean of the sample, p () is the probability of the occurrence of an outlier;
when E (X) -epsilon is less than or equal to X and less than or equal to E (X) + epsilon, X is a normal value;
when X > E (X) + ε or X < E (X) - ε, X is an abnormal value.
5. The system of claim 4, wherein the Parkinson's detection system is based on wrist tremor signals and comprises: the data preprocessing module is used for carrying out filtering and synthesizing acceleration data preprocessing operations on the data after the missing values and the abnormal values are corrected; the specific process is as follows:
step 401, performing FA-VMD-WTD denoising processing on the data after the abnormal value and the missing value are corrected to obtain denoised signal data;
the FA-VMD-WTD method combines the variable mode decomposition denoising and the wavelet hard threshold processing after the firefly optimization algorithm is optimized;
the FA is a firefly algorithm;
the VMD is a variational modal decomposition;
the WTD is wavelet hard threshold denoising;
step 402, synthesizing the denoised signal data into acceleration:
wherein: a. The x 、A y 、A z The de-noised acceleration signals of the x axis, the y axis and the z axis are respectively.
6. The system of claim 5, wherein the system is configured to detect Parkinson's disease based on a wrist tremor signal, and is further configured to: in the step 401, the data after the correction of the abnormal value and the missing value is subjected to the FA-VMD-WTD denoising processing to obtain denoised signal data;
the FA-VMD-WTD method combines the variation modal decomposition denoising and the wavelet hard threshold processing after the firefly optimization algorithm is optimized;
the FA is a firefly algorithm; the VMD is a variational modal decomposition; the WTD is wavelet hard threshold denoising;
the specific process is as follows:
4011, solving optimal variational modal decomposition method parameters K and alpha based on a firefly algorithm FA; the specific process is as follows:
s1, setting value ranges of K and alpha in a variational modal decomposition VMD algorithm;
k is the modal number of the decomposition target, and alpha is a secondary penalty factor;
k is more than or equal to 3 and less than or equal to 20;
alpha is more than or equal to 500 and less than or equal to 2500;
s2, initializing basic parameters of firefly algorithm FA, namely problem dimension, group size and maximum attraction degree beta 0 The light absorption coefficient, the step factor, the iteration times and the value ranges of K and alpha;
k is the modal number of the decomposition target, and alpha is a secondary penalty factor;
k is more than or equal to 3 and less than or equal to 20;
alpha is more than or equal to 500 and less than or equal to 2500;
s3, randomly initializing the K value and the alpha value of the ith firefly as initial positions in space, wherein i =1,2, \ 8230;, 25;
s4, calculating a fitness function of the ith firefly as a fluorescein value l of the ith firefly at the time t i (t) as the ith onlyMaximum fluorescence intensity of firefly I 0 ;
S5, calculating the relative brightness I of the ith firefly in the population r And attraction degree beta, determining the maximum I r Value, I-th firefly to maximum I r A value moving direction movement;
r represents the distance between two fireflies, gamma represents the light intensity absorption coefficient, I 0 Denotes the maximum fluorescence intensity, I r Denotes the relative brightness of firefly, beta denotes the attraction degree, beta 0 Represents the maximum attraction degree;
s6, based on beta and I in S5 r Updating the spatial position of the ith firefly, namely the new K value and the new alpha value, and the formula is as follows:
said X is i (t + 1) represents the updated spatial position of the ith firefly, X i (t) represents the spatial position of the ith firefly, X j (t) represents the spatial position of the jth firefly,denotes the step size factor, rand denotes [0,1 ]]Obeying uniformly distributed random factors;
based on the new K value and the alpha value, re-executing S4 to obtain the maximum fluorescence brightness;
s7, judging whether the set maximum iteration times are reached, if so, outputting the brightness and the position of the optimal firefly, namely the optimal K value and the optimal alpha value (outputting the optimal firefly luminous intensity); if not, based on the new K value and the alpha value, executing S4-S6 again until the maximum iteration times is reached, and outputting the optimal firefly position, namely the optimal K value and the alpha value;
step 4012, performing variational modal decomposition on the collected wrist tremor signal data according to the K value and the alpha value obtained in step 4011, and determining a dominant signal component, an interference signal component and a noise signal component; the specific process is as follows:
introducing a variational modal decomposition module vmdpy into pyson, and introducing a K value and an alpha value obtained by an FA algorithm and the collected wrist tremor signal into the variational modal decomposition module to obtain K decomposed signals;
calculating the correlation between each decomposed signal and the collected wrist tremor signal, and taking a component with the correlation of 0.5-1.0 as a dominant signal component, a component with the correlation of 0.2-0.5 as an interference signal component and a component with the correlation of 0-0.2 as a noise signal component;
step 4013, performing wavelet denoising processing on the dominant signal component, the interference signal component and the noise signal component determined in step 4012; the specific process is as follows:
for noise signal components, directly discarding;
for the dominant signal component, directly preserving;
for interference signal components, denoising by adopting a wavelet hard threshold;
and step 4014, directly synthesizing the components subjected to wavelet denoising processing in the step 4013 to obtain denoised signal data.
7. The parkinsonism detection system according to claim 6, wherein: the characteristic data set acquisition module is used for performing characteristic extraction and characteristic dimension reduction on the signal subjected to noise removal by the FA-VMD-WTD method to obtain a characteristic data set; the specific process is as follows:
performing feature extraction on the signal after noise is removed by the FA-VMD-WTD method;
calculating the importance of the extracted features by adopting a principal component analysis algorithm, screening the importance of the features by using a principal component analysis method, and realizing feature dimension reduction on the features with calculated importance by TSNE visualization to obtain a feature data set.
8. The parkinsonism detection system according to claim 7, wherein: the detection module is used for inputting the to-be-detected wrist tremor signal into the trained training data model, classifying the to-be-detected wrist tremor signal, and if the to-be-detected wrist tremor signal is a normal signal, continuously acquiring the wrist tremor signal; if the signal is an abnormal signal, constructing a multi-dimensional linear regression equation by using the frequency characteristics, and evaluating the Parkinson disease degree; the specific process is as follows:
selecting weighted power spectrum density, main frequency, maximum value of power spectrum and peak value to construct a four-dimensional linear regression equation, wherein the expression is as follows:
Rank=a 1 ×PSD Mean +a 2 ×F 0 +a 3 ×PSD Max +a 4 ×Peak+b
wherein Rank is the score, a 1 、a 2 、a 3 、a 4 B is four-dimensional linear regression equation coefficient, PSD Mean To weight the power spectral density, F 0 Dominant frequency, PSD Max Peak is the Peak value for the maximum of the power spectrum.
9. The parkinsonism detection system according to claim 8, wherein: the system also comprises a drug consumption determining module, wherein the drug consumption determining module is used for determining the drug consumption, and the calculation mode of the drug consumption is as follows:
the severity of the wrist tremor is divided into five grades of no, mild, small, medium and large amplitude;
rank values are 0,1, 2, 3, 4;
0 corresponds to none in severity of wrist tremor;
1 mild to moderate severity of carpal tremor;
2 small amplitude in the severity of the corresponding wrist tremor;
3 moderate amplitude in the severity of carpal tremor;
4 large amplitude in the severity of wrist tremor;
the dosage is determined based on different Rank values and tremor onset times:
MT(t+1)=MT(t)+AT-OT×ω
wherein DosingN (t) represents the dosage of the t-th medicine taking in the absence of the disease,indicates the dosage of the i' th drug, alpha i′ Indicates the kind of medication;
wherein DosingY (t) represents the dosage of the t-th medicine at the onset of disease, G beta i″ () Indicates the dosage of the drug of the i' type; beta is a i″ The kind of medicine to be used;
MT (t + 1) indicates the dosing time of the t +1 th time, MT (t) indicates the dosing time of the t th time, AT indicates the dosing interval, OT indicates the onset time, ω indicates the influence of onset on the dosing time, ω =1 AT onset, and ω =0 AT no onset.
10. A Parkinson's detection equipment based on wrist tremor signal which characterized in that: the device comprises a processor and a memory having stored therein at least one instruction that is loaded and executed by the processor to implement a wrist tremor signal based parkinson's detection system according to any of claims 1 to 9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211294802.9A CN115500822A (en) | 2022-10-21 | 2022-10-21 | Parkinson detection system and equipment based on wrist tremor signal |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211294802.9A CN115500822A (en) | 2022-10-21 | 2022-10-21 | Parkinson detection system and equipment based on wrist tremor signal |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115500822A true CN115500822A (en) | 2022-12-23 |
Family
ID=84511066
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211294802.9A Pending CN115500822A (en) | 2022-10-21 | 2022-10-21 | Parkinson detection system and equipment based on wrist tremor signal |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115500822A (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109524083A (en) * | 2018-10-30 | 2019-03-26 | 平安科技(深圳)有限公司 | Based on the rational method of medical big data monitoring adjuvant drug and Related product |
CN113111893A (en) * | 2020-01-09 | 2021-07-13 | 中国移动通信集团四川有限公司 | Data processing method and system and electronic equipment |
CN113100756A (en) * | 2021-04-15 | 2021-07-13 | 重庆邮电大学 | Stacking-based Parkinson tremor detection method |
CN114224296A (en) * | 2022-01-13 | 2022-03-25 | 福州大学 | Parkinson motion symptom quantitative evaluation method based on wearable sensing device |
-
2022
- 2022-10-21 CN CN202211294802.9A patent/CN115500822A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109524083A (en) * | 2018-10-30 | 2019-03-26 | 平安科技(深圳)有限公司 | Based on the rational method of medical big data monitoring adjuvant drug and Related product |
CN113111893A (en) * | 2020-01-09 | 2021-07-13 | 中国移动通信集团四川有限公司 | Data processing method and system and electronic equipment |
CN113100756A (en) * | 2021-04-15 | 2021-07-13 | 重庆邮电大学 | Stacking-based Parkinson tremor detection method |
CN114224296A (en) * | 2022-01-13 | 2022-03-25 | 福州大学 | Parkinson motion symptom quantitative evaluation method based on wearable sensing device |
Non-Patent Citations (2)
Title |
---|
李定文: "优化的变分模态分解算法在信号去噪中的应用", 《中国优秀硕士学位论文全文数据库》, vol. 2022, no. 3, 16 February 2022 (2022-02-16), pages 5 - 8 * |
李道军: "基于萤火虫算法优化VMD 的滚动轴承故障检测", 《机床与液压》, vol. 49, no. 15, 15 August 2021 (2021-08-15), pages 195 - 199 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20200015694A1 (en) | Automatic method to delineate or categorize an electrocardiogram | |
US10332638B2 (en) | Methods and systems for pre-symptomatic detection of exposure to an agent | |
CN111009321A (en) | Application method of machine learning classification model in juvenile autism auxiliary diagnosis | |
Chang et al. | DeepHeart: A deep learning approach for accurate heart rate estimation from PPG signals | |
Kumaravel et al. | A simplified framework for the detection of intracranial hemorrhage in CT brain images using deep learning | |
JPWO2019216378A1 (en) | Arithmetic logic unit, detection device, arithmetic method, and computer program | |
Liu et al. | MGNN: A multiscale grouped convolutional neural network for efficient atrial fibrillation detection | |
CN115240803A (en) | Model training method, complication prediction system, complication prediction device, and complication prediction medium | |
Siuly et al. | SchizoGoogLeNet: The GoogLeNet‐Based Deep Feature Extraction Design for Automatic Detection of Schizophrenia | |
CN114343585B (en) | Cognitive and behavioral disorder early warning method, device, equipment and storage medium | |
CN110693510A (en) | Attention deficit hyperactivity disorder auxiliary diagnosis device and using method thereof | |
Yamada et al. | Characteristics of drawing process differentiate Alzheimer’s disease and dementia with lewy bodies | |
CN108597615A (en) | A kind of screening reference method of Patients with Mild Cognitive Impairment dementia conversion | |
Liu et al. | Semantic segmentation of qrs complex in single channel ecg with bidirectional lstm networks | |
CN115500822A (en) | Parkinson detection system and equipment based on wrist tremor signal | |
Chen et al. | An interpretable deep learning optimized wearable daily detection system for Parkinson’s disease | |
Verma et al. | Methodologies, Applications, and Challenges of Pneumonia Detection of Chest X-Ray images for COVID-19 using IoT-enabled Deep Learning | |
CN115376638A (en) | Physiological characteristic data analysis method based on multi-source health perception data fusion | |
Ramaiah et al. | DETECTION OF PARKINSON'S DISEASE VIA CLIFFORD GRADIENT-BASED RECURRENT NEURAL NETWORK USING MULTI-DIMENSIONAL DATA | |
Rajmohan et al. | G-Sep: A deep learning algorithm for detection of long-term sepsis using bidirectional gated recurrent unit | |
CN112686091A (en) | Two-step arrhythmia classification method based on deep neural network | |
Henderson et al. | Never a dull moment: Distributional properties as a baseline for time-series classification | |
Yousafzai et al. | Improved Neural Network-Based System for Early and Accurate Diagnosis of Alzheimer Disease | |
Noble-Nnakenyi et al. | Predicting Epileptic Seizures using Ensemble Method | |
Su et al. | An Interpretable Deep Learning Optimized Wearable Daily Monitoring System for Parkinson's Disease Patients |
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 |