CN117017277A - Identity recognition method, device, equipment and medium based on photoelectric volume pulse wave - Google Patents
Identity recognition method, device, equipment and medium based on photoelectric volume pulse wave Download PDFInfo
- Publication number
- CN117017277A CN117017277A CN202311000014.9A CN202311000014A CN117017277A CN 117017277 A CN117017277 A CN 117017277A CN 202311000014 A CN202311000014 A CN 202311000014A CN 117017277 A CN117017277 A CN 117017277A
- Authority
- CN
- China
- Prior art keywords
- model
- historical
- ppg signal
- identity
- identity recognition
- 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
- 238000000034 method Methods 0.000 title claims abstract description 79
- 238000012549 training Methods 0.000 claims abstract description 78
- 239000011159 matrix material Substances 0.000 claims abstract description 74
- 230000008569 process Effects 0.000 claims abstract description 27
- 230000006870 function Effects 0.000 claims description 29
- 238000004590 computer program Methods 0.000 claims description 22
- 238000004364 calculation method Methods 0.000 claims description 12
- 230000017105 transposition Effects 0.000 claims description 5
- 238000013186 photoplethysmography Methods 0.000 description 175
- 238000013527 convolutional neural network Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000012795 verification Methods 0.000 description 4
- 238000007781 pre-processing Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013500 data storage Methods 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 230000003068 static effect Effects 0.000 description 2
- 238000013526 transfer learning Methods 0.000 description 2
- OKTJSMMVPCPJKN-UHFFFAOYSA-N Carbon Chemical compound [C] OKTJSMMVPCPJKN-UHFFFAOYSA-N 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 229910021389 graphene Inorganic materials 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/117—Identification of persons
-
- 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 pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
-
- 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
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Theoretical Computer Science (AREA)
- Biomedical Technology (AREA)
- Pathology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- Physiology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Psychiatry (AREA)
- Signal Processing (AREA)
- Fuzzy Systems (AREA)
- Cardiology (AREA)
- Image Analysis (AREA)
Abstract
The application relates to an identity recognition method based on photoelectric volume pulse waves, which comprises the following steps: acquiring target PPG signal data; obtaining a target classification recognition result corresponding to the target PPG signal data according to the target PPG signal data and the target identity recognition model; the target identity recognition model is obtained by training a historical identity recognition model based on the current PPG signal database; the target identity model comprises an adjustment weight matrix, and the adjustment weight matrix is used for determining model parameters which need to be adjusted in the process of obtaining the target identity model based on the historical identity model training. By adopting the method, higher classification recognition accuracy can be maintained when the input PPG signal database changes.
Description
Technical Field
The application relates to the technical field of biological signal identification, in particular to an identity identification method, device, equipment and medium based on photoelectric volume pulse waves.
Background
The PPG (Photo Plethysmo Graphy) signal is a medical physiological signal with individual specificity, and the identification method based on the PPG signal is an identification method widely used at present.
In the related art, the identification method based on the PPG signal generally includes: firstly, a convolutional neural network for identity recognition is obtained based on the current PPG signal database training, and then the convolutional neural network is utilized to conduct feature extraction and classification recognition on the obtained PPG signals, so that the user identity corresponding to the PPG signals is obtained.
However, the identification method based on the PPG signal has low accuracy when being applied to an identification system with strong personnel mobility.
Disclosure of Invention
In view of the foregoing, it is desirable to provide a method, apparatus, device, and medium for identifying an identity based on a photoplethysmography pulse wave, which can provide high accuracy in an identity identification system with high mobility for personnel.
In a first aspect, the present application provides a method for identifying an identity based on a photoplethysmography pulse wave, the method comprising:
acquiring target PPG signal data;
obtaining a target classification recognition result corresponding to the target PPG signal data according to the target PPG signal data and the target identity recognition model; the target identity recognition model is obtained by training a historical identity recognition model based on the current PPG signal database; the target identity model comprises an adjustment weight matrix, and the adjustment weight matrix is used for determining model parameters which need to be adjusted in the process of obtaining the target identity model based on the historical identity model training.
In one embodiment, the method further comprises:
performing iterative training on the initial identity recognition model according to the historical PPG signal database to obtain a historical identity recognition model and recognition accuracy of the historical identity recognition model;
and obtaining an adjustment weight matrix according to the recognition accuracy of the historical identity recognition model.
In one embodiment, performing iterative training on the initial identity recognition model according to the historical PPG signal database to obtain the historical identity recognition model and the recognition accuracy of the historical identity recognition model, including:
acquiring a historical PPG signal database, wherein the historical PPG signal database comprises a plurality of historical PPG signal data and historical user identities corresponding to each historical PPG signal data;
inputting the historical PPG signal data into an initial identity recognition model aiming at each historical PPG signal data to obtain an initial recognition result output by the initial identity recognition model;
and carrying out iterative training on the initial identity recognition model according to preset iterative conditions to obtain a historical identity recognition model and recognition accuracy of the historical identity recognition model.
In one embodiment, obtaining the adjustment weight matrix according to the recognition accuracy of the historical identity recognition model includes:
Carrying out gradient calculation on each model parameter in the historical identity model according to the identification accuracy to obtain a first parameter gradient matrix;
performing transposition on the first parameter gradient matrix to obtain a second parameter gradient matrix;
and performing inner product calculation on the first parameter gradient and the second parameter gradient to obtain an adjustment weight matrix.
In one embodiment, the method further comprises:
acquiring a current PPG signal database and a historical identity recognition model; the current PPG signal database comprises a plurality of current PPG signal data and current user identities corresponding to the current PPG signal data;
aiming at each piece of current PPG signal data, classifying and identifying the current PPG signal data by using a historical identity identification model to obtain an intermediate identification result;
according to the intermediate identification result, the current user identity corresponding to the current PPG signal data and the loss function, adjusting each model parameter of the historical identity identification model to obtain an intermediate identity identification model; regularization terms of the loss function are determined based on the adjustment weight matrix and model parameters of the historical identity recognition model;
and carrying out iterative training on the intermediate identity recognition model until a preset iteration condition is reached, outputting the intermediate identity recognition model corresponding to the intermediate identity recognition model when the preset iteration condition is reached, and obtaining the target identity recognition model.
In one embodiment, according to the intermediate recognition result, the current user identity corresponding to the current PPG signal data, and the loss function, each model parameter of the historical identity recognition model is adjusted to obtain the intermediate identity recognition model, including:
aiming at each weight data in the adjustment weight matrix, if the weight data is smaller than a preset threshold value, adjusting model parameters corresponding to the weight data in the historical identity recognition model according to the loss function to obtain an intermediate identity recognition model;
if the weight data is greater than or equal to a preset threshold value, model parameters corresponding to the weight data in the historical identity recognition model are kept unchanged, and an intermediate identity recognition model is obtained;
wherein the weight data corresponds to model parameters of the historical identity model.
In a second aspect, the present application further provides an identification device based on a photoplethysmography pulse wave, where the device includes:
the data acquisition module is used for acquiring target PPG signal data;
the classification recognition module is used for obtaining a target classification recognition result corresponding to the target PPG signal data according to the target PPG signal data and the target identity recognition model; the target identity recognition model is obtained by training a historical identity recognition model based on the current PPG signal database; the target identity model comprises an adjustment weight matrix, and the adjustment weight matrix is used for determining model parameters which need to be adjusted in the process of obtaining the target identity model based on the historical identity model training.
In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor implementing the steps of the method according to the first aspect when the processor executes the computer program.
In a fourth aspect, the present application also provides a computer-readable storage medium. The computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method according to the first aspect.
In a fifth aspect, the present application also provides a computer program product. The computer program product comprising a computer program which, when executed by a processor, implements the steps of the method according to the first aspect.
The identity recognition method, the device, the equipment and the medium based on the photoplethysmogram waves acquire target PPG signal data; obtaining a target classification recognition result corresponding to the target PPG signal data according to the target PPG signal data and the target identity recognition model; the target identity recognition model is obtained by training a historical identity recognition model based on the current PPG signal database; the target identity recognition model comprises an adjustment weight matrix, wherein the adjustment weight matrix is used for determining model parameters which need to be adjusted in the process of obtaining the target identity recognition model based on the history identity recognition model training; in this way, each parameter of the target identity recognition model is obtained by adjusting the model parameters and the adjustment weight matrix of the history identity recognition model according to different adjustment weights, so that the problem that the identity recognition model is forgotten catastrophically due to interference of new information to knowledge learned by the original identity recognition model when the input database of the identity recognition model based on transfer learning changes in the prior art is avoided, and the weight sharing among the model parameters of the original identity recognition model causes the interference information. According to the application, different adjustment weights are set for different model parameters, so that in the process of training to obtain the target identity recognition model, the model parameters of the history identity recognition model are adjusted in different ranges, and the obtained target identity recognition model can keep higher classification recognition accuracy when the input PPG signal database changes.
Drawings
FIG. 1 is a diagram showing an application environment of an identification method based on photoplethysmography in one embodiment;
FIG. 2 is a flow chart of an identification method based on photoplethysmography in one embodiment;
FIG. 3 is a flowchart illustrating steps for obtaining an adjustment weight matrix in one embodiment;
FIG. 4 is a flowchart illustrating steps for obtaining a historical identity model and recognition accuracy in one embodiment;
FIG. 5 is a flowchart illustrating steps for obtaining an adjustment weight matrix according to one embodiment;
FIG. 6 is a flowchart illustrating steps for obtaining a target identity model in one embodiment;
FIG. 7 is a flowchart illustrating steps for obtaining an intermediate identity model in one embodiment;
FIG. 8 is a flowchart of an identification method based on photoplethysmography in another embodiment;
FIG. 9 is a block diagram of an identification device based on photoplethysmography in one embodiment;
fig. 10 is an internal structural view of a computer device in one embodiment.
Detailed Description
The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.
The identity recognition method based on the photoplethysmogram pulse wave provided by the embodiment of the application can be applied to an application environment shown in fig. 1. Wherein the terminal 102 communicates with the server 104 via a network. The data storage system may store data that the server 104 needs to process. The data storage system may be integrated on the server 104 or may be located on a cloud or other network server. The terminal 102 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, internet of things devices, and portable wearable devices, where the internet of things devices may be smart speakers, smart televisions, smart air conditioners, smart vehicle devices, and the like. The portable wearable device may be a smart watch, smart bracelet, headset, or the like. The server 104 may be implemented as a stand-alone server or as a server cluster of multiple servers.
Terminal 102 obtains target PPG signal data; obtaining a target classification recognition result corresponding to the target PPG signal data according to the target PPG signal data and the target identity recognition model; the target identity recognition model is obtained by training a historical identity recognition model based on the current PPG signal database; the target identity model comprises an adjustment weight matrix, and the adjustment weight matrix is used for determining model parameters which need to be adjusted in the process of obtaining the target identity model based on the historical identity model training.
In one embodiment, as shown in fig. 2, an identity recognition method based on a photoplethysmography pulse wave is provided, and the method is applied to the terminal 102 in fig. 1 for illustration, and includes the following steps:
step 202, target PPG signal data is acquired.
The target PPG signal data is PPG signal data obtained by preprocessing an initial PPG signal which is directly acquired in a preset mode.
And 204, obtaining a target classification recognition result corresponding to the target PPG signal data according to the target PPG signal data and the target identity recognition model.
The target identity recognition model is obtained by training a historical identity recognition model based on a current PPG signal database; the target identity model includes an adjustment weight matrix.
The historical identity recognition model is obtained based on historical PPG signals in a historical PPG signal database and historical user identities corresponding to the historical PPG signals.
In the process of training to obtain a target identity recognition model, inputting a current PPG signal in a current PPG signal database and a current user identity corresponding to the current PPG signal into a historical identity recognition model for training, and obtaining the model after reaching a preset condition, namely the target identity recognition model.
By way of example, the historical identity recognition model may be a deep learning model such as a convolutional neural network, a recurrent neural network, or a deep belief network. The historical identity recognition model comprises a feature extraction layer and a decision layer, the current PPG signal database is input into the historical identity recognition model for training, and the model obtained after the preset condition is reached is the target identity recognition model. When the scales of the historical PPG signal database and the current PPG signal database are different, the decision layer parameters of the historical identity recognition model need to be adaptively adjusted, and then the historical identity recognition model is trained based on the current PPG signal database.
The adjustment weight matrix is used for determining model parameters which need to be adjusted in the process of training based on the historical identity recognition model to obtain the target identity recognition model.
And (3) evaluating the importance of each model parameter in the historical identity recognition model by adjusting the weight matrix, wherein each weight data in the weight matrix represents the importance of the corresponding model parameter. In the process of obtaining the target identity recognition model based on the historical identity recognition model training, the weight matrix is adjusted to restrict each model parameter to be adjusted in different ranges according to different importance of each model parameter, so that the target identity recognition model obtained through the control training can keep the knowledge learned by the historical identity recognition model from the historical PPG signal database.
Illustratively, the adjustment weight matrix may be a second order bias matrix of each model parameter pair loss function of the historical identity recognition model, and may also be defined by a first order method.
In the identity recognition method based on the photoplethysmogram, target PPG signal data are obtained; obtaining a target classification recognition result corresponding to the target PPG signal data according to the target PPG signal data and the target identity recognition model; the target identity recognition model is obtained by training a historical identity recognition model based on the current PPG signal database; the target identity recognition model comprises an adjustment weight matrix, wherein the adjustment weight matrix is used for determining model parameters which need to be adjusted in the process of obtaining the target identity recognition model based on the history identity recognition model training; in this way, each parameter of the target identity recognition model is obtained by adjusting the model parameters and the adjustment weight matrix of the history identity recognition model according to different adjustment weights, so that the problem that the identity recognition model is forgotten catastrophically due to interference of new information to knowledge learned by the original identity recognition model when the input database of the identity recognition model based on transfer learning changes in the prior art is avoided, and the weight sharing among the model parameters of the original identity recognition model causes the interference information. According to the application, different adjustment weights are set for different model parameters, so that in the process of training to obtain the target identity recognition model, the model parameters of the history identity recognition model are adjusted in different ranges, and the obtained target identity recognition model can keep higher classification recognition accuracy when the input PPG signal database changes.
In one embodiment, based on the embodiment shown in fig. 2, as shown in fig. 3, the method further comprises:
step 302, performing iterative training on the initial identity recognition model according to the historical PPG signal database to obtain a historical identity recognition model and recognition accuracy of the historical identity recognition model.
The initial identification model may be a deep learning model such as a convolutional neural network, a recurrent neural network, or a deep belief network, for example.
And carrying out iterative training on the initial identity recognition model, outputting the model reaching a preset iterative condition as a historical identity recognition model, and taking the accuracy of the recognition result of the historical identity recognition model based on the historical PPG signal database as the recognition accuracy of the historical identity recognition model.
And step 304, obtaining an adjustment weight matrix according to the recognition accuracy of the historical identity recognition model.
In this embodiment, the recognition accuracy of the historical identity recognition model is utilized to obtain the adjustment weight matrix, and the recognition result based on the historical PPG signal database is used as the constraint condition in the process of training the target identity recognition model, so that the target identity recognition model can maintain the memory of the old task of the historical PPG signal database when learning the new task of the current PPG signal database, thereby improving the stability of classifying and recognizing the target identity recognition model when the input database is changed.
In one embodiment, based on the embodiment shown in fig. 3, as shown in fig. 4, the process of performing iterative training on the initial identity recognition model according to the historical PPG signal database to obtain the historical identity recognition model and the recognition accuracy of the historical identity recognition model according to the embodiment includes:
step 402, a historical PPG signal database is obtained.
The historical PPG signal database comprises a plurality of historical PPG signal data and historical user identities corresponding to the historical PPG signal data. The historical PPG signal data is PPG signal data with similar edge characteristics and geometric characteristics, which is obtained by preprocessing original PPG signals directly collected from each historical user in a unified mode.
Step 404, for each historical PPG signal data, the historical PPG signal data is input into the initial identity recognition model, so as to obtain an initial recognition result output by the initial identity recognition model.
The initial recognition result may be that the recognition is passed, i.e. the historical PPG signal data corresponds to a certain historical user identity in the historical PPG signal database; the initial recognition result may also be that the recognition is not passed, i.e. the historical PPG signal data does not correspond to any one of the historical user identities in the historical PPG signal database.
And step 406, performing iterative training on the initial identity recognition model according to preset iterative conditions to obtain a historical identity recognition model and recognition accuracy of the historical identity recognition model.
After each training, the model parameters of the initial identity recognition model are adjusted according to the obtained recognition result until the preset iteration condition is met, and the historical identity recognition model and the recognition accuracy of the historical identity recognition model when the preset iteration condition is met are output.
In this embodiment, iterative training is performed on the initial identity recognition model according to the historical PPG signal database to obtain a historical identity recognition model and recognition accuracy of the historical identity recognition model, where the historical identity recognition model is used for training to obtain a target identity recognition model, and the recognition accuracy of the historical identity recognition model is used as a calculation basis for adjusting the weight matrix, so that model parameters of the target identity recognition model are constrained by model parameters of the historical identity recognition model, and stability of classification recognition of the target identity recognition model based on the historical PPG signal database can be maintained.
In one embodiment, based on the embodiment shown in fig. 3, as shown in fig. 5, the process of obtaining the adjustment weight matrix according to the history recognition accuracy of the history recognition model according to the present embodiment includes:
Step 502, performing gradient calculation on each model parameter in the historical identity model according to the historical identification accuracy, and obtaining a first parameter gradient.
The first parameter gradient is a matrix of prior probability gradients obtained by gradient calculation of each model parameter in the historical identity model by the historical identification accuracy.
And 504, performing transposition on the first parameter gradient matrix to obtain a second parameter gradient matrix.
Step 506, performing inner product calculation on the first parameter gradient and the second parameter gradient to obtain an adjustment weight matrix.
The adjustment weight matrix F may be expressed as a first parameter gradientAnd a second parameter gradient->Can be expressed as:
wherein x is n Representing historical PPG signal data in a historical PPG signal database, y n Represents the correctly classified historical PPG signal data in the historical recognition results,is a model parameter in the historical identity model.
In this embodiment, the adjustment weight matrix is obtained according to the history recognition accuracy of the history recognition model, the importance of each parameter in the history recognition model is determined by using the history recognition accuracy, and whether the parameter needs to be adjusted in the process of training the target recognition model is judged based on the importance of the parameter, so that the stability of classifying and recognizing the target recognition model when the input database is changed is improved.
In one embodiment, based on the embodiment shown in fig. 3, as shown in fig. 6, the method provided in this embodiment further includes:
step 602, obtaining a current PPG signal database and a historical identity model.
The current PPG signal database comprises a plurality of current PPG signal data and target user identities corresponding to the target PPG signal data. The current PPG signal data is PPG signal data with similar edge characteristics and geometric characteristics, which are obtained by preprocessing original PPG signals directly collected from each current user in a unified mode.
Step 604, for each target PPG signal data, classifying and identifying the target PPG signal data by using the historical identity identification model, so as to obtain an intermediate identification result.
The intermediate recognition result may be that the recognition is passed, i.e. the target PPG signal data corresponds to a certain target user identity in the current PPG signal database; the intermediate recognition result may also be that the recognition is not passed, i.e. the target PPG signal data does not correspond to any one of the target user identities in the current PPG signal database.
The historical identity recognition model can be a convolutional neural network, and comprises a feature extraction layer and a decision layer, the target PPG signal data is input into the historical identity recognition model, feature extraction is carried out through the feature extraction layer, and then the extracted features are input into the decision layer for classification recognition, so that an intermediate recognition result that recognition passes or fails is obtained.
Step 606, according to the intermediate recognition result, the target user identity corresponding to the target PPG signal data, and the loss function, each model parameter of the history identity recognition model is adjusted, so as to obtain the intermediate identity recognition model.
The loss function comprises a basic loss term and a regularization term, and the basic loss term of the loss function can be obtained according to the intermediate recognition result and the user identity corresponding to the PPG signal data; regularization terms of the loss function may be determined based on adjusting the weight matrix and model parameters of the historical identity model.
Exemplary, loss function L B Can be expressed as:
wherein L (θ) i ) The basic loss term representing the loss function can be in the form of cross entropy loss, perceptual loss, mean square error loss and other functions, lambda is a weight adjustment factor, L (theta) and lambda can be selected according to application scenes, and the application is not limited to the above; θ i An ith model parameter representing an identity recognition model that performs the current classification recognition procedure.
And adjusting each model parameter of the historical identity recognition model according to the value of the loss function to obtain an intermediate identity recognition model.
And 608, performing iterative training on the intermediate identity recognition model until a preset iteration condition is reached, and outputting the intermediate identity recognition model corresponding to the preset iteration condition to obtain the target identity recognition model.
And determining regularization items of the loss function according to model parameters of the intermediate identity recognition model of the current iteration process and an adjustment weight matrix during each iteration training, obtaining basic loss items of the loss function according to the intermediate recognition results and user identities corresponding to the PPG signal data, further obtaining values of the loss function of the current iteration process, and adjusting each model parameter of the intermediate identity recognition model of the current iteration process according to the values of the loss function to obtain an intermediate identity recognition model of the next iteration process until preset iteration conditions are reached.
In this embodiment, according to the intermediate recognition result, the target user identity corresponding to the target PPG signal data, and the loss function, the model parameters of the history identity recognition model and the intermediate identity recognition model are adjusted in the iterative training process to obtain a target identity recognition model conforming to a preset iteration condition, so that the weight adjustment matrix is used as a part of a regularization term of the loss function, and affects the adjustment of the model parameters of the intermediate identity recognition model in each iterative training process, and the model parameters of the target identity recognition model can be controllably adjusted when the PPG signal database input into the target identity recognition model changes through the weight adjustment matrix, thereby maintaining stable classification recognition accuracy; according to the embodiment, training is performed based on the historical identity recognition model and the current PPG signal database, when the current PPG signal database and the historical PPG signal database are in the overlapped part, the training time of the overlapped part can be saved, so that the training time of the target identity recognition model is shortened, and the training efficiency of the target identity recognition model is improved.
In one embodiment, based on the embodiment shown in fig. 6, as shown in fig. 7, the process for adjusting each model parameter of the history identity recognition model to obtain the intermediate identity recognition model according to the intermediate recognition result, the target user identity corresponding to the target PPG signal data, and the loss function provided in this embodiment includes:
step 702, for each weight data in the adjusted weight matrix, determining whether the weight data is smaller than a preset threshold.
Wherein the weight data corresponds to model parameters of the historical identity model.
And step 704, if the weight data is smaller than the preset threshold value, adjusting the model parameters of the historical identity recognition model according to the loss function to obtain an intermediate identity recognition model.
When the weight data is smaller than a preset threshold value, the importance of the model parameter of the historical identity recognition model corresponding to the weight data is lower, and after the parameter is adjusted, the difference between the intermediate recognition result obtained by recognizing the historical PPG signal in the historical PPG signal database and the historical recognition result is smaller.
And step 706, if the weight data is greater than or equal to the preset threshold value, maintaining the model parameters of the initial identity recognition model unchanged, and obtaining the intermediate identity recognition model.
When the weight data is greater than or equal to a preset threshold value, the importance of the model parameter of the historical identity recognition model corresponding to the weight data is higher, and if the parameter is adjusted, the difference between the intermediate recognition result and the historical recognition result of the intermediate identity recognition model is larger, so that the target recognition accuracy of the target identity recognition model is affected.
In this embodiment, each model parameter to be adjusted in the historical identity recognition model is determined according to the weight data, so that parameters with less influence on the historical PPG signal database are selectively updated in the process of performing iterative training according to the historical identity model to obtain the target identity recognition model, and the recognition accuracy of the target identity recognition model applied to classification recognition tasks based on the PPG signal database with strong mobility is improved.
In one embodiment, as shown in fig. 8, there is provided an identification method based on photoplethysmography, the method including:
step 802, a historical PPG signal database is obtained.
Step 804, for each historical PPG signal data, the historical PPG signal data is input into the initial identity recognition model, so as to obtain an initial recognition result output by the initial identity recognition model.
And step 806, performing iterative training on the initial identity recognition model according to preset iterative conditions to obtain a historical identity recognition model and recognition accuracy of the historical identity recognition model.
Step 808, obtaining an adjustment weight matrix according to the recognition accuracy of the historical identity recognition model.
Optionally, performing gradient calculation on each model parameter in the historical identity model according to the historical identification accuracy to obtain a first parameter gradient; performing transposition on the first parameter gradient matrix to obtain a second parameter gradient matrix; and performing inner product calculation on the first parameter gradient and the second parameter gradient to obtain an adjustment weight matrix.
Step 810, obtaining a current PPG signal database. The current PPG signal database comprises a plurality of current PPG signal data and current user identities corresponding to the current PPG signal data.
Step 812, for each current PPG signal data, classifying and identifying the current PPG signal data by using the historical identity identification model to obtain an intermediate identification result.
Step 814, for each weight data in the adjusted weight matrix, determining whether the weight data is less than a preset threshold.
And step 816, if the weight data is smaller than the preset threshold value, adjusting the model parameters of the historical identity recognition model according to the loss function to obtain an intermediate identity recognition model.
And step 818, if the weight data is greater than or equal to the preset threshold value, keeping model parameters of the initial identity recognition model unchanged, and obtaining the intermediate identity recognition model.
And step 820, performing iterative training on the intermediate identity recognition model until a preset iteration condition is reached, and outputting the intermediate identity recognition model corresponding to the preset iteration condition to obtain the target identity recognition model.
Step 822, obtain target PPG signal data.
Step 824, obtaining a target classification recognition result corresponding to the target PPG signal data according to the target PPG signal data and the target identity recognition model.
Example 1, 12 historical PPG signal data are included in the historical PPG signal database, and the current PPG signal database is that 2 historical PPG signal data are deleted from the historical PPG signal database, i.e., 10 current PPG signal data are included.
Setting iteration conditions as iteration for 20 times, and carrying out iterative training on the initial identity recognition model based on a historical PPG signal database to obtain a historical identity recognition model and a historical identity recognition model with 99.86% recognition accuracy.
And performing iterative training on the historical identity recognition model based on the current PPG signal database and the adjustment weight matrix to obtain a first target identity recognition model with the accuracy rate of 99.99%.
Example 2, the historical PPG signal database includes 12 historical PPG signal data, and the current PPG signal database is based on adding 3 historical PPG signal data to the historical PPG signal database, i.e., includes 15 current PPG signal data.
Setting iteration conditions as iteration for 20 times, and carrying out iterative training on the initial identity recognition model based on a historical PPG signal database to obtain a historical identity recognition model and a historical identity recognition model with 99.86% recognition accuracy.
And performing iterative training on the historical identity recognition model based on the current PPG signal database and the adjustment weight matrix to obtain a second target identity recognition model with the accuracy rate of 97.73%.
Training result data of the history identification model in example 1, the first target identification model, and the second target identification model in example 2 are shown in table 1.
TABLE 1 training results data for historical identity model, first target identity model, and second target identity model
| Model | Accuracy (average) | Training rounds | Duration of training |
| Historical identity recognition model | 97.4％ | 100 | 4750 seconds |
| First target identity recognition model | 98.1％ | 50 | 540 seconds |
| Second target identity recognition model | 95.1％ | 50 | 700 seconds |
As can be seen from the comparison of the table 1, the training speed of the target identity recognition model obtained by training the historical identity recognition model is obviously improved, meanwhile, the accuracy is not greatly reduced, and the higher recognition performance is maintained.
The first target identity recognition model and the second target identity recognition model can converge more rapidly by means of model parameters obtained from the historical identity recognition models, the training curve is smoother, and accuracy fluctuation is small. At 20 iterations, the first and second target identity models have already tended to steady state, while the historical identity models have not reached a converged state. From the aspect of training accuracy, the accuracy of the first target identity recognition model can reach 99.99 percent, which is superior to the highest accuracy of the historical identity recognition model by 99.86 percent. The highest accuracy of the second target identity model is 97.73% and slightly lower than that of the historical identity model, but the accuracy is not greatly reduced due to the fact that the historical identity model with smaller scale is migrated to the second target identity model with larger scale.
In addition, a verification data set is set, and verification accuracy of the historical identity recognition model, the first target identity recognition model and the second target identity recognition model in the example 1 is measured and calculated.
When the historical identity recognition model starts training, the verification accuracy fluctuates to a certain extent; the verification accuracy of the first target identity recognition model and the second target identity recognition model is relatively stable, meanwhile, higher accuracy is maintained, and the recognition performance is improved.
In this embodiment, by performing experiments on the historical identity recognition model and the target identity recognition model, the actual experimental result indicates that the method of this embodiment can improve the training speed of the model and can maintain higher recognition accuracy when the PPG signal database changes.
It should be understood that, although the steps in the flowcharts related to the embodiments described above are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.
Based on the same inventive concept, the embodiment of the application also provides an identity recognition device based on the photoelectric volume pulse wave for realizing the identity recognition method based on the photoelectric volume pulse wave. The implementation of the solution provided by the device is similar to the implementation described in the above method, so the specific limitation in the embodiment of the identification device based on the photoplethysmography wave provided below may be referred to the limitation of the identification method based on the photoplethysmography wave hereinabove, and will not be repeated here.
In one embodiment, as shown in fig. 9, there is provided an identity recognition device based on photoplethysmography, including: a data acquisition module 902 and a classification recognition module 904, wherein:
a data acquisition module 902, configured to acquire target PPG signal data.
The classification recognition module 904 is configured to obtain a target classification recognition result corresponding to the target PPG signal data according to the target PPG signal data and the target identity recognition model; the target identity recognition model is obtained by training a historical identity recognition model based on the current PPG signal database; the target identity model comprises an adjustment weight matrix, and the adjustment weight matrix is used for determining model parameters which need to be adjusted in the process of obtaining the target identity model based on the historical identity model training.
In one embodiment, the device further comprises a model training module, which is used for performing iterative training on the initial identity recognition model according to the historical PPG signal database to obtain a historical identity recognition model and recognition accuracy of the historical identity recognition model; and obtaining an adjustment weight matrix according to the recognition accuracy of the historical identity recognition model.
In one embodiment, the model training module is further configured to obtain a historical PPG signal database, where the historical PPG signal database includes a plurality of historical PPG signal data and historical user identities corresponding to each of the historical PPG signal data; inputting the historical PPG signal data into an initial identity recognition model aiming at each historical PPG signal data to obtain an initial recognition result output by the initial identity recognition model; and carrying out iterative training on the initial identity recognition model according to preset iterative conditions to obtain a historical identity recognition model and recognition accuracy of the historical identity recognition model.
In one embodiment, the model training module is further configured to perform gradient calculation on each model parameter in the historical identity model according to the identification accuracy, so as to obtain a first parameter gradient matrix; performing transposition on the first parameter gradient matrix to obtain a second parameter gradient matrix; and performing inner product calculation on the first parameter gradient and the second parameter gradient to obtain an adjustment weight matrix.
In one embodiment, the model training module is further configured to obtain a current PPG signal database and a historical identity model; the current PPG signal database comprises a plurality of target PPG signal data and target user identities corresponding to the target PPG signal data; aiming at each target PPG signal data, classifying and identifying the target PPG signal data by using a historical identity identification model to obtain an intermediate identification result; according to the intermediate identification result, the target user identity corresponding to the target PPG signal data and the loss function, adjusting each model parameter of the historical identity identification model to obtain an intermediate identity identification model; regularization terms of the loss function are determined based on the adjustment weight matrix and model parameters of the historical identity recognition model; and carrying out iterative training on the intermediate identity recognition model until a preset iteration condition is reached, outputting the intermediate identity recognition model corresponding to the intermediate identity recognition model when the preset iteration condition is reached, and obtaining the target identity recognition model.
In one embodiment, the model training module is further configured to adjust, for each weight data in the adjustment weight matrix, model parameters corresponding to the weight data in the historical identity recognition model according to the loss function if the weight data is smaller than a preset threshold value, so as to obtain an intermediate identity recognition model; if the weight data is greater than or equal to a preset threshold value, model parameters corresponding to the weight data in the historical identity recognition model are kept unchanged, and an intermediate identity recognition model is obtained; wherein the weight data corresponds to model parameters of the historical identity model.
The modules in the identity recognition device based on the photoplethysmography wave can be fully or partially realized by software, hardware and a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a terminal, and an internal structure diagram thereof may be as shown in fig. 10. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless mode can be realized through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. The computer program is executed by a processor to realize an identification method based on photoelectric volume pulse waves. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, can also be keys, a track ball or a touch pad arranged on the shell of the computer equipment, and can also be an external keyboard, a touch pad or a mouse and the like.
It will be appreciated by those skilled in the art that the structure shown in FIG. 10 is merely a block diagram of some of the structures associated with the present inventive arrangements and is not limiting of the computer device to which the present inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.
In an embodiment, there is also provided a computer device comprising a memory and a processor, the memory having stored therein a computer program, the processor implementing the steps of the method embodiments described above when the computer program is executed.
In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when executed by a processor, carries out the steps of the method embodiments described above.
In an embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the steps of the method embodiments described above.
The user information (including but not limited to user equipment information, user personal information, etc.) and the data (including but not limited to data for analysis, stored data, presented data, etc.) related to the present application are information and data authorized by the user or sufficiently authorized by each party.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magnetic random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (Phase Change Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in the form of a variety of forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), and the like. The databases referred to in the embodiments provided herein may include at least one of a relational database and a non-relational database. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processor referred to in the embodiments provided in the present application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, or the like, but is not limited thereto.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The foregoing examples illustrate only a few embodiments of the application and are described in detail herein without thereby limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of the application should be assessed as that of the appended claims.
Claims (10)
1. An identity recognition method based on photoelectric volume pulse waves is characterized by comprising the following steps:
acquiring target PPG signal data;
obtaining a target classification recognition result corresponding to the target PPG signal data according to the target PPG signal data and the target identity recognition model; the target identity recognition model is obtained by training a historical identity recognition model based on a current PPG signal database; the target identity model comprises an adjustment weight matrix, and the adjustment weight matrix is used for determining model parameters which need to be adjusted in the process of obtaining the target identity model based on the historical identity model training.
2. The method according to claim 1, wherein the method further comprises:
performing iterative training on an initial identity recognition model according to a historical PPG signal database to obtain a historical identity recognition model and recognition accuracy of the historical identity recognition model;
and obtaining the adjustment weight matrix according to the recognition accuracy of the historical identity recognition model.
3. The method according to claim 2, wherein the iterative training of the initial identity model according to the historical PPG signal database to obtain the historical identity model and the recognition accuracy of the historical identity model comprises:
the historical PPG signal database is obtained, and comprises a plurality of historical PPG signal data and historical user identities corresponding to the historical PPG signal data;
inputting the historical PPG signal data into the initial identity recognition model aiming at each historical PPG signal data to obtain an initial recognition result output by the initial identity recognition model;
and carrying out iterative training on the initial identity recognition model according to a preset iterative condition to obtain the historical identity recognition model and the recognition accuracy of the historical identity recognition model.
4. The method of claim 2, wherein the deriving the adjustment weight matrix based on the recognition accuracy of the historical identity model comprises:
performing gradient calculation on each model parameter in the historical identity model according to the identification accuracy to obtain a first parameter gradient matrix;
performing transposition on the first parameter gradient matrix to obtain a second parameter gradient matrix;
and carrying out inner product calculation on the first parameter gradient and the second parameter gradient to obtain the adjustment weight matrix.
5. The method according to claim 2, wherein the method further comprises:
acquiring the current PPG signal database and the historical identity recognition model; the current PPG signal database comprises a plurality of current PPG signal data and current user identities corresponding to the current PPG signal data;
for each piece of current PPG signal data, classifying and identifying the current PPG signal data by utilizing the historical identity identification model to obtain an intermediate identification result;
according to the intermediate identification result, the current user identity corresponding to the current PPG signal data and the loss function, adjusting each model parameter of the historical identity identification model to obtain an intermediate identity identification model; the regularization term of the loss function is determined based on the adjustment weight matrix and model parameters of the historical identity recognition model;
And carrying out iterative training on the intermediate identity recognition model until a preset iteration condition is reached, outputting the intermediate identity recognition model corresponding to the preset iteration condition, and obtaining the target identity recognition model.
6. The method according to claim 5, wherein the adjusting each model parameter of the historical identity model according to the intermediate identification result, the current user identity corresponding to the current PPG signal data, and the loss function to obtain the intermediate identity model includes:
for each weight data in the adjustment weight matrix, if the weight data is smaller than a preset threshold value, adjusting model parameters corresponding to the weight data in the historical identity recognition model according to the loss function to obtain the intermediate identity recognition model;
if the weight data is greater than or equal to the preset threshold value, model parameters corresponding to the weight data in the historical identity recognition model are kept unchanged, and the intermediate identity recognition model is obtained;
wherein the weight data corresponds to model parameters of the historical identity recognition model.
7. An identification device based on photoelectric volume pulse waves, characterized in that the device comprises:
The data acquisition module is used for acquiring target PPG signal data;
the classification recognition module is used for obtaining a target classification recognition result corresponding to the target PPG signal data according to the target PPG signal data and the target identity recognition model; the target identity recognition model is obtained by training a historical identity recognition model based on a current PPG signal database; the target identity model comprises an adjustment weight matrix, and the adjustment weight matrix is used for determining model parameters which need to be adjusted in the process of obtaining the target identity model based on the historical identity model training.
8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 6 when the computer program is executed.
9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.
10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311000014.9A CN117017277A (en) | 2023-08-09 | 2023-08-09 | Identity recognition method, device, equipment and medium based on photoelectric volume pulse wave |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311000014.9A CN117017277A (en) | 2023-08-09 | 2023-08-09 | Identity recognition method, device, equipment and medium based on photoelectric volume pulse wave |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117017277A true CN117017277A (en) | 2023-11-10 |
Family
ID=88601729
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311000014.9A Pending CN117017277A (en) | 2023-08-09 | 2023-08-09 | Identity recognition method, device, equipment and medium based on photoelectric volume pulse wave |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117017277A (en) |
-
2023
- 2023-08-09 CN CN202311000014.9A patent/CN117017277A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022042123A1 (en) | Image recognition model generation method and apparatus, computer device and storage medium | |
| CN111079780B (en) | Training method for space diagram convolution network, electronic equipment and storage medium | |
| CN110807529A (en) | Training method, device, equipment and storage medium of machine learning model | |
| JPWO2019146189A1 (en) | Neural network rank optimizer and optimization method | |
| CN112348079B (en) | Data dimension reduction processing method and device, computer equipment and storage medium | |
| CN109993026B (en) | Training method and device for relative recognition network model | |
| CN116188878A (en) | Image classification method, device and storage medium based on neural network structure fine adjustment | |
| CN113065593A (en) | Model training method and device, computer equipment and storage medium | |
| CN115797781A (en) | Crop identification method and device, computer equipment and storage medium | |
| CN116030312B (en) | Model evaluation method, device, computer equipment and storage medium | |
| CN115392594B (en) | Electrical load model training method based on neural network and feature screening | |
| CN116894721A (en) | Index prediction method and device and computer equipment | |
| CN116258923A (en) | Image recognition model training method, device, computer equipment and storage medium | |
| CN116384471A (en) | Model pruning method, device, computer equipment, storage medium and program product | |
| CN117017277A (en) | Identity recognition method, device, equipment and medium based on photoelectric volume pulse wave | |
| CN115601550B (en) | Model determination method, model determination device, computer equipment and computer readable storage medium | |
| CN116776966A (en) | Training method, training device, computer equipment and storage medium of image classification network | |
| CN117633425A (en) | Power equipment image processing method, device, equipment and medium | |
| CN117638882A (en) | Loss prediction method, device, equipment, medium and product of electric power system | |
| CN116976464A (en) | Unbiased federal learning training method, unbiased federal learning training apparatus, computer device, and storage medium | |
| CN116170335A (en) | Quality information calculating method, device, equipment, medium and product of edge equipment | |
| Liu et al. | Network Pruning via Weight Enhancement of Adaptive Exemplar Filters | |
| CN117312892A (en) | User clustering method, device, computer equipment and storage medium | |
| CN116861071A (en) | Information pushing method, information pushing device, computer equipment, storage medium and program product | |
| CN115564560A (en) | Risk assessment method, apparatus and storage medium |
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 |