CN113111726A - Vibration motor equipment fingerprint extraction and identification method based on homologous signals - Google Patents
Vibration motor equipment fingerprint extraction and identification method based on homologous signals Download PDFInfo
- Publication number
- CN113111726A CN113111726A CN202110292483.7A CN202110292483A CN113111726A CN 113111726 A CN113111726 A CN 113111726A CN 202110292483 A CN202110292483 A CN 202110292483A CN 113111726 A CN113111726 A CN 113111726A
- Authority
- CN
- China
- Prior art keywords
- layer
- cloud server
- motor
- convolution kernels
- signals
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/082—Learning methods modifying the architecture, e.g. adding, deleting or silencing nodes or connections
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Signal Processing (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a vibration motor equipment fingerprint extraction and identification method based on homologous signals, which comprises the steps of collecting acceleration and angular velocity signals in a motor vibration process through an inertial sensing unit of a terminal, segmenting, filtering and aligning the signals by the terminal, sending the signals to a cloud server, and sending an authentication or registration request to the cloud server by the terminal; the cloud server simultaneously inputs signals into a trained two-channel fusion network consisting of a residual block, a Dropout layer, a full connection layer and a loss function to obtain a motor fingerprint; classifying the motor fingerprints by adopting a classifier; if the classification is successful, the cloud server outputs a classification result; if the category does not exist, at the moment, if the terminal sends a registration request, the cloud server stores the motor fingerprint and updates the database; and if the terminal sends the authentication request, the cloud server directly refuses the authentication. The invention can overcome the influence of various noises on the identification result and improve the authentication accuracy and stability.
Description
Technical Field
The invention relates to the field of computer application, in particular to a vibration motor equipment fingerprint extraction and identification method based on homologous signals.
Background
Smart devices are already ubiquitous in our daily lives. Therefore, public attention has been paid to security issues in the use of smart devices. Existing solutions can solve the authentication problem of verifying the identity of an individual (e.g., fingerprint, PIN, face recognition). For high security scenarios (e.g., electronic payment, account login), multiple factor authentication is used in addition to the single factor authentication described above. The user needs to enter a received text message code or answer a phone call to verify that the operation is on a trusted device. However, the authentication method is very complicated because of the high cost of manual operation.
In recent years, the device fingerprint attracts people's extensive attention, and how to provide a device fingerprint with high security and simplicity in application becomes a great challenge to be solved.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a vibration motor equipment fingerprint extraction and identification method based on homologous signals, which provides a set of complete identity authentication process for equipment through the vibration of a built-in motor of the equipment, the acquisition of an inertia measurement unit and cloud analysis and identification, and has the advantages of low cost, high usability and high accuracy.
The purpose of the invention is realized by the following technical scheme:
a vibration motor device fingerprint extraction and identification method based on homologous signals comprises the following steps:
acquiring acceleration and angular velocity signals of a terminal in a motor vibration process through an inertial sensing unit of the terminal, segmenting, filtering and aligning the acceleration and angular velocity signals by the terminal, sending the signals to a cloud server, and sending an authentication request or a registration request to the cloud server by the terminal;
the cloud server simultaneously inputs the acceleration and angular velocity signals into a trained two-channel fusion network composed of a residual block, a Dropout layer, a full connection layer and a loss function to obtain a motor fingerprint; classifying the motor fingerprints by adopting a classifier; if the classification is successful, the cloud server outputs a classification result; if the type does not exist, at the moment, if the terminal sends a registration request, the cloud server stores the motor fingerprint and updates a motor fingerprint storage database; and if the terminal sends an authentication request, the cloud server directly refuses authentication.
Further, the two-channel fusion network comprises a convolution layer, a pooling layer, a flattening layer and a full-connection layer, the network has two inputs, the flattening layer expands two outputs into a one-dimensional vector, and the loss function is a generalized loss function.
Further, the two-channel fusion network comprises a convolutional layer 1, a pooling layer, a convolutional layer 2, a convolutional layer 3, a convolutional layer 4, a convolutional layer 5, a flattening layer and three full-connection layers;
the convolutional layer 1 is composed of 64 convolutional kernels with the size of 5 multiplied by 1, and the convolution step length is 2;
the pooling layer is composed of a largest pooling layer with the size of 3 multiplied by 1, and the convolution step length is 2;
the convolutional layer 2 is composed of 3 residual blocks with 64 1 × 1 convolution kernels, 64 1 × 3 convolution kernels and 256 1 × 1 convolution kernels arranged in sequence, and a Dropout layer with p equal to 0.2 is added at the end of each residual block;
the convolution layer 3 is composed of 4 residual blocks with 128 1 × 1 convolution kernels, 128 1 × 3 convolution kernels and 512 1 × 1 convolution kernels arranged in sequence, and a Dropout layer with p equal to 0.2 is added at the end of each residual block;
the convolutional layer 4 is composed of 6 residual blocks with 256 1 × 1 convolution kernels, 256 1 × 3 convolution kernels and 1024 1 × 1 convolution kernels, wherein the residual blocks are sequentially arranged, and a Dropout layer with p equal to 0.2 is finally added to each residual block;
the convolutional layer 5 is composed of 3 residual blocks with 512 1 × 1 convolution kernels, 512 1 × 3 convolution kernels and 2048 1 × 1 convolution kernels arranged in sequence, and a Dropout layer with p equal to 0.2 is added at the end of each residual block;
the flattening layer spreads all the outputs into a one-dimensional vector;
the three fully-connected layers are composed of fully-connected layers with different neuron numbers.
The invention has the following beneficial effects:
(1) according to the identification method, the motor vibration characteristics are analyzed through the two-channel fusion network and the generalized loss function, and the description part of the two-channel signal on the motor is extracted, so that the influence of noise and environmental noise of the terminal inertial sensing unit on the identification result is overcome, and the authentication accuracy and stability are improved.
(2) Most of the existing terminals comprise motors and inertia measurement units, so the identification method of the invention has low application cost and wide application range.
Drawings
FIG. 1 is a schematic diagram of a two-channel converged network architecture of the present invention;
FIG. 2 is a system authentication process of the present invention;
FIG. 3 is a system registration process of the present invention;
FIG. 4 is a system reset flow of the present invention;
FIG. 5 is a schematic diagram of several forms of a signal reset scheme;
FIG. 6 is a schematic diagram of a confusion matrix obtained from testing using the system of the present invention.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and preferred embodiments, and the objects and effects of the present invention will become more apparent, it being understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.
The invention relates to a vibration motor equipment fingerprint extraction and identification method based on homologous signals, which needs to relate to a terminal and a cloud server, wherein a motor and an inertia sensing unit are arranged in the terminal, and the inertia sensing unit is used for collecting acceleration and angular velocity signals in the vibration process of the motor; the terminal also comprises a preprocessing module and a terminal communication module. The preprocessing module is used for preprocessing acceleration and angular velocity signals in the vibration process of the motor, dividing the signals by the start time mark and the end time mark provided during collection, filtering the signals through a high-pass filter, and aligning the signals through cubic spline interpolation; the preprocessing module outputs a signal input by a two-channel fusion network for the cloud server, and sends the signal to the cloud server through the terminal communication module, and meanwhile, the terminal sends an authentication request and a registration request to the cloud server, as shown in fig. 2 and 3.
The cloud server simultaneously inputs the acceleration and angular velocity signals into a trained two-channel fusion network consisting of a residual block, a Dropout layer, a full connection layer and a loss function to obtain a motor fingerprint;
the cloud server comprises a fingerprint extraction module, a fingerprint registration identification module and a cloud communication module; the cloud communication module receives a signal sent by the terminal, transmits the signal to the fingerprint extraction module, inputs the signal into a two-channel fusion network consisting of a residual block, a Dropout layer, a full connection layer and a loss function, realizes mapping between the signal and a motor fingerprint, outputs the motor fingerprint to the fingerprint registration identification module, and classifies the motor fingerprint through a classifier in the fingerprint registration identification module. As shown in fig. 1.
The two-channel fusion network of the cloud server comprises a convolution layer 1, a pooling layer, a convolution layer 2, a convolution layer 3, a convolution layer 4, a convolution layer 5, a flattening layer and 1-3 full-connection layers; the network has two inputs of acceleration and angular velocity signals;
convolutional layer 1 is composed of 64 convolutional kernels of size 5 × 1, with a convolution step size of 2;
the pooling layer is composed of a largest pooling layer with a size of 3 × 1, and the convolution step length is 2;
the convolution layer 2 is composed of 3 residual blocks with 64 1 × 1 convolution kernels, 64 1 × 3 convolution kernels and 256 1 × 1 convolution kernels arranged in sequence, and a Dropout layer with p equal to 0.2 is added at the end of each residual block;
the convolution layer 3 is composed of 4 residual blocks with 128 1 × 1 convolution kernels, 128 1 × 3 convolution kernels and 512 1 × 1 convolution kernels arranged in sequence, and a Dropout layer with p equal to 0.2 is added at the end of each residual block;
the convolution layer 4 is composed of 6 residual blocks having 256 1 × 1 convolution kernels, 256 1 × 3 convolution kernels, and 1024 1 × 1 convolution kernels, which are sequentially arranged, and a Dropout layer with p equal to 0.2 is finally added to each residual block;
the convolution layer 5 is composed of 3 residual blocks with 512 1 × 1 convolution kernels, 512 1 × 3 convolution kernels and 2048 1 × 1 convolution kernels arranged in sequence, and a Dropout layer with p equal to 0.2 is added at the end of each residual block;
the flattening layer spreads all the outputs into a one-dimensional vector;
the three fully-connected layers 1-3 are composed of fully-connected layers with different numbers of neurons. See table 1 for details.
In order to achieve the effects of similar aggregation and heterogeneous repulsion during fingerprint identification, a generalized loss function is selected. The generalized loss function is as follows:
wherein e isijIs the j data of the i device in batch processing, Sij，kIs eijAnd the cosine similarity between the average of all vectors of device No. k.
The two-channel fusion network respectively inputs the acceleration signals and the angular speed signals into a network sharing parameters through the network structure, and the signals are combined into a one-dimensional vector after the flattening layer and enter the full connection layer.
The classifier can select artificial intelligence methods such as a support vector machine, a decision tree and a Bayesian network, and identification methods based on threshold values such as cosine similarity and Euclidean distance.
Therefore, the method for extracting and identifying the fingerprint of the vibration motor equipment based on the homologous signal comprises the following specific steps:
acquiring acceleration and angular velocity signals of a terminal in a motor vibration process through an inertial sensing unit of the terminal, segmenting, filtering and aligning the acceleration and angular velocity signals by the terminal, sending the signals to a cloud server, and sending an authentication request or a registration request to the cloud server by the terminal;
the cloud server simultaneously inputs the acceleration and angular velocity signals into a trained two-channel fusion network consisting of a residual block, a Dropout layer, a full connection layer and a loss function to obtain a motor fingerprint; classifying the motor fingerprints by adopting a classifier; if the classification is successful, the cloud server outputs a classification result; if the category does not exist, at the moment, if the terminal sends a registration request, the cloud server stores the motor fingerprint and updates a motor fingerprint storage database; and if the terminal sends the authentication request, the cloud server directly refuses the authentication.
The motor fingerprint needs to be reset because the motor fingerprint is forgotten or leaked in reality. In order to conveniently realize the fingerprint resetting of the motor, the terminal further includes a fingerprint resetting module, configured to receive a signal resetting scheme input by a user, reset the vibration mode of the motor according to the resetting scheme, and send a resetting request to the cloud server, as shown in fig. 4. The signal resetting scheme comprises any one or more of frequency peak value resetting, frequency change resetting and vibration interval resetting;
the frequency peak value is reset to change the fingerprint of the motor by changing the main frequency of the excitation signal;
the frequency change is reset to change the fingerprint of the motor by changing the frequency change of the excitation signal;
the vibration interval resetting changes the fingerprint of the motor by changing the output interval of the excitation signal, including dividing the acquisition time into equal-length time periods and setting whether the excitation signal is output or not. Specifically, as shown in fig. 5, four reset schemes are provided, in which the upper left corner is an excitation signal before resetting, the upper right corner is an excitation signal after setting a frequency peak, the lower left corner is an excitation signal after setting a frequency change, and the lower right corner is an excitation signal after setting a vibration interval; since the vibration signal of the motor changes along with the change of the excitation signal, the resetting of the fingerprint of the motor can be realized through the scheme.
A specific example is given below to illustrate the advantages of the identification method of the present invention.
In this embodiment, acceleration and angular velocity signals of 90 motors (including 80 independent motors and 10 mobile phones) in the vibration process are collected and preprocessed to be 3 × 200, the acceleration and angular velocity signals are successively input to the convolution layers 1-5 in the same training or use process, the single signal output is 1 × 6144, and the signals are combined into a vector with the size of 1 × 12288 on the flattening layer. As shown in fig. 6, the confusion matrix demonstrates that for a 90 motor test, the abscissa of the confusion matrix is the predicted equipment number and the ordinate is the actual equipment number, the darker the grid representing the greater the number of results falling in this grid. The accuracy of the identification reached 98.5%.
It will be understood by those skilled in the art that the foregoing is only a preferred embodiment of the present invention, and is not intended to limit the invention, and although the invention has been described in detail with reference to the foregoing examples, it will be apparent to those skilled in the art that various changes in the form and details of the embodiments may be made and equivalents may be substituted for elements thereof. All modifications, equivalents and the like which come within the spirit and principle of the invention are intended to be included within the scope of the invention.
Claims (3)
1. A vibration motor device fingerprint extraction and identification method based on homologous signals is characterized by comprising the following steps:
acquiring acceleration and angular velocity signals of a terminal in a motor vibration process through an inertial sensing unit of the terminal, segmenting, filtering and aligning the acceleration and angular velocity signals by the terminal, sending the signals to a cloud server, and sending an authentication request or a registration request to the cloud server by the terminal;
the cloud server simultaneously inputs the acceleration and angular velocity signals into a trained two-channel fusion network composed of a residual block, a Dropout layer, a full connection layer and a loss function to obtain a motor fingerprint; classifying the motor fingerprints by adopting a classifier; if the classification is successful, the cloud server outputs a classification result; if the type does not exist, at the moment, if the terminal sends a registration request, the cloud server stores the motor fingerprint and updates a motor fingerprint storage database; and if the terminal sends an authentication request, the cloud server directly refuses authentication.
2. The vibration motor apparatus fingerprint extraction and identification method based on homologous signals according to claim 1, wherein the two-channel fusion network comprises a convolution layer, a pooling layer, a flattening layer and a full-link layer, the network has two inputs, the flattening layer expands two outputs into a one-dimensional vector, and the loss function is a generalized loss function.
3. The vibration motor device fingerprint extraction and identification method based on homologous signals according to claim 1, wherein the two-channel fusion network comprises a convolutional layer 1, a pooling layer, a convolutional layer 2, a convolutional layer 3, a convolutional layer 4, a convolutional layer 5, a flattening layer and three fully-connected layers;
the convolutional layer 1 is composed of 64 convolutional kernels with the size of 5 multiplied by 1, and the convolution step length is 2;
the pooling layer is composed of a largest pooling layer with the size of 3 multiplied by 1, and the convolution step length is 2;
the convolutional layer 2 is composed of 3 residual blocks with 64 1 × 1 convolution kernels, 64 1 × 3 convolution kernels and 256 1 × 1 convolution kernels arranged in sequence, and a Dropout layer with p equal to 0.2 is added at the end of each residual block;
the convolution layer 3 is composed of 4 residual blocks with 128 1 × 1 convolution kernels, 128 1 × 3 convolution kernels and 512 1 × 1 convolution kernels arranged in sequence, and a Dropout layer with p equal to 0.2 is added at the end of each residual block;
the convolutional layer 4 is composed of 6 residual blocks with 256 1 × 1 convolution kernels, 256 1 × 3 convolution kernels and 1024 1 × 1 convolution kernels, wherein the residual blocks are sequentially arranged, and a Dropout layer with p equal to 0.2 is finally added to each residual block;
the convolutional layer 5 is composed of 3 residual blocks with 512 1 × 1 convolution kernels, 512 1 × 3 convolution kernels and 2048 1 × 1 convolution kernels arranged in sequence, and a Dropout layer with p equal to 0.2 is added at the end of each residual block;
the flattening layer spreads all the outputs into a one-dimensional vector;
the three fully-connected layers are composed of fully-connected layers with different neuron numbers.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110292483.7A CN113111726B (en) | 2021-03-18 | 2021-03-18 | Vibration motor equipment fingerprint extraction and identification method based on homologous signals |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110292483.7A CN113111726B (en) | 2021-03-18 | 2021-03-18 | Vibration motor equipment fingerprint extraction and identification method based on homologous signals |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113111726A true CN113111726A (en) | 2021-07-13 |
| CN113111726B CN113111726B (en) | 2022-10-28 |
Family
ID=76711781
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110292483.7A Active CN113111726B (en) | 2021-03-18 | 2021-03-18 | Vibration motor equipment fingerprint extraction and identification method based on homologous signals |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113111726B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106874852A (en) * | 2017-01-13 | 2017-06-20 | 浙江大学 | A kind of device-fingerprint based on acceleration transducer is extracted and recognition methods |
| US20170346815A1 (en) * | 2016-05-31 | 2017-11-30 | International Business Machines Corporation | Multifactor authentication processing using two or more devices |
| CN109766683A (en) * | 2019-01-16 | 2019-05-17 | 中国科学技术大学 | A kind of guard method of intelligent movable device sensor fingerprint |
| CN109766855A (en) * | 2019-01-16 | 2019-05-17 | 中国科学技术大学 | A kind of intelligent movable device sensor fingerprint identification method |
| US20200005019A1 (en) * | 2018-06-28 | 2020-01-02 | Beijing Kuangshi Technology Co., Ltd. | Living body detection method, system and non-transitory computer-readable recording medium |
| CN110796175A (en) * | 2019-09-30 | 2020-02-14 | 武汉大学 | Electroencephalogram data online classification method based on light convolutional neural network |
| CN111984960A (en) * | 2020-07-13 | 2020-11-24 | 深圳市捷讯云联科技有限公司 | Privacy protection equipment identification model design and use method based on homomorphic encryption |
| CN112187373A (en) * | 2020-08-28 | 2021-01-05 | 浙江大学 | Concealed channel communication method based on gyroscope resonance |
| CN112218294A (en) * | 2020-09-08 | 2021-01-12 | 深圳市燃气集团股份有限公司 | 5G-based access method and system for Internet of things equipment and storage medium |
-
2021
- 2021-03-18 CN CN202110292483.7A patent/CN113111726B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20170346815A1 (en) * | 2016-05-31 | 2017-11-30 | International Business Machines Corporation | Multifactor authentication processing using two or more devices |
| CN106874852A (en) * | 2017-01-13 | 2017-06-20 | 浙江大学 | A kind of device-fingerprint based on acceleration transducer is extracted and recognition methods |
| US20200005019A1 (en) * | 2018-06-28 | 2020-01-02 | Beijing Kuangshi Technology Co., Ltd. | Living body detection method, system and non-transitory computer-readable recording medium |
| CN109766683A (en) * | 2019-01-16 | 2019-05-17 | 中国科学技术大学 | A kind of guard method of intelligent movable device sensor fingerprint |
| CN109766855A (en) * | 2019-01-16 | 2019-05-17 | 中国科学技术大学 | A kind of intelligent movable device sensor fingerprint identification method |
| CN110796175A (en) * | 2019-09-30 | 2020-02-14 | 武汉大学 | Electroencephalogram data online classification method based on light convolutional neural network |
| CN111984960A (en) * | 2020-07-13 | 2020-11-24 | 深圳市捷讯云联科技有限公司 | Privacy protection equipment identification model design and use method based on homomorphic encryption |
| CN112187373A (en) * | 2020-08-28 | 2021-01-05 | 浙江大学 | Concealed channel communication method based on gyroscope resonance |
| CN112218294A (en) * | 2020-09-08 | 2021-01-12 | 深圳市燃气集团股份有限公司 | 5G-based access method and system for Internet of things equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113111726B (en) | 2022-10-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109086669B (en) | Face recognition identity verification method and device and electronic equipment | |
| KR101629224B1 (en) | Authentication method, device and system based on biological characteristics | |
| CN111914812B (en) | Image processing model training method, device, equipment and storage medium | |
| CN107483416A (en) | The method and device of authentication | |
| CN113515988B (en) | Palm print recognition method, feature extraction model training method, device and medium | |
| CN105160739A (en) | Automatic identification equipment, automatic identification method and door control system | |
| CN109741477A (en) | Work attendance management system, method and electronic equipment | |
| US10997609B1 (en) | Biometric based user identity verification | |
| CN106303599A (en) | A kind of information processing method, system and server | |
| CN111625793B (en) | Identification, order payment and sub-face library establishment method, device and equipment and order payment system | |
| CN111177469A (en) | Face retrieval method and face retrieval device | |
| CN110443181A (en) | Face identification method and device | |
| CN110489659A (en) | Data matching method and device | |
| CN111738199A (en) | Image information verification method, image information verification device, image information verification computing device and medium | |
| CN107656959B (en) | Message leaving method and device and message leaving equipment | |
| CN113869398B (en) | Unbalanced text classification method, device, equipment and storage medium | |
| CN110874602A (en) | Image identification method and device | |
| CN113111726B (en) | Vibration motor equipment fingerprint extraction and identification method based on homologous signals | |
| CN116386091B (en) | Fingerprint identification method and device | |
| CN113111725B (en) | Vibration motor equipment fingerprint extraction and identification system based on homologous signals | |
| CN116246303A (en) | Sample construction method, device, equipment and medium for model cross-domain training | |
| CN111062345A (en) | Training method and device of vein recognition model and vein image recognition device | |
| CN115563597A (en) | Artificial intelligence operation system and operation method based on big data | |
| CN112905816B (en) | Iris search recognition method and device, processor and electronic device | |
| CN113641844A (en) | Identity verification method, device, terminal and storage medium for marital registration system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |