CN110414373B - Deep learning palm vein recognition system and method based on cloud edge-side cooperative computing - Google Patents
Deep learning palm vein recognition system and method based on cloud edge-side cooperative computing Download PDFInfo
- Publication number
- CN110414373B CN110414373B CN201910609754.XA CN201910609754A CN110414373B CN 110414373 B CN110414373 B CN 110414373B CN 201910609754 A CN201910609754 A CN 201910609754A CN 110414373 B CN110414373 B CN 110414373B
- Authority
- CN
- China
- Prior art keywords
- palm vein
- model
- user
- training
- deep learning
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/94—Hardware or software architectures specially adapted for image or video understanding
- G06V10/95—Hardware or software architectures specially adapted for image or video understanding structured as a network, e.g. client-server architectures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/14—Vascular patterns
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Computer Interaction (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a deep learning palm vein recognition system based on cloud edge-end cooperative computing, which mainly comprises a cloud computing layer, an edge computing layer and a terminal layer, wherein the computing and storage capacity of the whole recognition system is determined by the combination of the three layers; the back-end cloud storage module can also realize rapid deployment and unified management of the palm vein recognition system according to the resource configuration and management capability of the cloud platform.
Description
Technical Field
The invention relates to the technical field of pattern recognition, in particular to a deep learning palm vein recognition system and method based on cloud-edge cooperative computing.
Background
In recent years, with the advent of the "intelligent age", AI technology is increasingly applied to the field of biometric identification, which has led to great improvements in biometric identification technology in terms of security, convenience, accuracy, and the like. The palm vein recognition is used as a globally unique second-generation biological characteristic recognition technology and has the unique advantages of anti-counterfeiting performance, accuracy and anti-interference performance.
CNN is called Convolutional Neural Networks, and Chinese is interpreted as a Convolutional Neural network, which is a deep feedforward Neural network with characteristics of local connection, weight sharing, convergence and the like. As a branch of deep neural networks, CNNs have good applications in computer vision, natural language processing, and data mining in software engineering. Particularly, the performance is more prominent in computer vision, and the aspects of Object detection (Object detection), Image classification (Image classification), Image retrieval (Image retrieval), Image semantic segmentation (Image semantic segmentation) and the like show better performance compared with the traditional method and even some other deep neural network models.
The inventor of the present application finds that the method of the prior art has at least the following technical problems in the process of implementing the present invention:
at present, a palm vein recognition system which is popular in the market is mainly carried on a local server or a remote cloud server, so that attendance checking, entrance guard recognition and identity authentication of a user are realized. Meanwhile, the deep learning identification engine with high identification precision is relatively high in calculation complexity, and the large cloud computing server is located at the far end of a terminal generated data and application model, so that the identification model trained through deep learning cannot be rapidly issued and applied to the terminal identification equipment, the feedback of the terminal cannot be rapidly obtained and timely modified, and finally the palm vein identification system cannot rapidly respond to the user requirements.
Disclosure of Invention
In view of this, the invention provides a deep learning palm vein recognition system and method based on cloud-edge collaborative computing, which are used for solving or at least partially solving the problems of low system recognition accuracy and low response speed in the prior art.
In order to solve the above technical problem, a first aspect of the present invention provides a deep learning palm vein recognition system based on cloud-edge collaborative computing, including:
the cloud storage module is used for deploying a deep learning palm vein recognition algorithm, training a deep convolution neural network model to obtain a deep learning palm vein recognition model, and storing user information;
the palm vein collection and identification device comprises a collection module, an edge calculation module and a receiving module, wherein the collection module is used for collecting a palm vein image of a user, the edge calculation module is used for calling a deep learning palm vein identification model to verify the collected palm vein image, if the verification is successful, verification information is uploaded to a cloud storage module, if the verification is failed, information input by the user is received through the receiving module, the information input by the user is matched with user information stored in the cloud storage module, whether the user is a newly registered user who does not collect the verified palm vein is judged, if the user is the newly registered user, the palm vein deep learning palm vein identification model of the newly registered user is used for updating, and the updated palm vein identification model is obtained;
and the client is used for providing a login interface for the user to inquire the corresponding identification information.
In one embodiment, the cloud storage module is further configured to:
and processing the request of the client, and making a corresponding response by judging different requests.
Based on the same inventive concept, a second aspect of the present invention provides a palm vein recognition method based on the recognition system of the first aspect, including:
step S1: deploying a deep learning palm vein recognition algorithm through a cloud storage module, training a deep convolution neural network model to obtain a deep learning palm vein recognition model, and storing user information;
step S2: collecting a palm vein image of a user through a collection module, calling a deep learning palm vein recognition model through an edge calculation module to verify the collected palm vein image, wherein verification information is uploaded to a cloud storage module when verification is successful, receiving information input by the user through a receiving module when verification is failed, matching the information input by the user with user information stored in the cloud storage module, judging whether the user is a newly registered user who does not collect the verification palm vein, and if the user is the newly registered user, updating the deep learning palm vein recognition model by using the newly registered user palm vein image to obtain an updated palm vein recognition model;
step S3: and providing a login interface through the client for the user to inquire corresponding identification information.
In one embodiment, the method further comprises:
identifying the palm vein image of the new registered user based on the updated palm vein identification model, and returning the palm vein image information of the new registered user to the cloud storage module;
matching the palm vein image information of the newly registered user with the registration information of the user, and if the matching is successful, returning to the palm vein acquisition and identification equipment to prompt that the input is successful; and if the matching fails, returning to the palm vein acquisition and identification equipment, prompting the input failure, and displaying the instruction information of re-acquiring the palm vein image.
In one embodiment, in step S1, the obtaining process of the deep learning palm vein recognition model includes:
step S1.1: inputting the obtained palm vein data set into a cloud storage module, wherein the palm vein data set is divided into a training set, a testing set and a verification set;
step S1.2: in the model training stage, inputting a training set and labels corresponding to images in the training set into a deep convolutional neural network model, and outputting results after training is finished;
step S1.3: in the model test stage, inputting a test set of the palm veins into a deep convolutional neural network model to test the recognition rate of the model;
step S1.4: in the model verification stage, the recognition condition of the model is verified in the deep convolutional neural network model, and the classifier parameters and the complexity of the model are adjusted according to the recognition condition to obtain a deep learning palm vein recognition model with the classifier parameters and the complexity of the model meeting the conditions.
In one embodiment, the step S2 of the edge calculation module invoking a deep learning palm vein recognition model to verify the collected palm vein image specifically includes:
step S2.1: preprocessing the collected palm vein image of the user;
step S2.2: calling a deep learning palm vein recognition model, and performing similarity calculation on the preprocessed palm vein image;
step S2.3: and according to the similarity calculation result, verifying whether the similarity of the palm veins reaches a set threshold value through a decision maker, and verifying the collected palm vein image.
In one embodiment, step S1.2 specifically includes:
step S1.2.1: initializing variable training times of the deep convolutional neural network model;
step S1.2.2: selecting a part of palm vein training set for batch processing;
step S1.2.3: when the model is trained, the whole model is subjected to forward propagation to obtain a predicted value;
step S1.2.4: after the forward propagation is finished, the model performs backward propagation and updates the variable in the training process;
step S1.2.5: detecting whether a training target is reached, and if the training target is reached, finishing the training; if not, go to step S1.2.6;
step S1.2.6: testing whether the training times are reached, and if the training times are reached, finishing the training; if not, go to step S1.2.7;
step S1.2.7: and (4) increasing the training times, and then repeating the steps S1.2.1-S1.2.6 again until the training requirements are met, and finishing the training.
One or more technical solutions in the embodiments of the present application have at least one or more of the following technical effects:
the invention provides a deep learning palm vein recognition system based on cloud edge cooperative computing, which comprises: the method comprises the steps of training a deep convolutional neural network model to obtain a deep learning palm vein recognition model, storing user information, acquiring palm vein images of a user, calling the deep learning palm vein recognition model to verify and update the acquired palm vein images, providing a login interface, and providing a client side for the user to inquire corresponding recognition information. Further provides a palm vein recognition method based on the deep learning palm vein recognition system.
In the system and the method provided by the invention, the cloud storage module can provide computing resources required by training for the deep convolutional neural network model, the trained model is transmitted to the edge computing module of the palm vein acquisition equipment, the acquired palm vein image is verified and identified, the acquisition equipment collects the verification and identification conditions of the deep convolutional neural network model and feeds back the verification and identification conditions to the cloud storage module through the edge computing module for updating the deep learning palm vein identification model, so that the accuracy of the finally obtained palm vein identification model is greatly improved through continuous training, application, feedback and retraining and repeated effective iterative training; the back-end cloud storage module can also realize rapid deployment and unified management of the palm vein recognition system according to the resource configuration and management capability of the cloud platform.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a system architecture diagram of the present invention;
fig. 2 is a flowchart of a palm vein recognition method provided in the present invention;
FIG. 3 is a flow chart of the entire palm vein recognition system of the present invention;
FIG. 4 is a schematic view of a cloud storage module of the present invention;
FIG. 5 is a schematic diagram of an edge calculation module according to the present invention.
FIG. 6 is a schematic diagram of a palm vein recognition model architecture according to the present invention;
FIG. 7 is a flow chart of palm vein recognition model training of the present invention;
Detailed Description
The invention aims to provide a cloud-edge-collaborative-computation-based deep learning palm vein recognition system and method, aiming at the problems of low recognition accuracy and low response speed of the system in the prior art, the deep learning palm vein recognition model with high recognition accuracy is applied to common embedded terminal equipment for palm vein collection and recognition through a cloud-edge-end collaborative computation framework, and recognized information is issued to a mobile phone client or a webpage end in real time through a cloud server, so that the technical effects of improving the recognition accuracy and the response speed of the model are achieved.
In order to achieve the technical effects, the invention mainly comprises the following concepts:
the invention discloses a cloud-edge-side-collaborative-computation-based deep-learning palm vein recognition system, wherein a cloud-edge-side-collaborative-computation framework mainly comprises a cloud computing layer (a cloud storage module), an edge computing layer (palm vein acquisition equipment) and a terminal layer (a client), and the computing and storage capacities of the whole recognition system are determined through the combination of the cloud computing layer and the edge computing layer. The whole system comprises a cloud storage module, palm vein acquisition and identification equipment and a client. The invention also discloses a deep learning palm vein identification method based on cloud edge end cooperative computing. The cloud storage module can provide computing resources required by training for the deep convolutional neural network model, the trained model is transmitted to the edge computing module of the palm vein acquisition equipment, the palm vein acquisition equipment collects the recognition condition of the deep convolutional neural network model and feeds the recognition condition back to a deep convolutional neural network model chip on the cloud storage module through the edge computing module, and therefore the accuracy of the finally obtained palm vein recognition model is greatly improved through continuous training, application, feedback and retraining and repeated effective iterative training; the back-end cloud computing storage module can also realize rapid deployment and unified management of the palm vein recognition system according to the resource configuration and management capability of the cloud platform.
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Example one
The embodiment provides a deep learning palm vein recognition system based on cloud-edge collaborative computing, and the system comprises:
the cloud storage module is used for deploying a deep learning palm vein recognition algorithm, training a deep convolution neural network model to obtain a deep learning palm vein recognition model, and storing user information;
the palm vein collection and identification device comprises a collection module, an edge calculation module and a receiving module, wherein the collection module is used for collecting a palm vein image of a user, the edge calculation module is used for calling a deep learning palm vein identification model to verify the collected palm vein image, if the verification is successful, verification information is uploaded to a cloud storage module, if the verification is failed, information input by the user is received through the receiving module, the information input by the user is matched with user information stored in the cloud storage module, whether the user is a newly registered user who does not collect the verified palm vein is judged, if the user is the newly registered user, the palm vein deep learning palm vein identification model of the newly registered user is used for updating, and the updated palm vein identification model is obtained;
and the client is used for providing a login interface for the user to inquire the corresponding identification information.
Specifically, a chip (namely an edge calculation module) for edge calculation is built in the palm vein acquisition and identification device, so that the preprocessing of the palm vein image and the updating of the deep learning palm vein identification model can be realized.
The edge computing module mainly provides computing resources required by recognition for the deep convolutional neural network model, the model trained by the cloud computing center is transmitted to the terminal device to be used for recognizing users, the terminal device collects the recognition conditions of the deep convolutional neural network model and feeds the recognition conditions back to a deep convolutional neural network model chip on the edge computing module, and therefore continuous training, application, feedback and retraining are achieved, and the accuracy of the finally obtained palm vein recognition model is greatly improved.
And the back-end cloud storage module is mainly used for training the deep neural network model and storing user data, and can also realize the deployment and management of the palm vein recognition system according to the resource configuration and management capability of the cloud platform.
The client comprises a webpage end or a palm vein recognition management system of the client, the system mainly comprises an administrator interface and a common user interface, the administrator can check, update, delete and the like the information of the common user after registering and logging in the client, and after registering and logging in the common user, the common user can only inquire the registration information, the login time, the time record of the palm vein recognition verification success and other related information.
Wherein, the cloud storage module is further configured to:
and processing the request of the client, and making a corresponding response by judging different requests.
Specifically, if the administrator of the web page end and the authority of the ordinary user are different, when the server at the cloud end interacts with the web page end, the cloud computing storage module processes the corresponding request according to the message protocol header.
As shown in fig. 1, the present invention provides a system architecture diagram for deep learning palm vein recognition based on "cloud-side" collaborative computing, which includes: the terminal palm vein collection and identification device comprises a terminal palm vein collection and identification device, an edge server (an edge computing module), a cloud computing center (a rear-end cloud storage module) and clients (a mobile client and a webpage end).
Based on the same inventive concept, the invention also provides an identification method based on the deep learning palm vein identification system in the first embodiment, which is specifically referred to in the second embodiment.
Example two
The embodiment provides a deep learning palm vein identification method based on cloud-edge collaborative computing, please refer to fig. 2, and the method includes:
step S1: deploying a deep learning palm vein recognition algorithm through a cloud storage module, training a deep convolution neural network model to obtain a deep learning palm vein recognition model, and storing user information;
step S2: collecting a palm vein image of a user through a collection module, calling a deep learning palm vein recognition model through an edge calculation module to verify the collected palm vein image, wherein verification information is uploaded to a cloud storage module when verification is successful, receiving information input by the user through a receiving module when verification is failed, matching the information input by the user with user information stored in the cloud storage module, judging whether the user is a newly registered user who does not collect the verification palm vein, and if the user is the newly registered user, updating the deep learning palm vein recognition model by using the newly registered user palm vein image to obtain an updated palm vein recognition model;
step S3: and providing a login interface through the client for the user to inquire corresponding identification information.
Specifically, the embodiment of the invention provides a deep learning palm vein recognition system based on cloud-side collaborative computing, which realizes rapid and efficient deployment and recognition of palm veins based on a deep learning model through mutual collaborative computing of a cloud computing layer, an edge computing layer and a terminal layer, and simultaneously issues recognition information of a terminal to a mobile client or a webpage end in real time by means of a cloud computing center. And the implementation II is a deep learning model training method based on cloud edge end cooperative computing. Related computing resources are provided through the cloud computing layer and the edge computing layer, and rapid training of the deep learning model is achieved.
In one embodiment, the method further comprises:
identifying the palm vein image of the new registered user based on the updated palm vein identification model, and returning the palm vein image information of the new registered user to the cloud storage module;
matching the palm vein image information of the newly registered user with the registration information of the user, and if the matching is successful, returning to the palm vein acquisition and identification equipment to prompt that the input is successful; and if the matching fails, returning to the palm vein acquisition and identification equipment, prompting the input failure, and displaying the instruction information of re-acquiring the palm vein image.
In one embodiment, the step S2 of the edge calculation module invoking a deep learning palm vein recognition model to verify the collected palm vein image specifically includes:
step S2.1: preprocessing the collected palm vein image of the user;
step S2.2: calling a deep learning palm vein recognition model, and performing similarity calculation on the preprocessed palm vein image;
step S2.3: and according to the similarity calculation result, verifying whether the similarity of the palm veins reaches a set threshold value through a decision maker, and verifying the collected palm vein image.
As shown in fig. 3, the present invention provides a flow chart of the whole system for deep learning palm vein recognition based on "cloud-side" collaborative computing. The specific process is as follows:
(I) the palm vein acquisition and identification equipment acquires a palm vein image of a user through a camera (namely an acquisition module) of a near infrared light source;
(II) the palm vein acquisition and identification equipment verifies the palm vein information of the user by calling an identification model in the edge computing module, whether the verification is successful is prompted on a display screen of the palm vein acquisition and identification equipment, and if the verification is successful, user verification information is uploaded to the cloud storage module, wherein the user verification information comprises user basic information and verification time; if the verification fails, the user can input the login account and the password of the user on the acquisition and identification device, then the information input by the user is matched with the user information stored in the cloud storage module database, whether the user is a newly registered user who does not acquire the verification palm vein is judged, and if the input account or the password is wrong, the palm vein acquisition device automatically exits the system and returns to the initial interface; if the prompt is successful, the step (III), (IV) and (V) are carried out;
(III) uploading the collected palm vein image of the user to an edge calculation module of the collection equipment;
(IV) calling a deep learning model of palm vein recognition trained in a cloud computing center by an edge computing module in the palm vein collection and recognition equipment to train collected palm vein images, completing the training of uploaded data (wherein a large amount of early training of a deep neural network model is iterated at the cloud end, so that the early training belongs to forward reasoning on an edge computing chip and is used for recognition and verification), and returning the result to the palm vein collection and recognition equipment;
(V) the palm vein collection and identification equipment uses the updated deep learning model to identify a palm vein image of a new user, returns image information to a management system database of the cloud storage module, and matches the image information with registration information of the user, wherein the registration information of the user comprises registered ID, name, user palm vein data, verification time and validity, and after matching is successful, the registration information is returned to the collection equipment to prompt that entry is successful; if the matching fails, returning to the acquisition equipment to prompt the input failure, and displaying instruction information for asking to acquire the palm vein image again;
(VI) the user can inquire the personal information in the cloud management system through a login webpage interface, and the login is divided into an administrator and a common user; the administrator can operate the information of the user, and if the administrator deletes the registration information of the user, the user cannot realize palm vein authentication. When a common user registers, the account with the ID number can be obtained, and the user can acquire a login interface of the identification device at a webpage end or a terminal and inquire the identification information of the user, including the verification times of the user and the specific time of each verification.
Fig. 4 is a schematic diagram of a cloud storage module. The cloud storage module has two main functions: firstly, carrying and training a deep neural network model, updating and issuing the model to edge equipment (namely palm vein acquisition and identification equipment) for the acquisition and identification equipment to verify palm vein information; and secondly, processing information transmitted by the acquisition and identification device and interaction information of a network terminal as a brain of the whole palm vein recognition system, and storing related data in a background server at the cloud.
The whole cloud storage module design specifically comprises the following parts:
(1) and deploying a deep learning palm vein recognition algorithm, training a palm vein image sample library through the neural network, and packaging the palm vein image sample library into a deep learning palm vein recognition model after continuous iteration. Then, the model is transmitted to an edge calculation module through a TCP/IP protocol, and the edge calculation module is used for rapidly verifying the palm vein information;
(2) processing information from an acquisition equipment end, and uploading verification information to a cloud storage module after the edge computing module passes verification; if the verification of the edge computing module is not passed, the edge computing module requests the cloud computing storage module to determine whether the cloud computing storage module is a newly registered user, and the newly registered user needs to upload palm vein image acquisition information passing the verification to the cloud storage module;
(3) and processing the request of the webpage end, and making a corresponding response by judging different requests. If the administrator of the webpage end and the authority of the common user are different, when the server of the cloud end interacts with the webpage end, the cloud computing storage module can process corresponding requests according to the message protocol header.
As shown in fig. 5, the edge calculation module provided by the present invention is schematically illustrated. The edge calculation module comprises a deep learning palm vein recognition model which is trained in a cloud computing center and is used for preprocessing collected images, downloading and updating the collected images, carrying out similarity calculation on the collected palm vein images and verifying whether the similarity of the palm veins reaches a set threshold through a decision maker, wherein the set threshold can be adjusted according to conditions, and is set to be 0.7, 0.8 and the like.
In one embodiment, in step S1, the obtaining process of the deep learning palm vein recognition model includes:
step S1.1: inputting the obtained palm vein data set into a cloud storage module, wherein the palm vein data set is divided into a training set, a testing set and a verification set;
step S1.2: in the model training stage, inputting a training set and labels corresponding to images in the training set into a deep convolutional neural network model, and outputting results after training is finished;
step S1.3: in the model test stage, inputting a test set of the palm veins into a deep convolutional neural network model to test the recognition rate of the model;
step S1.4: in the model verification stage, the recognition condition of the model is verified in the deep convolutional neural network model, and the classifier parameters and the complexity of the model are adjusted according to the recognition condition to obtain a deep learning palm vein recognition model with the classifier parameters and the complexity of the model meeting the conditions.
Specifically, as shown in FIG. 6, is an architecture for identifying a model. The matching condition is the convergence of the classifier parameters, the complexity of the model is low, and the model can be specifically adjusted according to the situation. In the model verification stage, according to the recognition condition, the classifier parameters and the complexity of the model are adjusted, the result is returned, and the model with high palm vein recognition rate can be obtained through continuous training.
In one embodiment, step S1.2 specifically includes:
step S1.2.1: initializing variable training times of the deep convolutional neural network model;
step S1.2.2: selecting a part of palm vein training set for batch processing;
step S1.2.3: when the model is trained, the whole model is subjected to forward propagation to obtain a predicted value;
step S1.2.4: after the forward propagation is finished, the model performs backward propagation and updates the variable in the training process;
step S1.2.5: detecting whether a training target is reached, and if the training target is reached, finishing the training; if not, go to step S1.2.6;
step S1.2.6: testing whether the training times are reached, and if the training times are reached, finishing the training; if not, go to step S1.2.7;
step S1.2.7: and (4) increasing the training times, and then repeating the steps S1.2.1-S1.2.6 again until the training requirements are met, and finishing the training.
As shown in fig. 7, a flow chart of training the palm vein recognition model according to the present invention is shown, and a recognition model meeting the requirements can be obtained through the above flow chart.
Generally, compared with the prior art, the technical scheme of the invention has the following advantages and beneficial effects:
computing resources required by training can be provided for the deep convolutional neural network model through the cloud storage module, the trained model is transmitted to the edge computing module of the palm vein acquisition equipment, the acquired palm vein image is verified and identified, the acquisition equipment collects the verification and identification conditions of the deep convolutional neural network model and feeds the verification and identification conditions back to the cloud storage module through the edge computing module to update the deep learning palm vein identification model, and therefore the accuracy of the finally obtained palm vein identification model is greatly improved through continuous training, application, feedback and retraining and repeated effective iterative training; the back-end cloud storage module can also realize rapid deployment and unified management of the palm vein recognition system according to the resource configuration and management capability of the cloud platform.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various modifications and variations can be made in the embodiments of the present invention without departing from the spirit or scope of the embodiments of the invention. Thus, if such modifications and variations of the embodiments of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to encompass such modifications and variations.
Claims (7)
1. The utility model provides a deep learning palm vein identification system based on cloud limit-side cooperative computing which characterized in that includes:
the cloud storage module is used for carrying a deep learning palm vein recognition model, training the deep convolution neural network model to obtain the deep learning palm vein recognition model and storing user information;
the palm vein collection and identification device comprises a collection module, an edge calculation module and a receiving module, wherein the collection module is used for collecting a palm vein image of a user, the edge calculation module is used for calling a deep learning palm vein identification model to verify the collected palm vein image, if the verification is successful, verification information is uploaded to a cloud storage module, if the verification is failed, information input by the user is received through the receiving module, the information input by the user is matched with user information stored in the cloud storage module, whether the user is a newly registered user who does not collect the verified palm vein is judged, if the user is the newly registered user, a palm vein image of the newly registered user is collected, the deep learning palm vein identification model is updated by using the palm vein image of the newly registered user, and an updated palm vein identification model is obtained;
and the client is used for providing a login interface for the user to inquire the corresponding identification information.
2. The system of claim 1, wherein the cloud storage module is further to:
and processing the request of the client, and making a corresponding response by judging different requests.
3. A palm vein recognition method based on the recognition system according to any one of claims 1 to 2, comprising:
step S1: carrying a deep learning palm vein recognition model through a cloud storage module, training a deep convolution neural network model to obtain the deep learning palm vein recognition model, and storing user information;
step S2: collecting a palm vein image of a user through a collection module, calling a deep learning palm vein recognition model through an edge calculation module to verify the collected palm vein image, wherein verification information is uploaded to a cloud storage module when verification is successful, receiving information input by the user through a receiving module when verification is failed, matching the information input by the user with user information stored in the cloud storage module, judging whether the user is a newly registered user who does not collect verification palm veins, if the user is the newly registered user, collecting the palm vein image of the newly registered user, updating the deep learning palm vein recognition model by using the palm vein image of the newly registered user, and obtaining an updated palm vein recognition model;
step S3: and providing a login interface through the client for the user to inquire corresponding identification information.
4. The method of claim 3, wherein the method further comprises:
identifying the palm vein image of the current new registered user based on the updated palm vein identification model, and returning the palm vein image information of the current new registered user to the cloud storage module;
matching the palm vein image information of the current newly registered user with the registration information of the user, and if the matching is successful, returning to the palm vein acquisition and identification equipment to prompt that the input is successful; and if the matching fails, returning to the palm vein acquisition and identification equipment, prompting the input failure, and displaying the instruction information of re-acquiring the palm vein image.
5. The method as claimed in claim 3, wherein in step S1, the procedure of obtaining the deep learning palm vein recognition model includes:
step S1.1: inputting the obtained palm vein data set into a cloud storage module, wherein the palm vein data set is divided into a training set, a testing set and a verification set;
step S1.2: in the model training stage, inputting a training set and labels corresponding to images in the training set into a deep convolutional neural network model, and outputting results after training is finished;
step S1.3: in the model test stage, inputting a test set of the palm veins into a deep convolutional neural network model to test the recognition rate of the model;
step S1.4: in the model verification stage, the recognition condition of the model is verified in the deep convolutional neural network model, and the classifier parameters and the complexity of the model are adjusted according to the recognition condition to obtain a deep learning palm vein recognition model with the classifier parameters and the complexity of the model meeting the conditions.
6. The method according to claim 3, wherein the step S2 of the edge calculation module invoking the deep learning palm vein recognition model to verify the collected palm vein image specifically comprises:
step S2.1: preprocessing the collected palm vein image of the user;
step S2.2: calling a deep learning palm vein recognition model, and performing similarity calculation on the preprocessed palm vein image;
step S2.3: and according to the similarity calculation result, verifying whether the palm vein similarity reaches a set threshold value through a decision-making device, and verifying the collected palm vein image of the user.
7. The method according to claim 5, characterized in that step S1.2 comprises in particular:
step S1.2.1: initializing variable training times of the deep convolutional neural network model;
step S1.2.2: selecting a part of palm vein training set for batch processing;
step S1.2.3: when the model is trained, the whole model is subjected to forward propagation to obtain a predicted value;
step S1.2.4: after the forward propagation is finished, the model performs backward propagation and updates the variable in the training process;
step S1.2.5: detecting whether a training target is reached, and if the training target is reached, finishing the training; if not, go to step S1.2.6;
step S1.2.6: testing whether the training times are reached, and if the training times are reached, finishing the training; if the training times are not reached, go to step S1.2.7;
step S1.2.7: and (4) increasing the training times, and then repeating the steps S1.2.1-S1.2.6 again until the training requirements are met, and finishing the training.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910609754.XA CN110414373B (en) | 2019-07-08 | 2019-07-08 | Deep learning palm vein recognition system and method based on cloud edge-side cooperative computing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910609754.XA CN110414373B (en) | 2019-07-08 | 2019-07-08 | Deep learning palm vein recognition system and method based on cloud edge-side cooperative computing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110414373A CN110414373A (en) | 2019-11-05 |
| CN110414373B true CN110414373B (en) | 2021-09-24 |
Family
ID=68360584
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910609754.XA Active CN110414373B (en) | 2019-07-08 | 2019-07-08 | Deep learning palm vein recognition system and method based on cloud edge-side cooperative computing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110414373B (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110719210B (en) * | 2019-12-05 | 2020-08-28 | 赣江新区智慧物联研究院有限公司 | Industrial equipment predictive maintenance method based on cloud edge cooperation |
| CN113128529A (en) * | 2019-12-30 | 2021-07-16 | 中移(成都)信息通信科技有限公司 | Livestock identity recognition model training method, recognition method, system and device |
| CN111274924A (en) * | 2020-01-17 | 2020-06-12 | 厦门中控智慧信息技术有限公司 | Palm vein detection model modeling method, palm vein detection method and palm vein detection device |
| CN111400040B (en) * | 2020-03-12 | 2023-06-20 | 重庆大学 | Industrial Internet system based on deep learning and edge calculation and working method |
| CN111461694B (en) * | 2020-03-17 | 2023-07-18 | 上海大学 | Dish identification and pricing system and method based on multi-level learning model |
| CN111368529B (en) * | 2020-03-17 | 2022-07-01 | 重庆邮电大学 | Mobile terminal sensitive word recognition method, device and system based on edge calculation |
| CN111582016A (en) * | 2020-03-18 | 2020-08-25 | 宁波送变电建设有限公司永耀科技分公司 | Intelligent maintenance-free power grid monitoring method and system based on cloud edge collaborative deep learning |
| CN111985650B (en) * | 2020-07-10 | 2022-06-28 | 华中科技大学 | Activity recognition model and system considering both universality and individuation |
| CN112732591B (en) * | 2021-01-15 | 2023-04-07 | 杭州中科先进技术研究院有限公司 | Edge computing framework for cache deep learning |
| CN113095506A (en) * | 2021-03-25 | 2021-07-09 | 北京大学 | Machine learning method, system and medium based on end, edge and cloud cooperation |
| CN113298055A (en) * | 2021-07-27 | 2021-08-24 | 深兰盛视科技（苏州）有限公司 | Vein identification method, vein identification device, vein identification equipment and computer readable storage medium |
| CN113965632A (en) * | 2021-11-03 | 2022-01-21 | 武汉大学 | ECG analysis system and method based on cloud-equipment terminal cooperation |
| CN114363984B (en) * | 2021-12-16 | 2022-11-25 | 苏州大学 | Cloud edge collaborative optical carrier network spectrum resource allocation method and system |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107437074A (en) * | 2017-07-27 | 2017-12-05 | 深圳市斑点猫信息技术有限公司 | A kind of identity identifying method and device |
| CN107832684A (en) * | 2017-10-26 | 2018-03-23 | 通华科技（大连）有限公司 | A kind of intelligent vein authentication method and system with independent learning ability |
| CN108446680A (en) * | 2018-05-07 | 2018-08-24 | 西安电子科技大学 | A kind of method for secret protection in face authentication system based on edge calculations |
| CN108615002A (en) * | 2018-04-22 | 2018-10-02 | 广州麦仑信息科技有限公司 | A kind of palm vein authentication method based on convolutional neural networks |
| CN109034034A (en) * | 2018-07-12 | 2018-12-18 | 广州麦仑信息科技有限公司 | A kind of vein identification method based on nitrification enhancement optimization convolutional neural networks |
| CN109492536A (en) * | 2018-10-12 | 2019-03-19 | 大唐高鸿信息通信研究院（义乌）有限公司 | A kind of face identification method and system based on 5G framework |
| CN109558827A (en) * | 2018-11-26 | 2019-04-02 | 济南浪潮高新科技投资发展有限公司 | A kind of finger vein identification method and system based on personalized convolutional neural networks |
-
2019
- 2019-07-08 CN CN201910609754.XA patent/CN110414373B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107437074A (en) * | 2017-07-27 | 2017-12-05 | 深圳市斑点猫信息技术有限公司 | A kind of identity identifying method and device |
| CN107832684A (en) * | 2017-10-26 | 2018-03-23 | 通华科技（大连）有限公司 | A kind of intelligent vein authentication method and system with independent learning ability |
| CN108615002A (en) * | 2018-04-22 | 2018-10-02 | 广州麦仑信息科技有限公司 | A kind of palm vein authentication method based on convolutional neural networks |
| CN108446680A (en) * | 2018-05-07 | 2018-08-24 | 西安电子科技大学 | A kind of method for secret protection in face authentication system based on edge calculations |
| CN109034034A (en) * | 2018-07-12 | 2018-12-18 | 广州麦仑信息科技有限公司 | A kind of vein identification method based on nitrification enhancement optimization convolutional neural networks |
| CN109492536A (en) * | 2018-10-12 | 2019-03-19 | 大唐高鸿信息通信研究院（义乌）有限公司 | A kind of face identification method and system based on 5G framework |
| CN109558827A (en) * | 2018-11-26 | 2019-04-02 | 济南浪潮高新科技投资发展有限公司 | A kind of finger vein identification method and system based on personalized convolutional neural networks |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110414373A (en) | 2019-11-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110414373B (en) | Deep learning palm vein recognition system and method based on cloud edge-side cooperative computing | |
| CN108491805B (en) | Identity authentication method and device | |
| KR101629224B1 (en) | Authentication method, device and system based on biological characteristics | |
| US6810480B1 (en) | Verification of identity and continued presence of computer users | |
| CN111652087B (en) | Car inspection method, device, electronic equipment and storage medium | |
| US10949517B2 (en) | Identification system enrollment and validation and/or authentication | |
| CN109522829B (en) | Smart phone face-brushing conference registration method based on deep learning | |
| CN112801054B (en) | Face recognition model processing method, face recognition method and device | |
| US20230359721A1 (en) | Methods and systems for facilitating secure authentication of user based on known data | |
| CN111695502A (en) | Feature updating method and device for face recognition and computer equipment | |
| CN110298240A (en) | A kind of user vehicle recognition methods, device, system and storage medium | |
| CN114387548A (en) | Video and liveness detection method, system, device, storage medium and program product | |
| CN113591603A (en) | Certificate verification method and device, electronic equipment and storage medium | |
| CN111062345A (en) | Training method and device of vein recognition model and vein image recognition device | |
| CN116312513A (en) | Intelligent voice control system | |
| CN114373213A (en) | Juvenile identity recognition method and device based on face recognition | |
| CN111209551B (en) | Identity authentication method and device | |
| CN115082873A (en) | Image recognition method and device based on path fusion and storage medium | |
| CN111753804A (en) | Recognition device, acquisition device, fee deduction device and non-contact biological recognition system for public transport | |
| CN112597468A (en) | Data interface management method and system based on stroke order, OCR (optical character recognition) and artificial intelligence | |
| CN114495188A (en) | Image data processing method and device and related equipment | |
| CN111368866A (en) | Picture classification method, device and system | |
| CN116612506A (en) | Training method of face recognition model, face verification method and device | |
| CN113988314B (en) | Clustering federation learning method and system for selecting clients | |
| CN116612358B (en) | Data processing method, related device, equipment 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |