CN111291628B - Face data distributed identification and storage architecture based on block chain technology - Google Patents
Face data distributed identification and storage architecture based on block chain technology Download PDFInfo
- Publication number
- CN111291628B CN111291628B CN202010049817.3A CN202010049817A CN111291628B CN 111291628 B CN111291628 B CN 111291628B CN 202010049817 A CN202010049817 A CN 202010049817A CN 111291628 B CN111291628 B CN 111291628B
- Authority
- CN
- China
- Prior art keywords
- node
- picture
- face
- operation node
- nodes
- 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
- 238000005516 engineering process Methods 0.000 title claims abstract description 31
- 238000013475 authorization Methods 0.000 claims description 17
- 238000012795 verification Methods 0.000 claims description 11
- 238000009434 installation Methods 0.000 claims description 8
- 239000000284 extract Substances 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 abstract description 4
- 238000013500 data storage Methods 0.000 abstract description 3
- 238000010586 diagram Methods 0.000 description 19
- 238000000034 method Methods 0.000 description 17
- 239000000306 component Substances 0.000 description 16
- 230000006870 function Effects 0.000 description 11
- 210000003128 head Anatomy 0.000 description 10
- 238000012544 monitoring process Methods 0.000 description 8
- 238000012545 processing Methods 0.000 description 8
- 238000000605 extraction Methods 0.000 description 6
- 230000007246 mechanism Effects 0.000 description 6
- 238000004891 communication Methods 0.000 description 5
- 238000007726 management method Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 230000003993 interaction Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000012423 maintenance Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 239000008358 core component Substances 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 210000004709 eyebrow Anatomy 0.000 description 1
- 230000001815 facial effect Effects 0.000 description 1
- 238000005242 forging Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012856 packing Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- 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
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L63/00—Network architectures or network communication protocols for network security
- H04L63/04—Network architectures or network communication protocols for network security for providing a confidential data exchange among entities communicating through data packet networks
- H04L63/0428—Network architectures or network communication protocols for network security for providing a confidential data exchange among entities communicating through data packet networks wherein the data content is protected, e.g. by encrypting or encapsulating the payload
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/104—Peer-to-peer [P2P] networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1095—Replication or mirroring of data, e.g. scheduling or transport for data synchronisation between network nodes
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1097—Protocols in which an application is distributed across nodes in the network for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/34—Network arrangements or protocols for supporting network services or applications involving the movement of software or configuration parameters
Landscapes
- Engineering & Computer Science (AREA)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Theoretical Computer Science (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Computer Security & Cryptography (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Hardware Design (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computing Systems (AREA)
- Image Processing (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides a face data distributed recognition and storage architecture based on a block chain technology, which comprises a system operation node, an operation node, a camera agent node and a camera group, wherein the block chain technology is adopted to integrate idle and scattered strange calculation resources together, the face recognition task is completed jointly in a mutually untrusted network formed by the nodes and the groups, meanwhile, the accuracy and the reliability of a recognition result are effectively ensured, and a set of face recognition and data storage platform with strong practicability, low cost, high safety, good expansibility and certain advancement is constructed.
Description
Technical Field
The invention relates to the technical field of blockchain technology and biological feature recognition, in particular to a framework for realizing distributed recognition and storage of face data by using the blockchain technology.
Background
The blockchain technology is a novel application mode integrating computer technologies such as data distributed storage, point-to-point transmission network, consensus mechanism, encryption algorithm and the like, and plays an important role in data privacy, data security, trust construction and the like.
The face recognition technology is mainly used for analyzing and recognizing the identity of the person. The human face is an inherent physiological characteristic of human body and has the characteristics of uniqueness, incapability of forging, portability, availability at any time and any place, and the like. Face recognition, fingerprint recognition, voiceprint recognition, iris recognition and the like belong to biological feature recognition technologies.
Along with the rapid popularization of video monitoring nowadays, in dealing with increasingly complex and severe public security situations, rapid identification of a photographed person under the condition that the photographed person is not matched with the video monitoring is required to be performed, so that intelligent early warning is realized. Face recognition technology is certainly the best choice over other biometric recognition technologies. The method can accurately find the human face from the monitoring video image in real time, and then compare the human face with the human face data in the human face database in real time, thereby rapidly confirming the identity of the photographed person.
The face recognition technology can be widely applied to the security field, is also an important component of a smart city, is a core support platform of a public service and emergency smart platform for city management, and can track each monitoring site in real time in time by a strict video monitoring network in a region-crossing space-time manner, thereby greatly improving the quick response capability, effectively striking crimes, improving the security level and the comprehensive management level of the city and being beneficial to the rapid development of various industries and economies.
Through practical inspection in recent years, the key affecting the popularization and application of the face recognition technology is not the advantages and disadvantages of the face recognition algorithm, is not the loss of massive training sample data, and is the computational problem which is regarded as solved subjectively. The face recognition algorithm has high requirements on computational power resources, and although the performance of a processing chip is continuously improved, the power consumption and the cost are still bottlenecks faced by the development of chip technology. With the rapid increase of the data volume of future video images, the application of the face recognition technology is restricted by computational effort more and more obviously. The huge technology can realize the improvement of calculation power by continuously accumulating hardware resources, but small and medium enterprises are difficult to bear high cost expenditure, so that the method is not beneficial to the commercial civilian popularization of the face recognition technology.
Disclosure of Invention
The invention aims to provide a face data distributed recognition and storage architecture based on a block chain technology, and the core idea is to integrate idle and dispersed unfamiliar computational resources (i.e. computational resources not controlled by a platform) by using the block chain technology, finish face recognition tasks together in a mutually-untrusted network, effectively ensure the accuracy and reliability of recognition results, and construct a set of face recognition and data storage platform with strong practicability, low cost, high safety, good expansibility and certain advancement. The platform can help small and medium enterprises not to invest excessive hardware overhead for building a high-power system, so that the operation cost of the enterprises is effectively reduced, and meanwhile, the rapid popularization of the face recognition technology in the whole society is facilitated.
In order to achieve the above objective, the present invention provides a face data distributed recognition and storage architecture based on a blockchain technology, which includes a system operation node, an operation node, a camera agent node, and a camera group.
The nodes formed by the system operation nodes, the operation nodes and the camera proxy nodes form a distributed P2P network by operating a P2P network protocol, so that point-to-point interconnection and intercommunication among the nodes are realized. The system operation node and the camera agent node are managed and deployed by a system operator, the operation node can be any group or individual willing to provide own computing power resources except the operator, and the operation node can be added into or withdrawn from the P2P network at will as long as the operation node and the camera agent node run the P2P network protocol on respective operation equipment; the operational node may also be an operator controlled computational resource. The computing device may be a notebook or desktop computer, or may be a server with superior performance.
The system operation node can be regarded as the unique representative of a system operator, and is mainly responsible for the management of a face recognition component and face data, the maintenance of P2P network communication and data security and the issuing of economic rewards, and comprises a face recognition module, a face database module, an encryption/decryption module, a service module, a network routing module, a rewarding module, a picture library, a picture chain and a registry. The face recognition module comprises a face data processing component and realizes two functions of face feature extraction and face feature comparison; the face database module is responsible for storing face data, the face data belongs to highly private information, and ownership is in a system operation node; the encryption/decryption module is responsible for encrypting and decrypting the data transmitted in the P2P network, for example, the face data is transmitted to each operation node through the P2P network after being encrypted, the identification result calculated by the operation node also needs to be encrypted and transmitted back to the system operation node, and the encrypted identification result needs to be decrypted after reaching the system operation node to obtain a final result; the service module is responsible for managing the face characteristic data and the face data processing component, monitoring the running state of the whole network and giving out rewards; the network routing module is responsible for nodes to join or leave the P2P network, and maintains data communication channels and connections with other nodes; the rewarding module mainly distributes economic rewards to the operation nodes providing computing power to complete the identification task, the operation nodes obtaining rewards are generated by voting of other operation nodes, and the rewarding module directly distributes the rewards to accounts of winning nodes; the picture library is used for storing unrecognized face pictures transmitted in the P2P network, and the face pictures are deleted from the picture library once the face pictures are recorded in the picture chain; the picture chain is used for storing the picture blocks which are identified and verified, and the picture blocks are hung behind the preamble blocks according to the contained preamble block hash values; the registry is used for recording the information of the operation node and the camera proxy node of the network.
The operation node comprises a network routing module, an encryption/decryption module, a face recognition module, a face database module, a picture library and a picture chain. The method adopts a workload proof mechanism to compete for the face picture identification right, an operation node obtaining the identification right needs to broadcast a picture block to a P2P network, a system operation node in the network and other operation nodes are used for verifying whether the picture block meets workload proof, a vote is accepted once the workload proof is met, and the system operation node gives winning operation node identification authorization according to a voting result; with identification authorization, the winning operation node is responsible for completing the identification task of the face picture contained in the picture block, wherein the task comprises face feature extraction and face feature comparison; the operation node without winning starts to pack new picture blocks and starts a new round of recognition right competition; after the identification task is completed, the winning operation node sends the identification result to the system operation node and other operation nodes through the P2P network, the nodes verify and vote the received identification result, the voting result is informed to the system operation node, the system operation node judges whether to finally give rewards to the winning operation node according to the voting result, and meanwhile, whether to give rewards to each operation node is informed. Each operation node locally maintains a picture chain, stores the picture block hanging chain which is demonstrated by the operation node which obtains rewards, and the system operation node also hangs the picture block on the picture chain. The image blocks which are demonstrated by the operation nodes which do not obtain rewards are discarded by each node, and the face images contained in the image blocks are packaged into new image blocks again to start a new round of competition. Each operation node is independent and peer-to-peer, and mainly works to perform face recognition operation, so that all operation nodes can be regarded as forming an operation resource network.
The camera proxy node contains a network routing module and an encryption/decryption module, which is managed and deployed by the system operator. The method has the main functions of receiving the face pictures transmitted by the camera group, encrypting the received face pictures and transmitting the encrypted face pictures to the P2P network.
The camera group is composed of cameras deployed at different places, and the purpose is to send face pictures collected on site to corresponding camera proxy nodes, and the face pictures are sent to an operation resource network after being processed by the proxy nodes, so that the face pictures finally reach a system operation node. The camera group is managed and deployed by the system operator.
Compared with the prior art, the technical scheme provided by the invention has the following advantages:
1. the practicability is strong: the technical scheme of the invention is well suitable for the face recognition technology and can meet the actual working requirements to the maximum extent. The node type and the function design fully consider the high efficiency and the stability of the execution of the face recognition task; the distributed P2P network ensures arbitrary joining and exiting of nodes.
2. The cost is low: for the recognition work of massive face pictures, a plurality of servers with excellent performance are required to be deployed, so that the investment and maintenance cost of operators are increased intangibly. According to the technical scheme, idle calculation forces except the operator can be attracted to be input into the identification task, so that the input cost of the operator is reduced; meanwhile, the rapid popularization of the face recognition technology in the whole society is promoted.
3. The safety is high: the face database module, the rewarding module and the transmission of data in the network channel are protected by adopting a safe encryption mechanism, so that the data is effectively prevented from being stolen and tampered. In addition, the face data is stored in other nodes in a scattered manner besides being stored locally in the system operator, and once the face data of one node is lost or tampered, the data can be synchronized from the other nodes.
4. The expansibility is good: after the operator deploys the system operation node and the camera agent node, the operator attracts the idle computing power to join the face recognition work through the rewarding mechanism, the operator decides the rewarding amount according to the actual work amount of the face recognition, and the higher the rewarding amount, the more the attractive idle computing power, and otherwise, the less. The size of the operation resource network is also in a dynamic change state.
5. Maintainability is high: the system architecture fully considers that the system architecture has good flexibility and expandability in the aspects of face data processing capability, network communication capability, local data storage capacity, data security, product upgrading and the like, and particularly an operator is not responsible for purchasing and maintaining face recognition hardware facilities any more, but the purchasing and maintaining work is transmitted to an operation node controlled by a non-operator; the product upgrade is also limited to the upgrade of software to the greatest extent, and a software upgrade message is sent to the operation node through the network, and the operation node decides whether to upgrade or not. The maintainability and portability are fully considered by the structure and the program module adopted by the system software, namely, the aim of modifying a certain component, upgrading a new function and reorganizing the structure of the system according to the requirement is fulfilled.
5. The method has certain advancement: at present, the blockchain technology is mainly applied to the field of financial science and technology, and the invention expands the blockchain technology to the field of security, and particularly provides a brand new thought and scheme for the application innovation of the face recognition technology.
Drawings
The drawings are not intended to be drawn to scale. In the drawings, each identical or nearly identical component that is illustrated in various figures may be represented by a like numeral. For purposes of clarity, not every component may be labeled in every drawing. Embodiments of various aspects of the invention will now be described, by way of example, with reference to the accompanying drawings, in which:
FIG. 1 is a schematic diagram of a face data distributed identification and storage architecture based on blockchain techniques in accordance with some embodiments of the present invention;
FIG. 2 is a schematic diagram of a distributed P2P network of system operation nodes, and camera proxy nodes;
FIG. 3 is a schematic diagram of the operation of the photo library and face pictures thereof;
FIG. 4 is a block diagram of a picture chain;
FIG. 5 is a schematic diagram of the internal modules and data flow of the system operation node;
FIG. 6 is a schematic diagram of a node voting method;
FIG. 7 is a schematic diagram of a startup procedure of a system operation node;
FIG. 8 is a schematic diagram of face recognition software sent to an operational node;
FIG. 9 is a schematic diagram of bonus data being sent to winning compute nodes;
FIG. 10 is a schematic diagram of a system operation node receiving a face picture;
FIG. 11 is a diagram illustrating a system operation node receiving a picture block and a recognition result;
FIG. 12 is a schematic diagram of a system operation node receiving voting results;
FIG. 13 is a schematic diagram of a system operator node receiving node status information;
FIG. 14 is a schematic diagram of internal modules of an operation node;
FIG. 15 is a schematic diagram of an operational node startup procedure;
FIG. 16 is a schematic diagram of a competition image identification process of an operation node;
FIG. 17 is a schematic diagram of a winning compute node winning a prize flow;
FIG. 18 is a schematic diagram of the internal modules of a camera proxy node;
FIG. 19 is a schematic diagram of a relationship between a camera group and a camera proxy node.
Detailed Description
For a better understanding of the technical content of the present invention, specific examples are set forth below, along with the accompanying drawings.
Aspects of the invention are described in this disclosure with reference to the drawings, in which are shown a number of illustrative embodiments. The embodiments of the present disclosure are not necessarily intended to include all aspects of the invention. It should be understood that the various concepts and embodiments described above, as well as those described in more detail below, may be implemented in any of a number of ways, as the disclosed concepts and embodiments are not limited to any implementation. Additionally, some aspects of the disclosure may be used alone or in any suitable combination with other aspects of the disclosure.
Referring to fig. 1, a face data distributed recognition and storage architecture based on a blockchain technology according to an embodiment of the present invention includes a system operation node, an operation node, a camera proxy node, and a camera group formed by a plurality of cameras.
With reference to fig. 2, the system operation node, the operation node and the camera proxy node are interconnected by an operation P2P network protocol to form a distributed P2P network. In this distributed P2P network, the functions of the individual nodes are as follows:
(1) The system operation node is mainly responsible for the management of a face recognition component and face data and economic rewards, and comprises a face recognition module, a face database module, an encryption/decryption module, a service module, a network routing module, a rewards module, a picture library, a picture chain and a registry. The face recognition module comprises a face data processing component and realizes two functions of face feature extraction and face feature comparison; the face database module is responsible for storing face feature data, wherein the face feature data belongs to highly private information, and ownership is in a system operation node; the encryption/decryption module is responsible for encrypting and decrypting the data transmitted in the P2P network, for example, the face data is transmitted to each operation node through the P2P network after being encrypted, the identification result calculated by the operation node also needs to be encrypted and transmitted back to the system operation node, and the encrypted identification result needs to be decrypted after reaching the system operation node to obtain a final result; the service module is responsible for managing the face data and the face data processing component, monitoring the running state of the whole network and giving out rewards; the network routing module is responsible for nodes to join or leave the P2P network, and maintains data communication channels and connections with other nodes; the rewarding module mainly distributes economic rewards to the operation nodes providing computing power to complete the identification task, the operation nodes obtaining rewards are generated by voting by the system operation nodes and other operation nodes, and the rewarding module directly distributes the rewards to accounts of winning nodes; the picture library is used for storing unrecognized face pictures transmitted in the P2P network, and the face pictures are deleted from the picture library once the face pictures are recorded in the picture chain; the picture chain is used for storing the picture blocks which are identified and verified, and the picture blocks are hung behind the preamble blocks according to the contained preamble block hash values; the registry is used for recording the information of the operation node and the camera proxy node of the network.
(2) The operation node comprises a network routing module, an encryption/decryption module, a face recognition module, a face database module, a picture library and a picture chain. The method comprises the steps that a workload proving mechanism is adopted to compete for face picture recognition rights, an operation node obtaining the recognition rights sends picture blocks to a system operation node and other operation nodes, the nodes verify and vote, and the system operation node collects the votes and then judges whether to authorize recognition for winning operation nodes; when the winning node receives authorized identification, the winning node is fully responsible for completing the face identification task, wherein the face identification task comprises face feature extraction and face feature comparison, and other operation nodes start a new round of identification right competition; the winning operation node sends the identification result to the system operation node and other operation nodes through the P2P network, the system operation node verifies and votes the identification result, the system operation node collects the votes and then judges whether to award the winning operation node, and once awarded, the system operation node and other operation nodes hang the picture block to a local picture chain. The picture blocks which are demonstrated by the winning operation nodes which do not obtain rewards are discarded by each node, and the face pictures contained in the picture blocks are packaged into new picture blocks again to start a new round of competition. Each operation node is independent and peer-to-peer, and mainly works to perform face recognition operation, so that all operation nodes form an operation resource network.
(3) The camera proxy node contains a network routing module and an encryption/decryption module, which is managed and deployed by the system operator. The function of the method is to receive the face picture transmitted by the camera group, and transmit the received face picture to the operation resource network after encryption.
In addition, the camera group is composed of cameras deployed at different places, and the purpose is to send face pictures collected on site to corresponding camera proxy nodes, and the face pictures are sent to the computing resource network after being processed by the proxy nodes.
With reference to fig. 3, other nodes except for the proxy node of the camera in the whole system locally maintain a picture library, which is a place for temporarily storing face pictures. And once the nodes receive the face pictures, the pictures are put in storage, and the face pictures in the storage are related together according to the sequence of the time stamps. Before competing for the picture block identification right each time, the operation node extracts a plurality of face pictures from the picture library to be packed into a picture block, and then starts competing for the identification right. All face pictures included in the picture chain are deleted from the picture library of each node.
As shown in connection with fig. 4, each node in the overall system maintains a chain of pictures locally. The picture chain is that each picture block is associated with the front block and the rear block according to the head hash value of the front block according to the time stamp sequence, so that the distributed storage of the face pictures captured by the camera at each node is realized. The picture block is composed of a picture block head and a picture block body, wherein the picture block body comprises picture data and a merkel tree thereof. The data structure of the picture block header mainly comprises: software version number, preamble block hash value, timestamp, difficulty coefficient, random number and picture merkel tree root; the preamble block hash value is a hash value obtained by carrying out SHA256 hash calculation on the preamble block head data, and the association of the preamble block and the rear block is realized by utilizing the unique unchanged characteristic of the hash value; the time stamp represents the time of generating the picture block head, so that the picture block can be traced back to the picture chain according to the time sequence; the difficulty coefficient is used for generating difficulty of the picture block, and the operation node is ensured to find a random number meeting the requirement of the difficulty coefficient in a certain period of time; the random number represents a changeable numerical value, and the operation node calculates a hash value of the block header by changing the value, so that the hash value meets the requirement of a difficulty coefficient; the picture merkel tree root represents that all face picture data contained in the picture block are related together through a merkel tree structure, and finally a hash value is calculated and obtained, wherein the hash value is the picture merkel tree root. Hash computation in the mekel tree still uses SHA256 hash functions. The photo chain can only be accessed by the system operation node or by a third party authorized by it.
The following describes the functions of each node and its internal modules, and the operation flow of the system in detail with reference to the following figures.
1. With reference to fig. 5, a network routing module of the system operation node provides a routing function, ensures that the node joins or exits the P2P network, and simultaneously receives and propagates face pictures, picture blocks, registration information, voting results, rewards and other data, and face recognition software (including a face recognition module and a face database module for the operation node to download and install), and discovers and maintains connection with other nodes.
All data received from the network or transmitted to the network through the network routing module need to be subjected to corresponding decryption processing through the encryption/decryption module, and all data are not opened to non-operators.
The picture library is a memory area for temporarily storing unidentified face pictures received from a network, and once a picture block reaches a node, the identified face pictures contained in the picture library are deleted from the picture library after block verification and identification result verification are performed successively; the picture chain stores the identified picture blocks, and the new picture blocks find out the associated preamble blocks according to the hash values of the preamble blocks and are hung behind the associated preamble blocks, so that the chain storage is realized.
The registry stores online operation node information and camera agent node information in the whole P2P network, nodes are added into the registry when the nodes are added, and the corresponding information is deleted from the registry when the nodes are withdrawn. The information in the registry is dynamically changed, the registry is stored in the hard disk of the computer, and once the system operation node fails and is restarted, the online node information can be directly obtained from the registry stored in the hard disk, and communication can be established again with the online node information.
The face recognition module comprises a face feature extraction component and a face feature comparison component. Face feature extraction refers to extracting feature data of key areas of a face (such as eyebrows, eyes, nose, mouth, chin, facial contours and the like), and the feature data reflects local relations of various parts in the key areas and interrelations between the parts, which are also called face feature data. The face feature comparison component searches and matches the feature data of the extracted face image with the feature data stored in the database, and when the similarity reaches or exceeds a comparison threshold value, the face matching is considered to be successful. The face recognition module is one of the core functions of the whole system, and is provided for the operation node to download and use by the system operation node.
Face feature data of people with important attention is stored in the face database module, and the data belongs to high confidentiality. The module and the face recognition module are matched to complete the recognition work of the face pictures acquired by the camera. The module is downloaded together with the face recognition module through the P2P network for the operation node to use.
The rewarding module confirms the legitimacy of the operation node obtaining the picture identification right, and then issues rewards to the winning operation node through the P2P network. The method for confirming the validity adopts a node voting method, and comprises the following steps: nodes with a face recognition module and a face database module participate in voting, winning operation nodes broadcast recognition results in the whole network, and other nodes carry out validity verification by the face recognition module after receiving the recognition results. Because the winning node is going to 1: n is compared, namely one with highest similarity is identified from N face database pictures, comparison is needed with the N pictures, so that time and labor are consumed, and verification of other nodes only needs to carry out 1 on the snap face picture and the identified library picture: 1, so that the verification can be completed quickly, the verification is passed to vote the operation node for checking the picture block, the verification is not passed to vote the operation node, each node with voting right sends the voting result to the system operation node, the system operation node maintains the on-line operation node information in the whole system, so that the reward module can count the percentage of the number of votes to vote. Of course, the system operation node also verifies the identification result and counts as a ticket, if the ticket is approved to be more than ninety five percent, the winning node is considered to be indeed obtained the identification right and the identification result is correct, and the rewarding module issues rewards to the winning node directly through the P2P network. The node voting method is shown in fig. 6.
The service module connects the inside and outside of the system. The user realizes the operation control of the system through the Web page, including the upgrade of each module in the system operation node, the update of the face characteristic data in the face database, the monitoring of the operation state of the whole system, and the like, all of which realize the normal operation through the service module.
The start-up flow of the system operation node is shown in fig. 7. The face recognition module and the face database module are used as a static data component for downloading and installing by the operation node; the picture library is stored in the memory, and the picture chain is stored in the local hard disk. The system operation node starts a service component firstly, because the service component controls the information interaction between the Web interface and other important modules in the node; and then starting the rewarding module and the encryption/decryption module in sequence, and then starting the network routing module, so that the system operation node accesses the P2P network, establishes connection with other nodes and maintains the connection, and waits for data interaction in the P2P network.
The data interaction between the system operation node and other nodes mainly comprises the following two aspects:
(1) From the system operation node to the P2P network side, face recognition software is loaded, wherein the face recognition software comprises a face recognition module and a face database module. When the operation node joins the P2P network for the first time or needs to upgrade the face recognition software, the operation node needs to load the latest face recognition software from the system operation node through the P2P network. The face recognition software is the software formed by packing the face recognition module and the face database module together by the system operation node, and needs to be encrypted at the operation node before being sent to the operation node, so that the face recognition software can prevent the face recognition module from being stolen or interfered in network transmission on one hand and can shield the operation node from the detailed information in the core component on the other hand, and the flow is shown in fig. 8. Secondly, the rewarding data is sent to the operation node obtaining rewards, the rewarding data belongs to core confidential data, and is also encrypted before being sent to the P2P network, the flow is shown in fig. 9, and other modules such as module upgrading information, identification authorization notification and rewarding notification are sent to relevant nodes in the P2P network through similar flows.
(2) The six types of data, namely face pictures, picture blocks, identification results, voting information after the operation nodes verify the picture blocks and the identification results and online information of the nodes, are transmitted to the system operation nodes from the P2P network. The face picture is generated by the camera, encrypted by the camera proxy node and sent to the P2P network, the system operation node decrypts the face picture after receiving the face picture, and then the face picture is put into a local picture library, and the flow is shown in the figure 10. After the picture block and the identification result are transmitted to the system operation node from the network, firstly decrypting the picture block, firstly judging whether the picture block meets the work load evidence for the picture block, after the picture block passes the verification, approving the vote for the winning operation node for proving the picture block, then the system operation node waits for the voting result of other operation nodes for the picture block, judging whether the winning operation node is given identification authorization according to the results, and informing all operation nodes of the identification authorization notification; and for the identification result, the system operation node sends the face picture corresponding to the identification result to the face recognition module and performs identification work in combination with the face database module, judges whether the identification result is accurate, votes and sends the voting result to the rewarding module, meanwhile, sends the received voting information of other operation nodes to the rewarding module, determines whether to give economic rewards to winning operation nodes by the rewarding module, and sends the rewarding information to all operation nodes. The picture block for the winning operation node is hung on the local picture chain, and the face picture contained in the picture block is deleted from the local picture library, and the flow is shown in fig. 11. In short, after verifying and voting the picture block and the identification result, the nodes with the voting rights send the voting result to the system operation node after encrypting, and after decrypting, the system operation node judges whether to give identification authorization to the picture block voting result; for the voting result of the recognition result, the voting result is sent to the rewarding module, and the rewarding module counts the number of votes and finally decides whether to issue rewards to the winning operation node, and the above flow is shown in fig. 12. Other nodes except the system operation node can send registration information to the system operation node when joining the network, unbind registration information is sent when exiting the network, and node keep-alive information needs to be sent in a fixed time in the network all the time, so that other nodes know the connection condition of the nodes, and the data flow is shown in figure 13.
2. The complete operation node comprises a network routing module, an encryption/decryption module, a face recognition module and a face database module, and is used for storing a picture library and a picture chain respectively by opening up a memory space and a hard disk space locally as shown in fig. 14.
The workflow of the operation node is shown in fig. 15, a user willing to provide his own computing power to participate in face recognition work to obtain rewards for the first time downloads an operation node installation package from an operator through the internet, the installation package comprises a network routing module and an encryption/decryption module, the operation node is started after the user installation is completed, the network routing module of the operation node helps the operation node to access a P2P network, then the operation node accesses the system operation node through the P2P network according to the routing information of the system operation node solidified in the network routing module, the system operation node is registered, and after the registration is successful, the operation node starts to download face recognition software formed by the face recognition module and the face database module and loads the face recognition software into the operation node, and then starts to open up space for a picture library and a picture chain. So far, the installation of the whole operation node is finished, and the competition picture identification right can be started. When the operation node exits the P2P network, the operation node needs to unbind the registration information for the system operation node, and the operation node information is deleted from the registry after the system operation node receives the unbind registration information, so that the system is convenient to maintain the online operation node information of the whole network. When the operation node which is not operated for the first time is started again, the operation node only needs to join the network through the network routing module, then registers with the operation node of the system, and then starts competing for the picture identification right. If the face recognition software upgrading information exists, whether to upgrade is determined by self.
The operator can deploy a plurality of operation nodes controlled by the operator, so that a network environment with stable operation can be constructed, and the problem that face recognition work is stagnated due to the fact that all operation nodes of non-operators exit the network is avoided.
The method for competing for picture identification rights adopts a workload proving mechanism in a block chain technology. The operation node firstly extracts a plurality of face pictures from a local picture library, then forms a picture block, meanwhile, based on the block competition picture identification right, calculates a block head hash value by changing random numbers in a block head, judges whether the hash value meets the block difficulty requirement or not, obtains the picture identification right, the winning operation node sends the picture block to the system operation node and other operation nodes, judges whether the picture block meets the difficulty requirement or not by adopting the voting method, sends the voting result to the system operation node, judges whether the winning node obtains the identification right according to the voting result of each node, once the identification right is obtained, the system operation node identifies the winning node with authorization, the winning node starts to identify the face picture in the picture block, and meanwhile, the system operation node also sends authorization notification to other operation nodes, and the other operation nodes know that the operation nodes have obtained the authorization after receiving the notification, so that the system operation nodes start a new round of picture identification right competition. The operation node which obtains the authorization first needs to identify the work, if no other node obtains the identification authorization after the identification work is completed, the operation node can be added into the round of the identification right competition. The above-described flow is shown in fig. 16. After receiving the authorization notice, the operation node which does not obtain the accounting right needs to mark the pictures contained in the picture blocks in the picture library and delete the pictures temporarily, and the pictures are permanently deleted from the picture library only after receiving the rewarding notice to confirm that the operation node finally wins.
After receiving the identification authorization, the winning operation node starts to consume own computing power to operate face recognition software to identify the face picture, and after the identification is completed, the identification result is sent to other operation nodes and system operation nodes through a network; and the other nodes verify after receiving the information, the verification is passed, the vote is accepted, the vote is not accepted, the voting result is fed back to the system operation node, the rewarding module of the system operation node determines whether to rewards the winning node according to the voting result, the rewarding module informs the winning node of the rewarding condition, and the rewarding module approves the winning operation node to finally win, so that the economic rewards can be directly issued to the winning operation node in a point-to-point mode. Other operation nodes know that winning nodes obtain rewards, the picture block is linked and the picture of the linked chain is deleted from the picture library, and the winning nodes update the picture chain and the picture library; otherwise, the operation node can learn that the identification result of the proved picture is not approved, and then repackage the picture contained in the operation node for next identification right competition. The above-described flow is shown in fig. 17. The identification result is supposed to contain three data of a face picture, a library picture and the similarity between the face picture and the library picture, which are convenient for other nodes to verify, but the face picture and the library picture can cause the data transmitted in the network to be too redundant.
3. As shown in fig. 18, the camera proxy node includes a network routing module and an encryption/decryption module, and mainly plays a role in receiving and forwarding a face picture. The node receives the face picture from the camera group, then encrypts the face picture internally, and then sends the encrypted picture into the network, and the encryption can effectively prevent illegal data from entering the network. After the camera proxy node is started, registration information is sent to the system operation node, the system operation node records the proxy node on line in a registry thereof, and the system operation node is informed that the proxy node exits the network, so that the proxy node information is deleted from the registry.
The camera proxy node does not participate in face recognition work, so that face recognition software, a picture library and a picture chain are not available. The camera proxy node is deployed and managed by the operator.
4. As shown in connection with fig. 19, multiple cameras form a camera group, each camera group corresponding to a camera proxy node. Multiple groups of cameras may be partitioned, one camera proxy node may follow one group or multiple groups, depending on the actual environment and processing power. The camera group is a data source, does not access the P2P network, and does not participate in face recognition.
5. The working steps of the face recognition platform constructed based on the face data distributed recognition architecture designed by the invention are as follows.
(1) The operator first deploys the cameras to make up a group of cameras.
(2) The operator deploys a system operation node, an operation node controlled by the operator and a camera proxy node. The three nodes are started up so that they form a stable P2P network.
(3) The user-controlled operation node downloads an operation node installation package from an operator through the Internet, and the operation node is accessed to the P2P network after the installation is completed; and then registering the face recognition software with the system operation node, and downloading the face recognition software from the system operation node after the registration is successful. So far, the distributed face recognition platform is successfully built.
(4) And the face pictures captured by the camera group are sent to the camera proxy node, encrypted by the camera proxy node and then sent to the P2P network.
(5) The face picture is firstly sent to a node adjacent to the camera agent node; if the neighboring node is also a camera proxy node, the node forwards the face picture to its neighboring node; if the adjacent node is a system operation node or an operation node, the face picture is saved in a local picture library and forwarded to the adjacent node. And the face pictures are stored in a local picture library according to the sequence of the time stamps.
(6) The operation node extracts a plurality of face pictures from a local picture library, and constructs a picture block head based on hash values of the face pictures; then changing random numbers in the picture block heads, calculating whether hash values of the picture block heads meet the difficulty requirement or not until a random number is found so that the hash values of the picture block heads meet the difficulty requirement, acquiring picture identification rights by the operation node, sending the picture block to other operation nodes and system operation nodes through a P2P network, verifying whether the picture block meets the difficulty requirement or not by the nodes, voting, and sending voting results to the system operation node; the system operation node judges whether the winning node really obtains the picture identification right according to the voting results, and informs each operation node of the identification right.
(7) The winning operation node obtains the identification authorization of the system operation node, and starts to carry out face identification on the face picture; and the other operation nodes are informed of authorization and know that the operation node fails in the competition identification right, new face pictures are extracted from the picture library again to form new picture blocks, and a new round of competition of the identification right is started.
(8) After the face recognition work of the winning operation node is finished, the recognition result is sent to the system operation node and other operation nodes, whether the recognition result is correct or not is verified by the nodes, meanwhile, voting is carried out on the verification result, and the voting result is sent to the system operation node; the rewarding module of the system operation node decides whether to issue rewards to winning operation nodes according to voting results, and notifies the rewards results to each operation node; if the winning operation node obtains the rewards, all operation nodes including the winning node and the system operation node hang the picture block to a local picture chain, and meanwhile, the face picture contained in the picture block is deleted from a picture library. If the winning operation node does not obtain rewards, the picture block which is demonstrated by the winning operation node is not approved, the system operation node and other operation nodes cannot hang the picture block to a local picture chain, and meanwhile, the other operation nodes repackage the face picture contained in the picture block into a new picture block for the next round of recognition right competition.
(9) The winning node joins the right competition of the identification of the current round after receiving the rewards.
While the invention has been described with reference to preferred embodiments, it is not intended to be limiting. Those skilled in the art will appreciate that various modifications and adaptations can be made without departing from the spirit and scope of the present invention. Accordingly, the scope of the invention is defined by the appended claims.
Claims (1)
1. The distributed face data recognition and storage architecture based on the blockchain technology is characterized in that the working steps of the whole platform are as follows:
step 1, an operator deploys cameras firstly to form a camera group;
step 2, an operator deploys a system operation node, an operation node controlled by the operator and a camera proxy node; starting the three nodes to form a stable P2P network;
step 3, the operation node controlled by the user downloads an operation node installation package from an operator through the Internet, and the operation node is accessed to the P2P network after the installation is completed; then registering the face recognition software with the system operation node, and downloading the face recognition software from the system operation node after the registration is successful; so far, the distributed face recognition platform is successfully built;
step 4, face pictures captured by the camera group are sent to the camera proxy node, and are sent to the P2P network after being encrypted by the camera proxy node;
step 5, the face picture is firstly sent to a node adjacent to the camera agent node; if the neighboring node is also a camera proxy node, the node forwards the face picture to its neighboring node; if the adjacent node is a system operation node or an operation node, the face picture is saved to a local picture library and is forwarded to the adjacent node; the face pictures are stored in a local picture library according to the sequence of the time stamps;
step 6, the operation node extracts a plurality of face pictures from a local picture library to construct picture blocks, competes for picture identification rights based on the picture blocks constructed by the operation node, sends the picture blocks to other operation nodes and system operation nodes through a P2P network, verifies whether the picture blocks meet requirements and vote by the nodes, and sends voting results to the system operation nodes; the system operation node judges whether the winning node really obtains the picture identification right according to the voting results, and informs each operation node of the identification right;
step 7, the winning operation node obtains the identification authorization of the system operation node, and then the face picture starts to be identified; the other operation nodes are informed of authorization and know that the operation nodes fail in the competition identification right, new face pictures are extracted from the picture library again to form new picture blocks, and a new round of competition of the identification right is started;
step 8, after the face recognition work of the winning operation node is finished, the recognition result is sent to the system operation node and other operation nodes, whether the recognition result is correct or not is verified by the nodes, meanwhile, the verification result is voted, and the voting result is sent to the system operation node; the rewarding module of the system operation node decides whether to issue rewards to winning operation nodes according to voting results, and notifies the rewards results to each operation node; if the winning operation node obtains the rewards, all operation nodes including the winning node and the system operation node hang the picture block to a local picture chain, and meanwhile, the face picture contained in the picture block is deleted from a picture library; if the winning operation node does not obtain rewards, the picture block which is held by the winning operation node is not approved, the system operation node and other operation nodes cannot hang the picture block to a local picture chain, and meanwhile, the other operation nodes repackage the face picture which is contained in the picture block into a new picture block for the next round of recognition right competition;
and 9, the winning node joins the right competition of the identification of the current round after receiving the rewards.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010049817.3A CN111291628B (en) | 2020-01-17 | 2020-01-17 | Face data distributed identification and storage architecture based on block chain technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010049817.3A CN111291628B (en) | 2020-01-17 | 2020-01-17 | Face data distributed identification and storage architecture based on block chain technology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111291628A CN111291628A (en) | 2020-06-16 |
| CN111291628B true CN111291628B (en) | 2024-02-06 |
Family
ID=71023406
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010049817.3A Active CN111291628B (en) | 2020-01-17 | 2020-01-17 | Face data distributed identification and storage architecture based on block chain technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111291628B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112055067B (en) * | 2020-08-28 | 2023-04-18 | 杭州复杂美科技有限公司 | Node connection method, device and storage medium |
| CN112241684A (en) * | 2020-09-16 | 2021-01-19 | 四川天翼网络服务有限公司 | Face retrieval distributed computing method and system |
| CN113408421B (en) * | 2021-06-21 | 2023-04-07 | 湖北央中巨石信息技术有限公司 | Face recognition method and system based on block chain |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106570087A (en) * | 2016-10-20 | 2017-04-19 | 中国电子科技集团公司第二十八研究所 | Distributed face recognition system |
| CN108509274A (en) * | 2018-03-28 | 2018-09-07 | 厦门益东智能科技有限公司 | A kind of method and system based on block chain data automatic marking and intellectual analysis |
| CN109146484A (en) * | 2018-08-31 | 2019-01-04 | 深圳付贝科技有限公司 | Common recognition verification method, digging mine machine and block catenary system based on block chain |
| CN109241111A (en) * | 2018-08-27 | 2019-01-18 | 武汉虹信技术服务有限责任公司 | A kind of distributed face identification system and method for database based on memory |
| CN109800691A (en) * | 2019-01-07 | 2019-05-24 | 深圳英飞拓科技股份有限公司 | Demographics method and system based on face recognition technology |
| CN109872152A (en) * | 2019-01-11 | 2019-06-11 | 平安科技（深圳）有限公司 | Block chain common recognition method and relevant device based on share authorisation verification mechanism |
| CN110138861A (en) * | 2019-05-15 | 2019-08-16 | 北京中星微人工智能芯片技术有限公司 | A kind of security protection Intelligentized method and device |
-
2020
- 2020-01-17 CN CN202010049817.3A patent/CN111291628B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106570087A (en) * | 2016-10-20 | 2017-04-19 | 中国电子科技集团公司第二十八研究所 | Distributed face recognition system |
| CN108509274A (en) * | 2018-03-28 | 2018-09-07 | 厦门益东智能科技有限公司 | A kind of method and system based on block chain data automatic marking and intellectual analysis |
| CN109241111A (en) * | 2018-08-27 | 2019-01-18 | 武汉虹信技术服务有限责任公司 | A kind of distributed face identification system and method for database based on memory |
| CN109146484A (en) * | 2018-08-31 | 2019-01-04 | 深圳付贝科技有限公司 | Common recognition verification method, digging mine machine and block catenary system based on block chain |
| CN109800691A (en) * | 2019-01-07 | 2019-05-24 | 深圳英飞拓科技股份有限公司 | Demographics method and system based on face recognition technology |
| CN109872152A (en) * | 2019-01-11 | 2019-06-11 | 平安科技（深圳）有限公司 | Block chain common recognition method and relevant device based on share authorisation verification mechanism |
| CN110138861A (en) * | 2019-05-15 | 2019-08-16 | 北京中星微人工智能芯片技术有限公司 | A kind of security protection Intelligentized method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111291628A (en) | 2020-06-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111291628B (en) | Face data distributed identification and storage architecture based on block chain technology | |
| Zhang et al. | Edge computing and its role in Industrial Internet: Methodologies, applications, and future directions | |
| EP3920119A1 (en) | Methods for splitting and recovering key, program product, storage medium, and system | |
| CN107528688A (en) | A kind of keeping of block chain key and restoration methods, device based on encryption commission technology | |
| CN111415163B (en) | Block chain-based service processing and verifying method, system and verifying node | |
| KR20190063796A (en) | Identification apparatus and method based on biometric data for blockchain system | |
| CN111291394B (en) | False information management method, false information management device and storage medium | |
| CN110602217A (en) | Block chain-based alliance management method, device, equipment and storage medium | |
| CN112231756A (en) | FL-EM-GMM medical user privacy protection method and system | |
| CN116527372B (en) | Internet-based data security interaction system and method | |
| CN111881487A (en) | Data application system and data application method based on block chain | |
| JP2022512202A (en) | Trust token for resource access | |
| US11790638B2 (en) | Monitoring devices at enterprise locations using machine-learning models to protect enterprise-managed information and resources | |
| CN113872751B (en) | Method, device and equipment for monitoring service data and storage medium | |
| CN114861211A (en) | Meta-universe scene-oriented data privacy protection method, system and storage medium | |
| WO2022201411A1 (en) | Face authentication application using homomorphic encryption | |
| CN112307440B (en) | Method for protecting online education and teaching video property by using blockchain technology | |
| Zhang et al. | Aegis: a trusted, automatic and accurate verification framework for vertical federated learning | |
| CN109067702A (en) | A kind of method that system of real name network identity is generated and protected | |
| Kaveri et al. | Blockchain based reliable electronic voting technology | |
| US11962697B2 (en) | Systems and methods for tracking chain of custody of body worn cameras | |
| Dong et al. | Bdfl: A blockchain-enabled fl framework for edge-based smart uav delivery systems | |
| CN113014540B (en) | Data processing method, device, equipment and storage medium | |
| Roy et al. | Fuzzy miner selection toward Blockchain-based secure communication using multifactor authentication | |
| Xu et al. | A method of federated learning based on blockchain |
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 |