CN105046234A - Invisible recognition method used for human face image in cloud environment and based on sparse representation - Google Patents

Invisible recognition method used for human face image in cloud environment and based on sparse representation Download PDF

Info

Publication number
CN105046234A
CN105046234A CN201510472454.3A CN201510472454A CN105046234A CN 105046234 A CN105046234 A CN 105046234A CN 201510472454 A CN201510472454 A CN 201510472454A CN 105046234 A CN105046234 A CN 105046234A
Authority
CN
China
Prior art keywords
client
image
vector
euclidean distance
server
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201510472454.3A
Other languages
Chinese (zh)
Other versions
CN105046234B (en
Inventor
金鑫
刘妍
赵耿
李晓东
郭魁
陈迎亚
田玉露
叶超尘
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shaoding Artificial Intelligence Technology Co.,Ltd.
Original Assignee
BEIJING ELECTRONIC SCIENCE AND TECHNOLOGY INSTITUTE
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by BEIJING ELECTRONIC SCIENCE AND TECHNOLOGY INSTITUTE filed Critical BEIJING ELECTRONIC SCIENCE AND TECHNOLOGY INSTITUTE
Priority to CN201510472454.3A priority Critical patent/CN105046234B/en
Publication of CN105046234A publication Critical patent/CN105046234A/en
Application granted granted Critical
Publication of CN105046234B publication Critical patent/CN105046234B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

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
    • G06V40/161—Detection; Localisation; Normalisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Multimedia (AREA)
  • General Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Human Computer Interaction (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Software Systems (AREA)
  • Collating Specific Patterns (AREA)
  • Processing Or Creating Images (AREA)

Abstract

The present invention discloses an invisible recognition method used for a human face image in a cloud environment and based on sparse representation, which can be carried out in a safe manner and also protects privacy of the image and invisibility of a database. An image database is stored in a cloud server; a client acquires the human face image and the server needs to determine whether the human face image is matched with a criminal suspect or not; meanwhile, the server does not acquire more information of both parties. According to the invisible recognition method, the sparse representation is applied to an invisible human face recognition security protocol for the first time, and a safe Euclidean distance algorithm is introduced; a Paillier homomorphic encryption and oblivious transferring algorithm is utilized, and sparse representation coefficient vectors of terminal and cloud human faces are invisibly compared, so that the number of dimensions of human face representation vectors is reduced and attacks based on image blocks are also avoided. Moreover, the method is very easy to realize through software; and the method can be widely applied and popularized to cloud computation, security authentication, criminal suspect tracing and the like.

Description

Based on the concealed recognition methods of facial image in the cloud environment of rarefaction representation
Technical field
The invention belongs to cryptography, computer vision field, the method for particularly concealed recognition of face, specifically based on the concealed recognition methods of facial image in the cloud environment of rarefaction representation.
Background technology
Recognition of face plays important role in video monitoring and safety.Along with the fast development of computer technology, cloud computing has changed the mode of traditional face identification system.Video and the large data of facial image and powerful face identification system are stored in high in the clouds and run beyond the clouds, and this provide one and apply such as face search widely, suspect searches for.But a large amount of detection cameras is distributed in public place, the privacy of people exposes undoubtedly completely.Suspect searches application may be utilized by lawless person the people going to search them and want to find.Once face identification system is connected to a general database as I.D., the removal search common people that some people just can make one's wish fulfilled.Suspect's database also may expose and even cause more crime on the other hand.
Paillier system be a kind of there is Semantic Security add homomorphism common key cryptosystem, so be usually used in structure multi-party computations basic agreement, as scalar product protocol, 0T agreement etc.If cryptographic operation in cipher system is designated as E pk(), decryption oprerations is designated as D pk(), if so semantic security refers to any message m 0, m 1, there is not any polynomial time algorithm and distinguish E pk(m 0) and E pk(m 1), add isomorphism and refer to E pk(x, r 1) E pk(y, r 2)=E pk(x+y, r 1r 2), r 1, r 2be random number, easily verify that Paillier system has semantic security and adds isomorphism according to the knowwhy of Paillier algorithm.
Oblivious transfer protocol is a kind of cipher protocol protecting privacy, is also a kind of intercommunication agreement protecting privacy, and communicating pair can be made to transmit message in a kind of mode of obfuscation of selecting.Can be interpreted as simply, Oblivious Transfer can make communicating pair transmit message in the casual mode of one.In specific occasion with under needing, it is the selection protecting the privacy Oblivious Transfer of user to provide a kind of reality.
Summary of the invention
The present invention wants technical solution problem to be: overcome the deficiencies in the prior art, there is provided a kind of based on the concealed recognition methods of facial image in the cloud environment of rarefaction representation, the method effectively can improve the counting yield of safe recognition of face, and effectively can resist the attack based on image block Recovery image.
The technical solution used in the present invention is: a kind of based on the concealed recognition methods of facial image in the cloud environment of rarefaction representation, performing step is as follows:
(1) training sample totally 100 width facial image training face dictionaries;
(2) client and server end calculates the rarefaction representation vector of its image respectively;
(3) client sends to server by after the encryption of the face of calculating vector, and server calculates and receives ciphertext that the is vectorial and Euclidean distance of any one image vector in self picture library, and ciphertext is sent back to client, and client solves Euclidean distance expressly;
(4) client is carried out alternately according to Euclidean distance recycling oblivious transfer protocol and server end;
(5) server end calculates the threshold value of each facial image Euclidean distance of database in advance by great many of experiments, the Euclidean distance calculated and corresponding threshold value can be compared and judge whether that the face of client and server mates in Oblivious Transfer.
Wherein, the image face dictionary of described step (1) and (2) and coefficient vector take following steps:
(11) training sample of training face dictionary is 5 different facial images of 20 people, totally 100 width training samples;
(12) after image compresses process in advance, such a matrix can be obtained: all pixels (being 18 pixel values after compression) order arrangement of piece image is shown in each list, 100 width facial images are arranged in order according to the mode classification of 20 different people again, after standardization computing, obtain the normalized matrix of 18 × 200, namely obtain face dictionary;
(13) facial image of k × j size is regarded as a column vector v ∈ R m(m=kj).Use matrix represent all training samples in the i-th class, a training sample in this classification, n are shown in its each list irepresent the number of all training samples in this classification, training sample matrix A = [ A 1 , A 1 , . . . , A K ] = [ v 1,1 , v 1,2 , . . . v 1 , n k ] , Test sample y can be expressed as y=Ax ∈ R again m;
(14) finally solve the most sparse solution and obtain rarefaction representation vector.
Wherein, the method for the compute euclidian distances described in step (3) takes following steps:
(21) first client carries out square operation by turn to vector, then is encrypted by turn respectively by a former vector sum square vector, and the result of encryption sends to server end;
(22) after received server-side to two vectors of secret, the Euclidean distance of the character utilizing homomorphism to add and Euclidean distance formulae discovery both sides coefficient vector, this carries out under ciphertext condition, and adds that a random number sends to client to result;
(23) client by the Euclidean distance under ciphertext state and random number and be decrypted, just obtain Euclidean distance under expressly state and random number with.
Wherein, step (4) client utilizes oblivious transfer protocol and server end to carry out reciprocal process and takes following steps:
(31) client generates the private key of symmetrical secret key, and server end generates the public private key pair of multiple asymmetric secret key, and client is selected a PKI to be encrypted its private key and sent to server end;
(32) its all private key of server is decrypted ciphertext, and with decrypted result as the encryption of secret key pair matching result, multiple encrypted result is returned to client;
(33) the client information of selecting its decrypt ciphertext that can decipher whether to be mated with symmetrical secret key.
Principle of the present invention is:
According to defect and the deficiency of current concealed face recognition scheme, design some rules of concealed face recognition algorithms based on rarefaction representation can be summed up, as described below:
(1) the same with any biometric data, between the image obtained by terminal and the list image of existence in same person, can not mate completely.Therefore, the face recognition algorithms that practical must be used.
(2) the coupling work of face must complete under the mode of a secret protection.That is, the input that high in the clouds and terminal all can not be known except terminal whether mate with face in the list of high in the clouds except any information.Realize this goal and need searching people's face recognition algorithms, this algorithm, in different light, can show under the conditions such as expression and well identify robustness, and can support safety compute agreement;
(3) safety of structure of concealed face recognition scheme is extremely important, and the defect of some structures can expose the information of facial image, and it is also possible that public database is exposed completely;
(4) the normally used data Biao Shi real number field of face recognition algorithms, and security protocol is operated in Galois field, existing face identification method is transformed into Galois field may be caused degenerating.The mathematical operation that can realize in existing cryptographic algorithm is also very conditional.
According to above-mentioned rule, the present invention utilizes rarefaction representation, Paillier homomorphic cryptography, Euclidean distance and Oblivious Transfer (OT), devises a kind of concealed face recognition scheme newly.In this scenario, rarefaction representation is used for generating facial image and represents vector, then generates vector and utilizes homomorphism to add Euclidean distance between compute vector, eventually passes OT operation and result is returned to client.In order to strengthen the robustness of identification, generating face representation vector by rarefaction representation, then carrying out similarity system design computing, this effectively can shorten the dimension of vector and improve the efficiency of encrypting.The realization of Oblivious Transfer utilizes symmetric cryptography and asymmetric encryption.Experimental analysis shows concealed rarefaction representation recognition of face, can be applicable to actual face recognition application.
The present invention compared with prior art, it is advantageous that:
(1) can show under the conditions such as algorithm is in different light, expression and well identify robustness, and safety compute agreement can be supported;
(2) utilize face sparse to represent to have broken with Euclidean distance the final conclusion only having binary vector can do the recognition of face under concealed condition, improve the efficiency of algorithm, reduce the dimension of face representation vector, and program shortens the time of test, the attack recovering original image based on image block also effectively can be resisted.
(3) face recognition scheme structure is simple, is easy to realize.
Accompanying drawing explanation
Fig. 1 is application scenarios figure of the present invention;
Fig. 2 is the present invention program's process flow diagram.
Embodiment:
Below in conjunction with the drawings and specific embodiments, the present invention is further detailed explanation.
Adding and counting multiplicative matter of Paillier homomorphism is described by formula (1) and formula (2):
E ( m 1 ) ≡ g m 1 · x 1 N ( mod N 2 )
E ( m 2 ) ≡ g m 2 · x 2 N ( mod N 2 )
E ( m 1 ) · E ( m 2 ) ≡ g m 1 · x 1 N · g m 2 · x 2 N ( mod N 2 ) ≡ g m 1 + m 2 ( x 1 · x 2 ) N mod N 2 - - - ( 1 )
E(km 1)≡E(m 1) k(2)
Wherein m 1, m 2represent two plaintexts, E () homomorphism equation, N=pq, p and q are two Big prime N ∈ Z, random number m ∈ Z n, G is mould n 2multiplicative group, namely stochastic choice g ∈ G, makes g meet gcd (L (g emodN 2), N)=1, then the PKI of this encryption system is (g, n), and private key is e (N).E (N) and L () is defined as follows:
Z N＝{x|x∈Z，0≤x≤N}，
Z N * = { x | x &Element; Z , 0 &le; x < N , gcd ( x , N ) = 1 } ,
e(N)＝1cm(p-1，q-1)，(3)
S＝{x＜N 2|x＝1modN}，
&ForAll; x &Element; S , L ( x ) = x - 1 N
Consult Fig. 2 protocol procedures figure, ciphering process of the present invention can be divided into following concrete steps:
Input step:
Client input one dimension face vector s=(s 0, s 1..., s l-1), l=200 in the present invention, can change according to actual conditions in the middle of other experiments.Client inputs one group Q one dimension face vector { s 1, s 2..., s qand random number (t 1, t 2..., t qeach s corresponding i, both sides are to Euclidean distance upper limit d max.
Export step:
Client knows index i, Euclidean distance ED (s, s i)≤t i, and server end does not know more information.
(1) client bitwise encryption face vector s=(s 0, s 1..., s l-1) and vector square (s) 2=((s 0) 2, (s 1) 2..., (s l-1) 2, send to server, received server-side is to encrypted result (E pk(s 0), E pk(s 1) ..., E pk(s l-1)) and (E pk((s 0) 2), E pk((s 1) 2) ..., E pk((s l-1) 2)), be repetition for the face following steps of often opening of one group of suspect of server end;
(2) for the jth element in the face vector of i-th in server, cloud server calculates E pk(v j) wherein:
v j = ( s j - s j i ) 2
E p k ( ( s j - s j i ) 2 ) = E p k ( ( s j ) 2 - 2 s j s j i + ( s j i ) 2 ) = E p k ( ( s j ) 2 ) &CenterDot; E p k ( ( s j ) 2 s j i ) &CenterDot; E p k ( ( s j i ) 2 ) - - - ( 4 )
(3) according to the character of homomorphism, the server in high in the clouds can pass through calculate d e=(ED (s, s i) 2, d e∈ [0, d max].Then a random number r is selected to each face iand calculate E p k ( ( E D ( s , s i ) ) 2 + r i ) Send to client;
(4) client receives E pk((ED (s, s i)) 2+ r i) and deciphered;
(5) both sides use agreement goes to judge whether (d e) i< t i, at client result of calculation R i
(6) draw whether mate, 1 represents coupling, and 0 expression is not mated.
In a word, the face sparse of the secret proposed in the present invention represents that recognition methods can be carried out under the agreement of a safety, meanwhile can protect the data security of client and server end both sides.Rarefaction representation is applied in the security protocol of concealed recognition of face by the present invention first, this not only lowers the dimension of face representation vector and the attack can resisted based on image block Recovery image information.In addition, we talk of safe Euclidean distance algorithm, solve the difficult problem that nonbinary vector can not carry out safe computing.The method demonstrating the present invention's proposition by experiment effectively can shorten the efficiency of the dimension raising encryption and decryption of vector, reduces calculated amount, improves recognition efficiency and shorten recognition time.And this encryption method is easily via software simulating, the present invention can be generalized in recognition of face safe storage and Transmission Encryption in widespread use.
The foregoing is only basic explanations more of the present invention, any equivalent transformation done according to technical scheme of the present invention, all should belong to protection scope of the present invention.

Claims (4)

1., based on the concealed recognition methods of facial image in the cloud environment of rarefaction representation, it is characterized in that performing step:
(1) training sample totally 100 width facial image training face dictionaries;
(2) client and server end calculates the rarefaction representation vector of its image respectively;
(3) client sends to server by after the encryption of the face of calculating vector, and server calculates and receives ciphertext that the is vectorial and Euclidean distance of any one image vector in self picture library, and ciphertext is sent back to client, and client solves Euclidean distance expressly;
(4) client is carried out alternately according to Euclidean distance recycling oblivious transfer protocol and server end;
(5) server end calculates the threshold value of each facial image Euclidean distance of database in advance by great many of experiments, the Euclidean distance calculated and corresponding threshold value can be compared and judge whether that the face of client and server mates in Oblivious Transfer.
2. according to claim 1 based on the concealed recognition methods of facial image in the cloud environment of rarefaction representation, it is characterized in that: step (1) and (2) described image face dictionary and coefficient vector take following steps:
(11) training sample of training face dictionary is 5 different facial images of 20 people, totally 100 width training samples;
(12) after image compresses process in advance, such a matrix can be obtained: all pixel order arrangements of piece image are shown in each list, be 18 pixel values after this all pixel compression, 100 width facial images are arranged in order according to the mode classification of 20 different people again, after standardization computing, obtain the normalized matrix of 18 × 200, namely obtain face dictionary;
(13) facial image of k × j size is regarded as a column vector v ∈ R m(m=kj), matrix is used represent all training samples in the i-th class, a training sample in this classification, n are shown in its each list irepresent the number of all training samples in this classification, training sample matrix test sample y can be expressed as y=Ax ∈ R again m;
(14) finally solve the most sparse solution and obtain rarefaction representation vector.
3. according to claim 1 based on the concealed recognition methods of facial image in the cloud environment of rarefaction representation, it is characterized in that: the method for the compute euclidian distances described in step (3) takes following steps:
(21) first client carries out square operation by turn to vector, then is encrypted by turn respectively by a former vector sum square vector, and the result of encryption sends to server end;
(22) after received server-side to two vectors of encryption, the Euclidean distance of the character utilizing homomorphism to add and Euclidean distance formulae discovery both sides coefficient vector, this carries out under ciphertext condition, and adds that a random number sends to client to result;
(23) client by the Euclidean distance under ciphertext state and random number and be decrypted, just obtain Euclidean distance under expressly state and random number with.
4. according to claim 1 based on the concealed recognition methods of facial image in the cloud environment of rarefaction representation, it is characterized in that: step (4) described client utilizes oblivious transfer protocol and server end to carry out reciprocal process and takes following steps:
(31) client generates the private key of symmetric cryptography, and server end generates the public private key pair of multiple asymmetric encryption, and client is selected a PKI to be encrypted its private key and sent to server end;
(32) its whole private key of server is decrypted ciphertext, and with decrypted result as the encryption of secret key pair matching result, multiple encrypted result is returned to client;
(33) the symmetrical secret key of client selects its decrypt ciphertext that can decipher to obtain the information of matching result.
CN201510472454.3A 2015-08-04 2015-08-04 Facial image secret recognition methods in cloud environment based on rarefaction representation Active CN105046234B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510472454.3A CN105046234B (en) 2015-08-04 2015-08-04 Facial image secret recognition methods in cloud environment based on rarefaction representation

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510472454.3A CN105046234B (en) 2015-08-04 2015-08-04 Facial image secret recognition methods in cloud environment based on rarefaction representation

Publications (2)

Publication Number Publication Date
CN105046234A true CN105046234A (en) 2015-11-11
CN105046234B CN105046234B (en) 2018-05-11

Family

ID=54452763

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510472454.3A Active CN105046234B (en) 2015-08-04 2015-08-04 Facial image secret recognition methods in cloud environment based on rarefaction representation

Country Status (1)

Country Link
CN (1) CN105046234B (en)

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105721140A (en) * 2016-01-27 2016-06-29 北京航空航天大学 K out of n oblivious transfer method and system
CN106096548A (en) * 2016-06-12 2016-11-09 北京电子科技学院 A kind of many intelligent terminal based on cloud environment share face secret recognition methods
CN106127666A (en) * 2016-06-12 2016-11-16 北京电子科技学院 Subject image secret detection method in a kind of cloud environment represented based on random subgraph
CN108171262A (en) * 2017-12-22 2018-06-15 珠海习悦信息技术有限公司 The recognition methods of ciphertext picture/mb-type, device, storage medium and processor
CN109359210A (en) * 2018-08-09 2019-02-19 中国科学院信息工程研究所 The face retrieval method and system of double blind secret protection
CN111241514A (en) * 2020-01-14 2020-06-05 浙江理工大学 Safety face verification method based on face verification system
US10713544B2 (en) 2018-09-14 2020-07-14 International Business Machines Corporation Identification and/or verification by a consensus network using sparse parametric representations of biometric images
CN112215158A (en) * 2020-10-13 2021-01-12 中山大学 Face recognition method fusing fully homomorphic encryption and discrete wavelet transform in cloud environment
CN112287375A (en) * 2020-11-21 2021-01-29 上海同态信息科技有限责任公司 Method for calculating dense state Euclidean distance
CN113946858A (en) * 2021-12-20 2022-01-18 湖南丰汇银佳科技股份有限公司 Identity security authentication method and system based on data privacy calculation
WO2024027066A1 (en) * 2022-08-04 2024-02-08 中国银联股份有限公司 Data matching method, apparatus and system, and device and medium

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101674299A (en) * 2009-10-16 2010-03-17 西安电子科技大学 Method for generating key based on amalgamation of multiple features in encryption area
CN101753304A (en) * 2008-12-17 2010-06-23 中国科学院自动化研究所 Method for binding biological specificity and key
CN101976321A (en) * 2010-09-21 2011-02-16 北京工业大学 Generated encrypting method based on face feature key

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101753304A (en) * 2008-12-17 2010-06-23 中国科学院自动化研究所 Method for binding biological specificity and key
CN101674299A (en) * 2009-10-16 2010-03-17 西安电子科技大学 Method for generating key based on amalgamation of multiple features in encryption area
CN101976321A (en) * 2010-09-21 2011-02-16 北京工业大学 Generated encrypting method based on face feature key

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
刘妍等: "基于稀疏表示的云环境中人脸图像隐秘识别方法", 《系统仿真学报》 *

Cited By (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105721140A (en) * 2016-01-27 2016-06-29 北京航空航天大学 K out of n oblivious transfer method and system
CN105721140B (en) * 2016-01-27 2019-03-15 北京航空航天大学 N takes the Oblivious Transfer method and system of k
CN106096548A (en) * 2016-06-12 2016-11-09 北京电子科技学院 A kind of many intelligent terminal based on cloud environment share face secret recognition methods
CN106127666A (en) * 2016-06-12 2016-11-16 北京电子科技学院 Subject image secret detection method in a kind of cloud environment represented based on random subgraph
CN106127666B (en) * 2016-06-12 2019-02-19 北京电子科技学院 It is a kind of based on random subgraph indicate cloud environment in subject image secret detection method
CN106096548B (en) * 2016-06-12 2019-05-24 北京电子科技学院 A kind of shared face secret recognition methods of more intelligent terminals based on cloud environment
CN108171262A (en) * 2017-12-22 2018-06-15 珠海习悦信息技术有限公司 The recognition methods of ciphertext picture/mb-type, device, storage medium and processor
CN109359210A (en) * 2018-08-09 2019-02-19 中国科学院信息工程研究所 The face retrieval method and system of double blind secret protection
US10922588B2 (en) 2018-09-14 2021-02-16 International Business Machines Corporation Identification and/or verification by a consensus network using sparse parametric representations of biometric images
US10713544B2 (en) 2018-09-14 2020-07-14 International Business Machines Corporation Identification and/or verification by a consensus network using sparse parametric representations of biometric images
CN111241514A (en) * 2020-01-14 2020-06-05 浙江理工大学 Safety face verification method based on face verification system
CN111241514B (en) * 2020-01-14 2022-05-31 浙江理工大学 Safety face verification method based on face verification system
CN112215158A (en) * 2020-10-13 2021-01-12 中山大学 Face recognition method fusing fully homomorphic encryption and discrete wavelet transform in cloud environment
CN112215158B (en) * 2020-10-13 2022-10-18 中山大学 Face recognition method fusing fully homomorphic encryption and discrete wavelet transform in cloud environment
CN112287375A (en) * 2020-11-21 2021-01-29 上海同态信息科技有限责任公司 Method for calculating dense state Euclidean distance
CN113946858A (en) * 2021-12-20 2022-01-18 湖南丰汇银佳科技股份有限公司 Identity security authentication method and system based on data privacy calculation
WO2024027066A1 (en) * 2022-08-04 2024-02-08 中国银联股份有限公司 Data matching method, apparatus and system, and device and medium

Also Published As

Publication number Publication date
CN105046234B (en) 2018-05-11

Similar Documents

Publication Publication Date Title
CN105046234A (en) Invisible recognition method used for human face image in cloud environment and based on sparse representation
Zhang et al. A review of compressive sensing in information security field
CN109145829A (en) A kind of safe and efficient face identification method based on deep learning and homomorphic cryptography
Zhang et al. Secure transmission of compressed sampling data using edge clouds
CN101984576B (en) Method and system for authenticating anonymous identity based on face encryption
CN109165581A (en) A kind of secret protection face identification method based on homomorphic cryptography
CN112215158B (en) Face recognition method fusing fully homomorphic encryption and discrete wavelet transform in cloud environment
CN105553980A (en) Safety fingerprint identification system and method based on cloud computing
CN104821942A (en) Face recognition method and system
Wang Side-channel analysis of AES based on deep learning
Tang et al. Robust multiple color images encryption using discrete Fourier transforms and chaotic map
CN104618113B (en) The method that the authentication of a kind of mobile terminal and safe lane are set up
CN116361649A (en) Efficient unbalanced PSI (program specific information) based on bloom filter and hash
Ahmad et al. A partial ligt-weight image encryption scheme
CN115913537A (en) Data intersection method and system based on privacy protection and related equipment
Jin et al. Privacy preserving face identification in the cloud through sparse representation
CN112380404B (en) Data filtering method, device and system
CN113556225A (en) Efficient PSI (program specific information) method based on Hash and key exchange
WO2023093278A1 (en) Digital signature thresholding method and apparatus
Jin et al. Efficient blind face recognition in the cloud
Singhai et al. An efficient image security mechanism based on advanced encryption standard
EP3883178A1 (en) Encryption system and method employing permutation group-based encryption technology
Gong Plaintext recovery attack on 3DES algorithm with different byte keys
Vanitha et al. Enhanced Security using Advanced Encryption Standards in Face Recognition
Li et al. Ciphertext face recognition system based on secure inner product protocol

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20191106

Address after: 311800 no.30-1 and no.30-2, Likai Road, Jiyang street, Zhuji City, Shaoxing City, Zhejiang Province

Patentee after: Zhejiang Jun Kai Electromechanical Technology Co. Ltd.

Address before: 100070 Beijing city Fengtai District Fung Fu Road Beijing Electronics Science and Technology Institute No. 7 Building 413

Patentee before: Beijing Electronic Science and Technology Institute

TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20191225

Address after: Room 302-9, Qunxian road 2003, Keqiao street, Keqiao District, Shaoxing City, Zhejiang Province

Patentee after: Zhejiang shaoding intelligent robot Co., Ltd

Address before: 311800 no.30-1 and no.30-2, Likai Road, Jiyang street, Zhuji City, Shaoxing City, Zhejiang Province

Patentee before: Zhejiang Jun Kai Electromechanical Technology Co. Ltd.

TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20200826

Address after: 610-2, floor 6, No. 8, caihefang Road, Haidian District, Beijing 100000

Patentee after: Shaoding Artificial Intelligence Technology Co.,Ltd.

Address before: Room 302-9, Qunxian road 2003, Keqiao street, Keqiao District, Shaoxing City, Zhejiang Province

Patentee before: Zhejiang shaoding intelligent robot Co.,Ltd.