CN103577738B - Based on hereditary automatic mold clustering analysis without template biological key generation method - Google Patents

Based on hereditary automatic mold clustering analysis without template biological key generation method Download PDF

Info

Publication number
CN103577738B
CN103577738B CN201310375864.7A CN201310375864A CN103577738B CN 103577738 B CN103577738 B CN 103577738B CN 201310375864 A CN201310375864 A CN 201310375864A CN 103577738 B CN103577738 B CN 103577738B
Authority
CN
China
Prior art keywords
features component
key
cluster
cluster result
stability
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
Application number
CN201310375864.7A
Other languages
Chinese (zh)
Other versions
CN103577738A (en
Inventor
盛伟国
白丽叶
应豪超
卢梦雅
陈胜勇
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.)
Zhejiang University of Technology ZJUT
Original Assignee
Zhejiang University of Technology ZJUT
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 Zhejiang University of Technology ZJUT filed Critical Zhejiang University of Technology ZJUT
Priority to CN201310375864.7A priority Critical patent/CN103577738B/en
Publication of CN103577738A publication Critical patent/CN103577738A/en
Application granted granted Critical
Publication of CN103577738B publication Critical patent/CN103577738B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
    • G06F21/31—User authentication
    • G06F21/32—User authentication using biometric data, e.g. fingerprints, iris scans or voiceprints
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/12—Computing arrangements based on biological models using genetic models
    • G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming

Abstract

Based on hereditary automatic mold clustering analysis without template biological key generation method, step is as follows: 1) obtain some biometric sample from user, calculate the statistical characteristics of each biometric sample, application heredity automatically fuzzy clustering algorithm carries out fuzzy cluster analysis to data.2) by fuzzy clustering result, calculate the average blur degree of membership of each features component and each bunch, determine the stability of features component.The stability that angle value weighs features component is subordinate to what obtain.3) selection of features component.Based on the features component stability result obtained in previous step, select the higher features component of stability in order to generate key.4) generation of biological secret key.After determining the features component of each user, for the cluster result in each features component, each bunch can mark by a secret key bits.Determine the corresponding home cluster of each selected features component and secret key bits, by combining all secret key bits obtained, namely can be each user and generating key.

Description

Based on hereditary automatic mold clustering analysis without template biological key generation method
Technical field
The present invention relates to technical field of biometric identification, be specifically related to a kind of based on hereditary automatic mold clustering analysis without template biological key generation method, for the authentication of security system.
Background technology
Authentication is key components of current many security systems, and particularly under the background of ecommerce widespread use, safe and reliable auth method is increasingly important.
The object (as I.D.) that traditional identity verification method mainly utilizes user to have or secret knowledge (as password) carry out identity verification.Although these technology have been widely used in existing many security systems, there is shortcomings in them, and such as, these systems cannot prevent the personnel's access system being obtained checking object or access code by back door.Recently, people start to adopt biometrics identification technology to carry out authentication.Auth method based on living things feature recognition can overcome the shortcoming existed in existing auth method, meanwhile, compares traditional auth method and has the advantages such as esy to use, safe and reliable.
At present, biometrics identification technology application in the security system receives much attention.Conventional biological feather recognition method gathers the biometric sample (i.e. physiological characteristic/behavioural characteristic) of user as template at registration phase usually, is then together stored in system database with user name, key and access rights etc.At Qualify Phase, extract the biometric sample of user and mate with the biological information stored in database, as the match is successful then by verifying and discharging key.But, store biological information and introduce many safety problems as template, there is the hidden danger revealing privacy simultaneously.First, the biological information of storage be easily stolen and in order to copy, thus trespass system.Secondly, the biological information of storage easily exposes individual privacy (such as, retina textural characteristics can reflect the relevant disease such as diabetes and apoplexy).In addition, in these methods, biometric matches and password release module are normally separated completely.Password release module determines whether discharging password by receiving successful or failed (the namely 1 or 0) signal of biometric matches.This simple two-value exports and is easy to victim rewriting, thus reaches the success attack rate of 50%.In order to solve these safety problems of existence and leak the hidden danger of privacy, the present invention proposes a kind of new method, directly from the statistical nature that biometric sample is extracted, generates key, realizes without the reliable authentication of template biological secret key safety.
Summary of the invention
The object of the invention is to solve the existing shortcoming and defect based on existing in the auth method of biometrics identification technology, propose a kind of based on hereditary automatic mold clustering analysis without template biological key generation method, to realize safe and reliable authentication.The method directly generates key based on the statistical nature extracted in biological attribute data, without the need to storing biometric templates and key.
The present invention solves the scheme that its technical matters adopts:
Based on hereditary automatic mold clustering analysis without a template biological key generation method, comprise the steps:
Step 1: gather user biological feature samples and extract statistical nature as training data, application heredity automatically fuzzy clustering algorithm carries out cluster analysis to each features component (i.e. single statistical nature or statistical nature collection).
Step 2: the stability determining features component.Calculated the average blur degree of membership of each features component by the cluster result of training data, the value obtained is in order to determine the stability of features component.
Step 3: the selection of features component.According to result obtained in the previous step, select the features component that stability is high, in selection course, need the compromise problem simultaneously considered between key stability and key length.
Step 4: the generation of key.From sample to be verified, extract selected statistical nature assembly and calculate its value, then determining corresponding bunch and secret key bits according to the value of these features.The secret key bits obtained represents with binary value, and all secret key bits are combined generation key.
Technical conceive of the present invention is: for potential problem potential in the shortcoming existed in traditional auth method and existing biological identification technology, propose a kind of based on hereditary automatic mold clustering analysis without template biological key generation method.
Heredity is fuzzy clustering algorithm automatically
Training dataset X={x is extracted from the biological specimen gathered 1, x 2..., x n, wherein, x ifor the eigenvector on d dimension space, n is the number of sample.K fuzzy clustering is comprised, if a kth fuzzy clustering center is C in data set X k, (k=1,2 ..., K), then cluster centre is C={C 1, C 2..., C k.Cluster result need meet with properties:
Σ i = 1 n u ki ≥ 1 ( k = 1 , . . . , K ) , Σ k = 1 K u ki ≥ 1 ( i = 1 , . . . , n ) , Σ k = 1 K Σ i = 1 n u ki = n ,
Wherein u kidata centralization i-th element x iwith a kth fuzzy clustering center C kdegree of membership.This clustering problem can be converted into cluster matrix U (X)=[u finding the given fuzzy clustering standard of optimization ki], 1≤k≤K, 1≤i≤n.The present invention adopts the automatic fuzzy clustering algorithm of heredity to solve this problem.
Following steps operation is carried out to each features component in training data:
1. random produce P cluster result (representing using bunch center) as initial population, each cluster result can comprise varying number bunch.
2. adopt the suitability degree of each cluster result in Xie-Beni (XB) index calculate initial population.
Xie-Beni (XB) index is:
The suitability degree of each cluster result is defined as: f=1/XB.
Wherein, s=min i ≠ j{ D 2(z i-z j), D () is the measuring distance based on Euclidean tolerance, z krepresent kGe Cu center, m is weighted index.
3. repeat following (a) to (d) operation until meet end condition (namely best in colony cluster result does not all change in n generation develops).
A) cluster result adopting tournament selection method to select two relative suitability degrees high from colony partners father and mother.Repeat this operation until select P/2 to father and mother.
B) on the father and mother selected, application interlace operation generates P offspring, then to each rear substitute performance mutation operation.
C) suitability degree of each offspring is calculated.
D) create colony of new generation, by the past generation colony and progeny population in select the cluster result of P high suitability degree, and replace last generation colony, get back to step (a).
4. export the best cluster result found in evolutionary process.
As can be seen from technique scheme, effect of the present invention is mainly manifested in: for each user selects the features component that stability is high individually, generates key to carry out authentication under the prerequisite not damaging security of system.The method, without the need to storing biometric templates and key, directly generates key from the statistical nature of biological attribute data, and the harm that the storage avoiding template or key brings has boundless application prospect.
Accompanying drawing explanation
Fig. 1 is the example of the present invention at the upper cluster result of a features component (comprising two features).
Fig. 2 is that the present invention is at data set DB 1and DB 2fRR and FAR during the different significance bit key of upper generation.
Fig. 3 is that the present invention is at data set DB 3fRR and FAR during the different significance bit key of upper generation.
Embodiment
Below in conjunction with accompanying drawing, the invention will be further described.
Specific embodiment of the invention process comprises the following steps:
Step 1: obtain some biometric sample from user, calculates the statistical characteristics of each biometric sample.Based on the statistical characteristics calculated (i.e. training data), application heredity automatically fuzzy clustering algorithm carries out fuzzy cluster analysis to data.
Because the data object from same user is normally similar, often by cluster to same bunch, especially when the stability of user characteristics assembly is higher.On the contrary, from different user data object often by cluster to different bunches.Therefore, cluster result can be used for simulating the difference between colony inside and colony.
Heredity is fuzzy clustering algorithm automatically
Training dataset X={x is extracted from the biological specimen gathered 1, x 2..., x n, wherein, x ifor the eigenvector on d dimension space, n is the number of sample.K fuzzy clustering is comprised, if a kth fuzzy clustering center is C in data set X k, (k=1,2 ..., K), then cluster centre is C={C 1, C 2..., C k.Cluster result need meet with properties:
Σ i = 1 n u ki ≥ 1 ( k = 1 , . . . , K ) , Σ k = 1 K u ki ≥ 1 ( i = 1 , . . . , n ) , Σ k = 1 K Σ i = 1 n u ki = n ,
Wherein u kidata centralization i-th element x iwith a kth fuzzy clustering center C kdegree of membership.This clustering problem can be converted into cluster matrix U (X)=[u finding the given fuzzy clustering standard of optimization ki], 1≤k≤K, 1≤i≤n.The present invention adopts the automatic fuzzy clustering algorithm of heredity to solve this problem.
Following steps operation is carried out to each features component in training data:
1. random produce P cluster result (representing using bunch center) as initial population, each cluster result can comprise varying number bunch.
2. adopt the suitability degree of each cluster result in Xie-Beni (XB) index calculate initial population.
Xie-Beni (XB) index is:
The suitability degree of each cluster result is defined as: f=1/XB.
Wherein, s=min i ≠ j{ D 2(z i-z j), D () is the measuring distance based on Euclidean tolerance, z krepresent kGe Cu center, m is weighted index.
3. repeat following (a) to (d) operation until meet end condition (namely best in colony cluster result does not all change in n generation develops).
A) cluster result adopting tournament selection method to select two relative suitability degrees high from colony partners father and mother.Repeat this operation until select P/2 to father and mother.
B) on the father and mother selected, application interlace operation generates P offspring, then to each rear substitute performance mutation operation.
C) suitability degree of each offspring is calculated.
D) create colony of new generation, by the past generation colony and progeny population in select the cluster result of P high suitability degree, and replace last generation colony, get back to step (a).
4. export the best cluster result found in evolutionary process.
Step 2: the stability determining features component.
By fuzzy clustering result, the average blur degree of membership of each features component and each bunch can be calculated.The maximum membership degree value obtained is by the stability in order to weigh these features component.
Step 3: the selection of features component.Based on the features component stability result obtained in previous step, select the higher features component of stability in order to generate key.In the selection process, there is certain trade-off relationship between the reliability of key and key length, this trade-off relationship controls by threshold value.If threshold value is less, then more features component is by selected, and the key obtained is longer, and key reliability will be lower.On the contrary, if threshold value is comparatively large, the key obtained is shorter, and the reliability of key will be higher.
It should be noted that selected features component should without any plyability, namely any features component can not repeatedly for generating key; Otherwise the easy victim of information of this overlap is in order to guess key.Meanwhile, in order to obtain longer key, features component as much as possible should be selected.Therefore, low-dimensional features component is first considered when selecting features component.
Step 4: the generation of biological secret key.
After determining the features component of each user, key can generate thus.For the cluster result in each features component, each bunch can mark by a secret key bits.For each user, by each selected features component, corresponding home cluster and secret key bits can be determined.By combining the secret key bits obtained from all features component, namely can be each user and generating key.
Such as, if the jth of a user i selected module diagnostic is CF ij.Suppose CF ijcomprise two features, and this features component there are seven bunches, as shown in Figure 1.Each bunch can from 000 to 111 numberings.In order to secret key bits be mapped in this features component, the sample first provided by user i calculates its selected statistical characteristics; Then, the value by calculating is determined home cluster and is determined secret key bits.In this way, we can determine the secret key bits of each selected features component, and in order to form key.
Such as:
The present invention adopts handwritten signature biological attribute data to verify the method proposed.Handwritten signature as biometric authentication a kind of mode accept by users and widely use, but there is an obvious shortcoming in it, namely there is larger difference at the handwriting of different time in same person.In such cases, how to determine that the feature that stability is high has larger challenge in order to generate stable key.
The handwritten signature data sample adopted in experiment, from 359 volunteers, by using the plotting sheet of traditional A4 size, being caught with the resolution of per inch 500 line each data sample that volunteer provides, being have collected 7430 handwritten signature data samples altogether.
In all volunteer's handwritten signatures, choose a part as experiment test sample data, these sample datas are divided into DB 1, DB 2and DB 3three data sets.First data set DB 1comprise 300 signatures from 30 users, other two data sets comprise 400 signatures of 40 users and 1000 signatures of 100 users respectively.For guaranteeing there are enough training datas, be selected into data set DB 1, DB 2and DB 3user everyone have 10 samples at least.
From handwriting samples extracting data 30 kinds of dynamic perfromances, comprise: sign T.T. used, lift a time, start to write the time, stroke number, the average velocity/acceleration of pen movement, in horizontal/vertical directions local minimum/maximal value, the min/max speed/acceleration of pen movement and time of origin in horizontal/vertical directions, and the T.T. etc. of cost is moved in direction up and down.Owing to employing character subset in the inventive method, any assemblage characteristic can be inferred from further feature (such as, the time and starting to write the time of lifting of given signature, then can derive represent lift pen and the feature of ratio between starting to write).The statistical characteristics calculated is by advanced column criterion process, and namely the mean value of each feature is 0, and standard deviation is 1.By these statistical natures, performance test will be carried out to said method below.
Performance test
False rejection rate (FalseRejectionRate, FRR) and false acceptance rate (FalseAcceptanceRate, FAR) are the indexs of the most general adopted measurement biometric identity Verification System performance.In the present invention, false rejection rate (FRR) refers to the probability of validated user authentication failed, and false acceptance rate (FAR) is although refer to that the signature provided not comes from legal family, still by the probability of checking.Except FRR and FAR, the length of key weighs another index of key safety.Key length and FRR and FAR have direct relation, as longer key can cause the low FAR of high FRR.Therefore, in testing, FRR with FAR is adopted to assess the performance of system during generation different effective key length.
Consider the key mapping mistake that may exist, be necessary to carry out error correction to secret key bits.In the methods of the invention, the difference between handwriting feature samples may cause some secret key bits to be mapped to the secret key bits of adjacent cluster.For this reason, introduce a strategy and correct this type of mistake, namely carry out substituting to corresponding secret key bits by the secret key bits of adjacent cluster and for checking.Such as, key=01 is supposed 100 2111 310 4001 5it is the primary key generated in five selected features component from user.If this key cannot by checking, another key for logon attempt then by replacing a secret key bits of primary key, such as second secret key bits (namely 00 2) secret key bits that replaces to adjacent cluster (is assumed to be 11 2), obtain key=01 111 2111 310 4001 5attempt checking.Suppose, this strategy is in order to correct r this type of mistake, then proving program will be attempted before returning the result refusing to log in individual key, ω is here the sum of optional features assembly.In embodiments of the present invention, with r=4 and r=5, performance test is carried out to proposed method.
First at data set DB 1and DB 3the method that application the present invention proposes is assessed FRR and FAR.Each experiment generates the key with b position significance bit.For each given b value, reduce the selection threshold value of each user gradually to generate the key about having b significance bit.By test individual significance bit, draws out the chart of FRR and FAR as a function of b.In an experiment, six samples of each user of random selecting are used for training, and other four samples are for calculating FRR.FAR then carries out Measurement and Computation by the signature that random selecting from other users 20 is different.At data set DB 1on test result (accompanying drawing 2) display, testing when b=20 and r=4 the system performance that obtains is FRR=14.5% and FAR=0%, and this result shows that the key with the inventive method generates has good reliability and security in practice.When r=5, adopt the party rule to have better performance: test result when b=20, FRR=13.1% and FAR=0%.As can be seen from accompanying drawing 3, at DB 3on carry out testing the experimental result that can obtain better performances equally.This shows that the performance of institute's extracting method in the present invention can not reduce because of increasing of customer volume.
What more than set forth is the specific embodiment that the present invention provides, and test result and performance evaluation embody technical scheme proposed by the invention and have good FRR and FAR performance for handwritten signature authentication, can carry out safe and reliable authentication.It is to be noted, the present invention is not only limited to above-described embodiment, for other biological feature, such as fingerprint, iris, face etc., also the technical scheme that the present invention proposes can be adopted, from the biological specimen feature gathered, extract statistical nature directly generate biological secret key, for the authentication of security system.

Claims (2)

1. based on hereditary automatic mold clustering analysis without template biological key generation method, comprise the following steps:
Step 1: obtain some biometric sample from user, calculate the statistical characteristics of each biometric sample, based on the statistical characteristics calculated, i.e. training data, application heredity automatically fuzzy clustering algorithm carries out fuzzy cluster analysis to data, by the difference between cluster result simulation colony inside and colony;
Step 2: the stability determining features component,
By fuzzy clustering result, the average blur degree of membership of each features component and each bunch can be calculated; The maximum membership degree value obtained is by the stability in order to weigh these features component;
Step 3: the selection of features component; Based on the features component stability result obtained in previous step, select the higher features component of stability in order to generate key;
Step 4: the generation of biological secret key;
After determining the features component of each user, key can generate thus; For the cluster result in each features component, each bunch can mark by a secret key bits; For each user, by each selected features component, corresponding home cluster and secret key bits can be determined; By combining the secret key bits obtained from all features component, namely can be each user and generating key.
2. as claimed in claim 1 based on hereditary automatic mold clustering analysis without template biological key generation method, it is characterized in that:
The automatic fuzzy clustering algorithm of heredity described in step 1 comprises:
Training dataset X={x is extracted from the biological specimen gathered 1, x 2..., x n, wherein, x ifor the eigenvector on d dimension space, n is the number of sample; K fuzzy clustering is comprised, if a kth fuzzy clustering center is C in data set X k, (k=1,2 ..., K), then cluster centre is C={C 1, C 2..., C k; Cluster result need meet with properties:
Σ i = 1 n u k i ≥ 1 , ( k = 1 , ... , K ) , Σ k = 1 K u k i ≥ 1 , ( i = 1 , ... , n ) , Σ k = 1 K Σ i = 1 n u k i = n ,
Wherein u kidata centralization i-th element x iabout a kth fuzzy clustering center C kdegree of membership; This clustering problem can be converted into cluster matrix U (X)=[u finding the given fuzzy clustering standard of optimization ki], 1≤k≤K, 1≤i≤n;
Above-mentioned training data is concentrated to each features component of sample data, the concrete steps of the automatic fuzzy clustering algorithm of the heredity described in step 1 are as follows:
1.1. produce the cluster result that P represents with bunch center at random, as initial population, each cluster result can comprise varying number bunch;
1.2. the suitability degree of each cluster result in Xie-Beni (XB) index calculate initial population is adopted;
Xie-Beni (XB) index is:
The suitability degree of each cluster result is defined as: f=1/XB;
Wherein, s=min i ≠ j{ D 2(z i-z j), D () is the measuring distance based on Euclidean tolerance, z krepresent kGe Cu center, m is weighted index;
1.3. repeat following a) to d) operation is until meet end condition, namely best in colony cluster result does not all change in n generation develops;
A) cluster result adopting tournament selection method to select two relative suitability degrees high from colony partners father and mother; Repeat this operation until select P/2 to father and mother;
B) on the father and mother selected, application interlace operation generates P offspring, then to each rear substitute performance mutation operation;
C) suitability degree of each offspring is calculated;
D) create colony of new generation, by the past generation colony and progeny population in select the cluster result of P high suitability degree, and replace last generation colony, get back to step a);
1.4. the best cluster result found in evolutionary process is exported.
CN201310375864.7A 2013-08-26 2013-08-26 Based on hereditary automatic mold clustering analysis without template biological key generation method Active CN103577738B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201310375864.7A CN103577738B (en) 2013-08-26 2013-08-26 Based on hereditary automatic mold clustering analysis without template biological key generation method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201310375864.7A CN103577738B (en) 2013-08-26 2013-08-26 Based on hereditary automatic mold clustering analysis without template biological key generation method

Publications (2)

Publication Number Publication Date
CN103577738A CN103577738A (en) 2014-02-12
CN103577738B true CN103577738B (en) 2016-04-13

Family

ID=50049502

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201310375864.7A Active CN103577738B (en) 2013-08-26 2013-08-26 Based on hereditary automatic mold clustering analysis without template biological key generation method

Country Status (1)

Country Link
CN (1) CN103577738B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101072100A (en) * 2006-05-12 2007-11-14 联想(北京)有限公司 Authenticating system and method utilizing reliable platform module
CN101295361A (en) * 2007-04-25 2008-10-29 中国科学院自动化研究所 Multi-biological characteristic authentication amalgamation method based on rejection region

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5343761B2 (en) * 2009-08-25 2013-11-13 株式会社デンソーウェーブ Optical information reader and authentication system using optical information reader

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101072100A (en) * 2006-05-12 2007-11-14 联想(北京)有限公司 Authenticating system and method utilizing reliable platform module
CN101295361A (en) * 2007-04-25 2008-10-29 中国科学院自动化研究所 Multi-biological characteristic authentication amalgamation method based on rejection region

Also Published As

Publication number Publication date
CN103577738A (en) 2014-02-12

Similar Documents

Publication Publication Date Title
Eskander et al. Hybrid writer‐independent–writer‐dependent offline signature verification system
TWI592820B (en) Man-machine recognition method and system
Yager et al. The biometric menagerie
CN105631272B (en) A kind of identity identifying method of multiple security
CN104820924B (en) A kind of online safety payment system based on handwriting verification
GB2437100A (en) Biometric security system using keystroke dynamics of a user's login attempt
CN116957049B (en) Unsupervised internal threat detection method based on countermeasure self-encoder
CN111062424A (en) Small sample food image recognition model training method and food image recognition method
CN112560710B (en) Method for constructing finger vein recognition system and finger vein recognition system
Saunders et al. Using automated comparisons to quantify handwriting individuality
Sae-Bae et al. Distinctiveness, complexity, and repeatability of online signature templates
Dass Fingerprint‐Based Recognition
WO2010028517A1 (en) System and method for generating/ identifying cipher code via artificial neural network
Yin et al. A feature selection method for improved clonal algorithm towards intrusion detection
CN102368291B (en) Personal authentication system based on invisible consciousness of fingerprint image
CN109918883B (en) Identity verification method of biological characteristics based on hand-drawn track
CN108830217B (en) Automatic signature distinguishing method based on fuzzy mean hash learning
CN103577738B (en) Based on hereditary automatic mold clustering analysis without template biological key generation method
Revett et al. On the use of rough sets for user authentication via keystroke dynamics
WO2015032303A1 (en) Character radical-based method for online handwriting authentication and template expansion
Wang et al. TempatMDS: a masquerade detection system based on temporal and spatial analysis of file access records
CN113434840B (en) Mobile phone continuous identity authentication method and device based on feature map
CN106682477B (en) Authentication method based on human body solid state characteristics
CN109086677A (en) A kind of auth method and system based on fingerprint recognition
JP4283864B1 (en) User authentication program and user authentication system

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant