CN102324031B - Latent semantic feature extraction method in aged user multi-biometric identity authentication - Google Patents
Latent semantic feature extraction method in aged user multi-biometric identity authentication Download PDFInfo
- Publication number
- CN102324031B CN102324031B CN 201110264487 CN201110264487A CN102324031B CN 102324031 B CN102324031 B CN 102324031B CN 201110264487 CN201110264487 CN 201110264487 CN 201110264487 A CN201110264487 A CN 201110264487A CN 102324031 B CN102324031 B CN 102324031B
- Authority
- CN
- China
- Prior art keywords
- image
- feature
- latent semantic
- extract
- semantic feature
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Collating Specific Patterns (AREA)
Abstract
The invention relates to a latent semantic feature extraction method in aged user multi-biometric identity authentication. Identity authentication is performed by performing multi-mode latent semantic analysis and data mining mapping on aged user multi-biometric images and extracting the latent semantic features of the images. According to the latent semantic feature extraction method in the aged user multi-biometric identity authentication, multiple local bottom features can be acquired by extracting the multi-biometric images of face, multiple fingerprints and palm prints and the like; the extracted features can be processed by using a multi-mode latent semantic analysis algorithm on three aspects of bottom feature-image matrix construction, two-dimensional matrix decomposition and clustering algorithm; the processed features are further mined and mapped through an 'intelligent black box model', so that the latent semantic features of the images can be effectively acquired; and the system is automatically adjusted by introducing an adaptive feedback structure with a genetic algorithm (GA), so that modification of the latent semantic features of the images is realized.
Description
Technical field
The present invention relates to technical field of biometric identification, especially the latent semantic feature extraction method in the aged user multi-biometric identity authentication.
Background technology
At present, China more than 60 years old population reach 1.8 hundred million people, account for total population 13.8%, weigh by international standard, China has entered the society of the aged, along with country's Speed-up Establishment and improve to cover the social security system of urban and rural residents energetically, granting such as the social old-age insurance gold, supplementary pension, medical insurance etc., elder user will become the main foreigner tourists of future society public service, the social old-age insurance gold, there is deception in the distribution process such as supplementary pension, the false claiming phenomenon becomes now social question of common concern, informationization, digitizing, network technology provides help for solving elder user authentication quagmire.At present, biometrics identification technology, long-distance video authentication have been successfully applied the identity of examining elder user in the social pension gold false claiming phenomenon.
Biometrics identification technology is by utilizing the intrinsic physiological characteristic of human body and behavior act to carry out identification and checking.According to the kind and the number that use biological characteristic, living things feature recognition can be divided into single living things feature recognition and multi-biological characteristic identification, as using the widest single bio-identification identity identifying technology, the problem of authentication had been subject to extensive concern when fingerprint recognition was put at solution elder user social pension golden hair.As far back as 1901, Britain began employing fingerprint and has identified to avoid the railway worker to falsely claim as one's own, lead more salary.At present, associated companies such as IBM, Microsoft, HP, Compaq, Changchun letter reach, are just waiting in the Hangzhou service field that entered society of the product of company." the payment pension fingerprint ID authentication system technical manual (trying) " of the social insurance career management center issue of China Ministry of Labour and Social Security also will be promulgated as the social public service standard based on the fingerprint identification method of minutiae point (minutiae), but, concerning elder user, owing to have experienced all sorts of hardships, fuzzy finger is very common, and traditional fingerprint recognition system based on minutiae point tends to cause system's misclassification rate to increase because the extraction minutiae point is undesirable even authentication was lost efficacy.In addition, recognition technology based on single biological characteristic exists not ubiquity: some biological characteristic disappearance (such as severed digit), damage (such as impaired finger), pathology (such as cataract) or collection apparatus second-rate (changing such as people's face light) all can cause robustness, the poor reliability of recognition system, a little less than the anti-duplicity, be difficult to satisfy the actual requirement of different occasions.
Image latent semantic feature (Image Latent Semantic Features, ILSF) obtained by low-level image feature-image array, has the information more abundanter than traditional image, semantic, but relative and low-level image feature, these features have stronger expression and classification capacity.Therefore, the feature of utilizing TLSA to extract can be used as the feature of a kind of " uniqueness ", and is proved to be and can be used in the biometric identity field of authentication.Simultaneously, compare traditional low-level image feature, owing to indirectly being used for Description Image, the image latent semantic feature is not very high for the quality requirements that gathers image, can better overcome the impact that some unfavorable factor is brought, fuzzy such as the image streakline of fingerprint, and the impact of human face light variation.
Summary of the invention
The technical problem to be solved in the present invention is: in order to overcome the problem of above-mentioned middle existence, latent semantic feature extraction method in a kind of aged user multi-biometric identity authentication is provided, utilizes image processing techniques and intellectual technology the user to be carried out the technology of authentication.
The technical solution adopted for the present invention to solve the technical problems is: the latent semantic feature extraction method in a kind of aged user multi-biometric identity authentication, by the aged user multi-biometric image being carried out multimode latent semantic analysis and data mining mapping, and extract the image latent semantic feature and carry out authentication, the concrete steps of its described multimode latent semantic analysis are as follows:
A. the image array of low-level image feature makes up: adopt multiple low-level image feature, make up the image array of each user's low-level image feature;
B. two-dimentional Algorithms of Non-Negative Matrix Factorization walks abreast: first the image array of low-level image feature carried out the diagonalization processing, again diagonalizable matrix is carried out the row matrix Directional Decomposition, and then former diagonalizable matrix is carried out transpose process obtain column direction information, the basis matrix that obtains is carried out basis matrix orthogonalization;
C. fuzzy C-means clustering: utilize the fuzzy C-means clustering method in the program-ming Toolbox to carry out cluster;
The concrete steps of described data mining mapping are as follows:
A. beta pruning Algorithm Analysis: the network structure predefine parameter of initialization fuzzy neural network (FNN) (such as convergence constant a, attenuation constant b, least error e, regular importance threshold value f), input the first stack features vector, produce article one fuzzy rule, any input feature value is calculated the distance of itself and the first stack features vector, draw minimum value d
MinThereby, calculate actual output error e
iIf, error e
iGreater than regular importance threshold value f, then produce new fuzzy rule, thereby adjust network architecture parameters (such as convergence constant a, attenuation constant b, least error e, regular importance threshold value f);
B. extract the image latent semantic feature: comprise off-line learning stage and on-line testing stage;
The concrete steps that described extraction image latent semantic feature is revised are as follows:
A. introduce self-adaptation dynamic feedback structure: in the intelligent blackbox model based on fuzzy neural network, utilize and extract extraction image latent semantic feature and the state parameter of identification with the self-adaptation dynamic feedback structure of GA optimized algorithm;
B. image latent semantic feature normalization: the state parameter by picking out intelligent blackbox model and compare with the normal condition that obtains by sample learning, draw difference as the input of intelligent blackbox model, thereby draw the image latent semantic feature deviation that causes because of environmental difference.
Described multi-biological characteristic image comprises people's face, refers to fingerprint and palmmprint more that the concrete steps of the multiple local low-level image feature of its extraction multi-biological characteristic image are as follows:
A. biometric image pre-service: the emerging system that is made of people's face, four finger fingerprints and palmmprint carries out pre-service;
B. extract low-level image feature: extract invariant moment features, Garbor filter feature, direction equalization feature and half-tone information entropy feature;
C. extract the local low-level image feature in the low-level image feature: successively by to the selection of reference point, based on the extraction of the ROI of reference point and the division of ROI, by extracting the local low-level image feature of image, the local bottom that extracts image is characterized as invariant moment features, Garbor filter feature, direction equalization feature and half-tone information entropy feature at last.
The described off-line learning stage is to be used for Training Fuzzy Neural Networks (FNN) and adopt the beta pruning algorithm that the network structure of fuzzy neural network (FNN) is done dynamic adjustment by learning sample.
The described on-line testing stage is to utilize the fuzzy neural network that trains that test sample book is tested, thereby it is semantic to extract image implicit expression.
The beneficial effect of the latent semantic feature extraction method in the aged user multi-biometric identity authentication of the present invention is: by extracting people's face, referring to the multi-biological characteristic images such as fingerprint and palmmprint more, can obtain multiple local low-level image feature; Utilize the multimode latent semantic analysis algorithm to decompose and the clustering algorithm three aspects: from low-level image feature-image array structure, two-dimensional matrix, can process the feature of extracting; Further the feature after processing is excavated mapping by " intelligent blackbox model ", can effectively obtain the image latent semantic feature; By introducing the self-adaptation feedback arrangement of band genetic algorithm (GA), system is adjusted automatically, realize the correction of image latent semantic feature, the method acquisition quality is good and reliability is strong, can satisfy the actual requirement of different occasions.
Description of drawings
The present invention is further described below in conjunction with drawings and Examples.
Fig. 1 is structural principle block diagram of the present invention;
Fig. 2 is the synoptic diagram of two-dimensional matrix diagonalization (a) row combination (b) row combination among Fig. 1;
Fig. 3 is FCM cluster synoptic diagram among Fig. 1;
Fig. 4 is self-adaptation dynamic feedback structured flowchart among Fig. 1;
Fig. 5 is the normalized structured flowchart of image latent semantic feature among Fig. 1.
Embodiment
In conjunction with the accompanying drawings, the present invention is further detailed explanation.These accompanying drawings are the synoptic diagram of simplification, basic structure of the present invention only is described in a schematic way, so it only show the formation relevant with the present invention.
Latent semantic feature extraction method in the aged user multi-biometric identity authentication as shown in Figure 1, by people's face, the emerging system that four finger fingerprints and palmmprint consist of carries out respectively pre-service, from pretreated biometric image, extract invariant moment features (comprising not bending moment and zernike bending moment not of hu), the Garborfilter feature, the image low-level image feature of direction equalization feature and half-tone information entropy feature, image low-level image feature after the extraction at first carries out multiple low-level image feature, make up the image array of each user's low-level image feature, secondly the image array of low-level image feature carried out the diagonalization processing, again diagonalizable matrix is carried out the row matrix Directional Decomposition, and then former diagonalizable matrix is carried out transpose process obtain column direction information, the basis matrix that obtains is carried out basis matrix orthogonalization, fuzzy C-means clustering method in the recycling program-ming Toolbox is carried out cluster, after data behind the process multimode latent semantic analysis are carried out the data mining mapping, in the intelligent blackbox model based on fuzzy neural network, utilize and extract semantic feature and the state parameter of identification with the self-adaptation feedback arrangement of GA optimized algorithm.
The concrete steps of the latent semantic feature extraction method in the aged user multi-biometric identity authentication of the present invention are as follows:
One, the extraction of low-level image feature: comprise the biometric image pre-service, extract low-level image feature and extract three key steps of local low-level image feature:
(1) biometric image pre-service: the pre-service of multi-biological characteristic image is one of committed step before the feature extraction, because emerging system is made of people's face, four finger fingerprints and palmmprint etc., therefore need to carry out respectively pre-service to it, pretreated key step comprises: area-of-interest (ROI) is cut apart, enhancing, normalization etc.At first, we will extract the ROI of image: 1. to people's face, mainly detect from video and cut apart facial image; 2. hand is gathered image, to cut apart first and refer to fingerprint and palmmprint more, by four fingers and palmmprint are located, the thinking (document sees reference) of reference Uhl etc., four finger fingerprints, palmmprint are split one by one, then, the ROI to images such as people's face of obtaining, fingerprint, palmmprints strengthens respectively and normalized on the basis of early-stage Study;
(2) extraction of low-level image feature: get indescribably invariant moment features (hu is bending moment and zernike orthogonally-persistent square not), Garbor filter feature, direction equalization feature, half-tone information entropy feature etc.
(a) invariant moment features: invariant moment features has rotation, yardstick and translation invariant feature, has the provincial characteristics ability of very strong Description Image.Not bending moment relatively commonly used as: hu is bending moment and zernike orthogonally-persistent square not, and its key step is as follows:
Step 1. is according to hu 7 hu eigenwert I of bending moment not of bending moment formulas Extraction not;
Step 2. is according to formula
Select zernike not exponent number n and the repeat number m of bending moment, extract A
NmValue;
I and the A of step 3. pair extraction
NmBe combined into invariant moment features.
(b) Garbor filter feature: Gabor filter has good set direction and frequency selective characteristic, can carry out time frequency analysis to image, extracts the texture value under different directions, the frequency, and its key step is as follows:
Step 1: according to formula
x'＝xsinθ+ycosθ，
Y'=xcos θ-ysin θ selects rational direction θ and frequency f parameter (to see that the applicant publishes thesis
[16]), extract the G value under different directions and the different frequency;
Step 2: the G value of extracting is combined into Gabor filter feature.
(c) direction equalization feature: the direction equalization has been described among the image direction figure central block and the relation of neighborhood piece on every side, the direction of neighborhood piece around if the direction of central block is more approaching, its value will be higher, therefore, can be used for the direction distribution situation of Description Image, its key step is as follows:
Step 1: the direction θ that calculates local streakline;
Step 2: according to formula
The calculated direction equilibrium value;
Step 3: the equilibrium value C that extracts is combined into direction equalization feature.
(d) half-tone information entropy feature: closely related average probabilistic the measuring that can be used as local gray level distribution in the image of information.If certain regional grey scale change is violent or frequent, illustrate that this regional picture material is complicated and relatively unstable, have higher uncertainty degree and higher half-tone information moisture in the soil value, on the contrary, lower uncertainty degree and half-tone information moisture in the soil value that the zone that picture material is unified and stable is corresponding, its key step is as follows:
Step 1: according to formula
The information entropy of computed image, wherein P
IjThe frequency that occurs in image in residing gray scale for gradation of image value f (i, j);
Step 2: the half-tone information entropy H that extracts is combined into half-tone information entropy feature.
(3) local bottom feature extraction: extract local feature and can accurately express the concrete information that image comprises, and can overcome the impact of noise and image deformation.By image being extracted area-of-interest (ROI) and divide, and the extraction of local feature is carried out in the zoning, its key step is as follows:
Step 1: the selection of reference point, the selection of general reference point is based on the central point with global characteristics of image, in order to extract accurately ROI and feature, needs to determine unique reference point.The central point that as fingerprint, refers to den and palm root distance according to central point, people's face based on the distance between two irises, palmmprint Main Basis;
Step 2: based on the extraction of the ROI of reference point.Determine ROI centered by the position of reference point and direction, every width of cloth image all can obtain the extraction that feature is convenient in unique candidate ROI zone;
The division of step 3:ROI.In order to overcome the factors such as non-linear deformation and noise to extracting the impact of feature, need to be with the ROI piecemeal.Can adopt dual mode: the method that 1. adopts rectangular node extracted region feature; 2. adopt annular partition to explain image-region, can overcome the shortcoming that the rectangular node field method needs accurate zoning, thereby reduce the time consumption of system, improve discrimination;
Step 4: extract local low-level image feature.Comprise invariant moment features (hu is bending moment and zernike orthogonally-persistent square not), Garbor filter feature, direction equalization feature, half-tone information entropy feature etc.
Two, multimode latent semantic analysis is processed:
(1) low-level image feature-image array makes up: adopt multiple low-level image feature, make up low-level image feature-image array (q biological characteristic merges altogether) of each user, its concrete steps are as follows:
Step 1: the ROI image unification of each biological characteristic is blocked into p size be the little image of n * n, q the individual local little image of the common p * q of biometric image;
Step 2: each local little image is comprised respectively the analyses such as invariant moment features (hu is bending moment and zernike orthogonally-persistent square not), Garbor filter feature, direction equalization feature, half-tone information entropy feature, and with the column vector of these features as low-level image feature-image array of each user;
Step 3: with the row vector of the local little image behind each piecemeal (q biological characteristic) as low-level image feature-image array of each user, add up each low-level image feature obtained in the previous step to the probability of its appearance, make up each user's low-level image feature and the feature-image array between the image, its size is p * q.
(2) parallel 2D-NNF algorithm: at first, low-level image feature-image array is carried out diagonalization to be processed, then, after diagonalizable matrix is carried out the row matrix Directional Decomposition, directly former diagonalizable matrix is carried out transpose process and obtain column direction information, at last, the basis matrix that obtains is carried out basis matrix orthogonalization: its key step as shown in Figure 2 is as follows:
Step 1 diagonalization of matrix: be m the low-level image feature-image array set I of p * q with size
P * q=[S
1, S
2..., S
m] represent S
nRepresent low-level image feature-image array of each user, m is user's quantity.1) if length p is not more than width q, so with the capable combination of matrix, and adopts the mode row combination image of Fig. 2 (a), and obtain diagonalizable matrix A from image array
n(zone shown in the shade).2) if length p greater than width q, is listed as combination with matrix so, and the mode row combination image of employing Fig. 2 (b), and the image array after combination obtains diagonalizable matrix A
n(zone shown in the shade).And the large young pathbreaker who obtains diagonalizable matrix is the same with original matrix, still is p * q;
Step 2: the image array line direction decomposes: be m the diagonalizable matrix set X of p * q with size
P * q=[A
1, A
2..., A
m] represent A
nRepresent the low-level image feature-image array after each user's diagonalization, m is user's quantity, and the matrix H that at first to utilize 1D-NMF to be decomposed into matrix L that size is p * d and size be d * q is long-pending, so that: X
P * q≈ L
P * dH
D * qHere d is with reference to dimension, and L is that matrix X decomposes the basis matrix that obtains in image row direction, and H is matrix of coefficients;
Step 3: the image array column direction decomposes: be m the diagonalizable matrix set Y of q * p with size
Q * p=[B
1, B
2..., B
m] represent, wherein
That former diagonalizable matrix is carried out transpose process.Similar above-mentioned algorithm utilizes 1D-NMF to find a size for nonnegative matrix R and the nonnegative matrix H that size is r * p of q * r, so that Y
Q * p≈ R
Q * rH
R * pHere, r is with reference to dimension, and R is the basis matrix that the decomposition of matrix Y on image column direction obtains, and H is matrix of coefficients;
Step 4: matrix X and Y consist of according to the former figure matrix of training sample and transposition figure matrix thereof, therefore can carry out simultaneously their decomposition.And, to any one user's diagonalization low-level image feature-image array A
n, its coefficient C on the row and column basis matrix
n=L
TA
nR, size is d * r, obviously, vectorial dimension reduces greatly.Utilize the basic L of row and Lie Ji R reconstruct low-level image feature-image array to be expressed as: A
n≈ LC
nR
T, n=1,2 ... m, then two-dimentional basis matrix is E=L R
T
Step 5: basis matrix orthogonalization: the basis matrix E=L R that parallel 2D-NMF method is obtained
TIn L and R matrix respectively orthogonalization: L '=orth (L) and R'=orth (R).The so two-dimentional basis matrix E ' after the orthogonalization=L ' (R ')
TConsisted of original image matrix A
nAn implicit expression semantic space, a semanteme in the corresponding subspace of every column vector.Project image onto in this semantic space, namely obtain the coefficient C by the semantic feature combination
nRepresented image latent semantic feature.
Three, data mining mapping: based on the design of FNN " intelligent blackbox model ", on the basis of early-stage Study, utilize the FNN construction data to excavate the model of mapping, wherein, obtaining of sample need to be by first three treatment step of Fig. 1, obtain the multimode latent semantic analysis feature, and with the multimode latent semantic analysis tagsort that obtains for study with train two groups of samples, at learning phase, by the beta pruning algorithm network structure of FNN is dynamically adjusted, made up rational FNN " intelligent blackbox model "; At test phase, carry out test sample (as shown in Figure 3 and Figure 4) with the network parameter that trains:
(1) step of beta pruning algorithm is as follows:
Step 1: the network structure predefine parameter of initialization FNN (such as convergence constant a, attenuation constant b, least error e, regular importance threshold value f), input the first stack features vector, produce the 1st fuzzy rule (if-then mode);
Step 2: any input feature value is calculated the distance of itself and the first stack features vector, seek minimum value d
Min, and calculate actual output error e
iIf, error e
iGreater than regular importance threshold value f, then produce new fuzzy rule (if-then mode), and calculating is based on the error rate of descent η of linear regression, otherwise, adjust network architecture parameters (such as convergence constant a, attenuation constant b, least error e, regular importance threshold value f) and jump to step 4;
Step 3: if rate of descent η less than least error e, then rejects the i rule, otherwise return step 2
Step 4: return many circulations of step 2, until all proper vector end of input.
(2) extract the image latent semantic feature: FNN is carried out off-line learning and two stages of on-line testing.At learning phase, a large amount of learning samples will be used for training FNN, and adopt the beta pruning algorithm that the network structure of FNN is done dynamic adjustment; At test phase, utilize the FNN that trains that test sample book is tested, extract the image latent semantic feature.
Four, image latent semantic feature correction (as shown in Figure 5)
(1) introduces self-adaptation dynamic feedback structure: in " intelligent blackbox model " based on FNN, in the sample learning stage, utilize and extract semantic feature and the state parameter of identification with the self-adaptation feedback arrangement of GA optimized algorithm.Its technical scheme mainly contains dual mode as shown in Figure 4: 1. FNN optimizes the mode of GA.On the early-stage Study basis, utilize the FL among the FNN dynamically to adjust crossover probability Pc and the variation probability P m parameter of GA and control evolutionary process, avoid precocious situation; 2. GA optimizes the mode of FNN, use for reference the thinking of Papadakis and He Suliang etc., utilize GA that the major parameter of the FL among the FNN and NN is adjusted respectively, wherein, the parameter of adjusting FL mainly comprises subordinate function and the fuzzy learning rule of fuzzy rule, and the major parameter of adjustment NN comprises Learning Step, network weight, hidden layer node numerical value etc.Continuous like this by feedback learning, after stability condition satisfies, will obtain the image latent semantic feature of identification, and write down the state parameter of the estimation of this moment;
(2) image latent semantic feature normalization: for eliminating the impact of external environment factor, obtain stable image latent semantic feature, this research is further revised and normalized the image latent semantic feature: namely decide the normal condition comparison according to the state parameter of " the intelligent blackbox model " that pick out with the institute that obtains by sample learning, its difference is as the input of model, try to achieve the image latent semantic feature deviation that causes because of environmental difference, thereby obtain normalized image latent semantic feature.
Take above-mentioned foundation desirable embodiment of the present invention as enlightenment, by above-mentioned description, the relevant staff can in the scope that does not depart from this invention technological thought, carry out various change and modification fully.The technical scope of this invention is not limited to the content on the instructions, must determine its technical scope according to the claim scope.
Claims (5)
1. the latent semantic feature extraction method in the aged user multi-biometric identity authentication, it is characterized in that: by the aged user multi-biometric image being carried out multimode latent semantic analysis and data mining mapping, and extract the image latent semantic feature and carry out authentication, the concrete steps of its described multimode latent semantic analysis are as follows:
A. the image array of low-level image feature makes up: adopt multiple low-level image feature, make up the image array of each user's low-level image feature;
B. two-dimentional Algorithms of Non-Negative Matrix Factorization walks abreast: first the image array of low-level image feature carried out the diagonalization processing, again diagonalizable matrix is carried out the row matrix Directional Decomposition, and then former diagonalizable matrix is carried out transpose process obtain column direction information, the basis matrix that obtains is carried out basis matrix orthogonalization;
C. fuzzy C-means clustering: utilize the fuzzy C-means clustering method in the program-ming Toolbox to carry out cluster;
The concrete steps of described data mining mapping are as follows:
A. beta pruning Algorithm Analysis: the network structure predefine parameter of initialization fuzzy neural network, input the first stack features vector, produce article one fuzzy rule, any input feature value is calculated the distance of itself and the first stack features vector, draw minimum value d
MinThereby, calculate actual output error e
iIf, error e
iGreater than regular importance threshold value f, then produce new fuzzy rule, thereby adjust network architecture parameters;
B. extract the image latent semantic feature: comprise off-line learning stage and on-line testing stage;
The concrete steps that described extraction image latent semantic feature is revised are as follows:
A. introduce self-adaptation dynamic feedback structure: in the intelligent blackbox model based on fuzzy neural network, utilize and extract extraction image latent semantic feature and the state parameter of identification with the self-adaptation dynamic feedback structure of GA optimized algorithm;
B. image latent semantic feature normalization: the state parameter by picking out intelligent blackbox model and compare with the normal condition that obtains by sample learning, draw difference as the input of intelligent blackbox model, thereby draw the image latent semantic feature deviation that causes because of environmental difference.
2. the latent semantic feature extraction method in the aged user multi-biometric identity authentication according to claim 1, it is characterized in that: described multi-biological characteristic image comprises people's face, refers to fingerprint and palmmprint more that the concrete steps of the multiple local low-level image feature of its extraction multi-biological characteristic image are as follows:
A. biometric image pre-service: the emerging system that is made of people's face, four finger fingerprints and palmmprint carries out pre-service;
B. extract low-level image feature: extract invariant moment features, Garbor filter feature, direction equalization feature and half-tone information entropy feature;
C. extract the local low-level image feature in the low-level image feature: successively by to the selection of reference point, based on the extraction of the ROI of reference point and the division of ROI, and extract image.
3. the latent semantic feature extraction method in the aged user multi-biometric identity authentication according to claim 1 is characterized in that: the concrete steps of the data behind the multimode latent semantic analysis being carried out the data mining mapping are as follows:
A. beta pruning Algorithm Analysis: the network structure predefine parameter of initialization fuzzy neural network, input the first stack features vector, produce article one fuzzy rule, any input feature value is calculated the distance of itself and the first stack features vector, draw minimum value d
MinThereby, calculate actual output error e
iIf, error e
iGreater than regular importance threshold value f, then produce new fuzzy rule, thereby adjust network architecture parameters;
B. extract the image latent semantic feature: comprise off-line learning stage and on-line testing stage.
4. the latent semantic feature extraction method in the aged user multi-biometric identity authentication according to claim 3 is characterized in that: the described off-line learning stage is to be used for Training Fuzzy Neural Networks and adopt the beta pruning algorithm that the network structure of fuzzy neural network is done dynamic adjustment by learning sample.
5. the latent semantic feature extraction method in the aged user multi-biometric identity authentication according to claim 3, it is characterized in that: the described on-line testing stage is to utilize the fuzzy neural network that trains that test sample book is tested, thereby it is semantic to extract image implicit expression.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110264487 CN102324031B (en) | 2011-09-07 | 2011-09-07 | Latent semantic feature extraction method in aged user multi-biometric identity authentication |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110264487 CN102324031B (en) | 2011-09-07 | 2011-09-07 | Latent semantic feature extraction method in aged user multi-biometric identity authentication |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102324031A CN102324031A (en) | 2012-01-18 |
| CN102324031B true CN102324031B (en) | 2013-04-24 |
Family
ID=45451770
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201110264487 Expired - Fee Related CN102324031B (en) | 2011-09-07 | 2011-09-07 | Latent semantic feature extraction method in aged user multi-biometric identity authentication |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102324031B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103886283A (en) * | 2014-03-03 | 2014-06-25 | 天津科技大学 | Method for fusing multi-biometric image information for mobile user and application thereof |
| CN103985113B (en) * | 2014-03-12 | 2017-09-29 | 浙江工商大学 | Tongue is as dividing method |
| CN104063721B (en) * | 2014-07-04 | 2017-06-16 | 中国科学院自动化研究所 | A kind of human behavior recognition methods learnt automatically based on semantic feature with screening |
| CN105975440A (en) * | 2016-05-05 | 2016-09-28 | 浙江理工大学 | Matrix decomposition parallelization method based on graph calculation model |
| CN106779090B (en) * | 2016-12-15 | 2019-03-08 | 南开大学 | A kind of feature learning model based on adaptive Dropout Non-negative Matrix Factorization |
| CN108427923B (en) * | 2018-03-08 | 2022-03-25 | 广东工业大学 | Palm print identification method and device |
| CN108629744B (en) * | 2018-04-09 | 2021-07-30 | 北京无线电计量测试研究所 | Image enhancement method |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100395770C (en) * | 2005-06-27 | 2008-06-18 | 北京交通大学 | Hand-characteristic mix-together identifying method based on characteristic relation measure |
| KR101185525B1 (en) * | 2005-07-29 | 2012-09-24 | 텔레콤 이탈리아 소시에떼 퍼 아찌오니 | Automatic biometric identification based on face recognition and support vector machines |
| CN101226653A (en) * | 2007-01-18 | 2008-07-23 | 中国科学院自动化研究所 | Rapid go-aboard system and method based on id card and biological characteristic recognition technique |
-
2011
- 2011-09-07 CN CN 201110264487 patent/CN102324031B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN102324031A (en) | 2012-01-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102324031B (en) | Latent semantic feature extraction method in aged user multi-biometric identity authentication | |
| CN101246543B (en) | Examiner identity identification method based on bionic and biological characteristic recognition | |
| Manjunatha et al. | Online signature verification based on writer dependent features and classifiers | |
| CN101763503B (en) | Face recognition method of attitude robust | |
| CN105760821B (en) | The face identification method of the grouped accumulation rarefaction representation based on nuclear space | |
| CN101840511B (en) | Method for extracting, matching and recognizing characteristics of finger veins | |
| CN101866421B (en) | Method for extracting characteristic of natural image based on dispersion-constrained non-negative sparse coding | |
| CN1322465C (en) | Image segmentation and fingerprint line distance getting technique in automatic fingerprint identification method | |
| CN102622610B (en) | Handwritten Uyghur character recognition method based on classifier integration | |
| CN101777117B (en) | Method for extracting finger vein feature for matching identification | |
| CN105844285A (en) | Cucumber disease identification method and apparatus based on image information | |
| CN101789075B (en) | Finger vein identifying method based on characteristic value normalization and bidirectional weighting | |
| CN102982322A (en) | Face recognition method based on PCA (principal component analysis) image reconstruction and LDA (linear discriminant analysis) | |
| CN103336960A (en) | Human face identification method based on manifold learning | |
| CN101604376A (en) | Face identification method based on the HMM-SVM mixture model | |
| CN101874738A (en) | Method for biophysical analysis and identification of human body based on pressure accumulated footprint image | |
| Casanova et al. | IFSC/USP at ImageCLEF 2012: Plant Identification Task. | |
| CN102938055A (en) | Hand bone identification system | |
| CN103164689A (en) | Face recognition method and face recognition system | |
| CN102339384B (en) | Face recognition method | |
| CN109508692A (en) | A kind of gender identification method and system based on 3D fingerprint image | |
| CN104361339B (en) | Slap shape Graph Extraction and recognition methods | |
| CN103942545A (en) | Method and device for identifying faces based on bidirectional compressed data space dimension reduction | |
| CN103942572A (en) | Method and device for extracting facial expression features based on bidirectional compressed data space dimension reduction | |
| CN114937298A (en) | Micro-expression recognition method based on feature decoupling |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130424 |