CN107330395A - A kind of iris image encryption method based on convolutional neural networks - Google Patents
A kind of iris image encryption method based on convolutional neural networks Download PDFInfo
- Publication number
- CN107330395A CN107330395A CN201710498746.3A CN201710498746A CN107330395A CN 107330395 A CN107330395 A CN 107330395A CN 201710498746 A CN201710498746 A CN 201710498746A CN 107330395 A CN107330395 A CN 107330395A
- Authority
- CN
- China
- Prior art keywords
- mrow
- msup
- iris
- image
- mfrac
- 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
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
- G06V40/18—Eye characteristics, e.g. of the iris
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/50—Maintenance of biometric data or enrolment thereof
- G06V40/53—Measures to keep reference information secret, e.g. cancellable biometrics
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Human Computer Interaction (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Biomedical Technology (AREA)
- Ophthalmology & Optometry (AREA)
- Image Analysis (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention discloses a kind of iris image encryption method based on convolutional neural networks, iris feature is proposed by CNN models, and then RS codes coding obtains encryption key, then carries out AES computings with image array corresponding grey scale value to be encrypted, that is, realizes ciphering process.Because iris sample is fewer in CNN training process, to ensure the high security of encryption, the generation of key must be using overall iris, it is impossible to extract iris progress image block to make up the problem of sample is not enough.The present invention carries out image block extraction to iris first, trains SAE models, and CNN parameters are initialized using SAE, and CNN model trainings are carried out using view picture iris image.It is so designed that and not only solves the problem of iris sample is few, and has guarantee to the security of image encryption.
Description
Technical field
The invention belongs to area of pattern recognition, and in particular to a kind of iris image encryption side based on convolutional neural networks
Method.
Background technology
Iris recognition technology is a kind of human biological's identification technology.Iris recognition technology is the knowledge in eyes based on iris
Not, the place applied to safety means (such as access control) and the confidentiality requirement of height.Human eyes structure is by sclera, rainbow
The positions such as film, pupil, retina are constituted.Iris is the annular section between black pupil and white sclera, and it dissipates comprising many
Layout, the details such as filament, coronal, striped, crypts.And the whole life process iris after prenatal development stage shaping will be kept
It is constant.These characteristics determine the uniqueness of iris feature, and further define the uniqueness of identification.Therefore, the iris of eyes
Characteristic can be applied to individual identity identification.
Iris is one of most robust feature in various biological identification technologies, is used as identity characteristic special with other biological
Levy and compare with more excellent property:Uniqueness, stability, collectable property, Noninvasive and security.In bio-identification
In, Noninvasive is the inexorable trend of identity research and application development, with other non contact angle measurement methods (such as facial harmony
Sound) to compare, iris recognition has higher accuracy.According to statistics, iris recognition is the minimum bio-identification of error rate, based on rainbow
The identity recognizing technology of film is gradually of interest by academia and business circles.
Iris such as can be widely applied for recognizing, encrypt at the numerous areas, have benefited from iris with some different from other biologies
The specific characteristic of tissue, mainly including the following aspects:
(1) stability:Among all one's life of people, iris information change is very small, only runs into some special diseases
Such as physical damnification, cataract, therefore iris has extremely strong stability;
(2) uniqueness:Iris is formed in embryonic development, and randomness is strong, is not influenceed by gene.Development of fetus experience one
Individual chaos morphogenetic process, this process causes the twinborn iris of DNA identicals also to differ, to unique performance of iris
Improve very big.So far, the iris for not occurring any two people is the same.As can be seen from Figure 2, iris is not smooth
Curved surface, surface distributed many depressions, spot and shrinks the details such as ditch, and these minutias determine iris information number
According to composition Texture eigenvalue, is the external embodiment of iris uniqueness.
(3) non-infringement:The non-property invaded is also referred to as untouchable, refers to connect when carrying out information gathering to iris
Touch with regard to readable information.Iris is located on rear side of cornea, and protected situation is fine.During being acquired to iris, it is not necessary to
Contact iris can be realized as iris capturing, it is therefore prevented that the harmful effect such as transmission, iris damage, user's acceptance compares
It is high.Relative to biological characteristics such as palmmprint, fingerprints, iris need not contact collection, and sense of discomfort is lower, be more beneficial for the general of application
Change.
(4) antifalsification:The change difficulty of iris feature is very big, is difficult to change in the case of to vision without major injury,
The duplication and conversion that the iris of one people is carried out with special object are nearly impossible, due to the hair of In vivo detection technology
Exhibition, carries out iris processing come counterfeit live body using irises such as image, video recordings and is all likely to detect.
Biological identification technology optimal selection is iris, and reason is that, relative to other biological feature, iris has highest unique
Property, accuracy of identification is high, speed is fast, bioactivity is strong, antifalsification is strong and has the advantages that unchangeable stability.
The variable item up to 260 of iris is constituted, internal structure is sufficiently complex, repeat possibility very little.Statistics shows, two
The same probability of individual human iris is almost nil, and the left and right eye iris of same person is also incomplete same.Iris is in development
It is constant to almost just stablizing when two or three years old, therefore the iris with labyrinth can be used as the unique mark of human body.
Iris encryption technology refers to use feature extraction algorithm, feature extraction is carried out to iris image and produces data flow, work
For the key of AES, it is decrypted during decryption using same key, so that it is whole with decryption to complete encryption
Process.Iris encryption technology is extracted using biological characteristic, and to produce unique key, key is unique, it is impossible to by other
Bion carries out imitated generation.Iris encryption technology is fusion living things feature recognition and encryption technology, and generation key is added
The new technology of close and decryption processing.
The content of the invention
Goal of the invention:For above-mentioned prior art, a kind of iris image encryption method based on convolutional neural networks is proposed,
Ensure encryption high security on the premise of, solve in CNN training process due to iris sample than it is few the problem of.
Technical scheme:A kind of iris image encryption method based on convolutional neural networks, ciphering process comprises the following steps:
Step 1, the iris image in iris database is pre-processed respectively, iris image collection X is drawn, to X
Image block extraction is carried out, iris image block data set X' is drawn;
Step 2, the iris image block data set X' is used to train four layers of SAE models, by study per layer parameter W=
{W(1),W(2),W(3), W(4), for initializing CNN model convolution layer parameter K={ K(1),K(2)And full connection layer parameter W '=
{W′(3),W′(4), CNN model trainings are then carried out using the iris image collection X;
Step 3, encryption side's collection iris image, inputs the CNN models trained after pretreatment, realizes to characteristic vector V1
Extraction, characteristic vector V1Dimension according to use resume image be adjusted;
Step 4, using RS codes to characteristic vector V1Encoded, draw encryption key Vk1With RS error correcting codes;
Step 5, encryption key V is utilizedk1AES computings are carried out with image array corresponding pixel points gray value to be encrypted, are drawn
Encrypted image, then complete whole ciphering process;
Decrypting process comprises the following steps:
Step 1, iris image acquiring is carried out to decryption side, the CNN models trained is input to after pretreatment, realized and extract
Iris feature vector V2;
Step 2, using RS error correcting codes to characteristic vector V2Error correction is carried out, decruption key V is obtainedk2;
Step 3, decruption key V is utilizedk2AES inverse operations are carried out with the encrypted image matrix corresponding pixel points gray value,
Decrypted image is drawn, then completes whole decrypting process.
Further, in the step 2, in CNN models, interval restrictive condition is increased initiation parameter:SAE is instructed
The parameter W practised is limited inInterval is interior, then convolution kernel K={ K in CNN(1),K(2)And connect entirely
Meet a layer weights W '={ W '(3),W′(4)Computational methods be:
nkRepresent kth layer SAE nodes, nk+1Represent the node layer number of kth+1.
Further, iris preprocessing is comprised the following steps:
(1) iris image I0Rim detection carries out Iris Location, realizes iris segmentation I1;
(2) the iris image I extracted1I is normalized2;
(3) by I2Carry out ZCA whitening processings I3。
Beneficial effect:Iris image AES of the invention based on convolutional neural networks, first the iris figure to collection
As data are pre-processed, characteristic vector pickup is then carried out to iris image using the CNN models trained.The feature of extraction
Vector is used for the generation of encryption key, finally uses encryption key with original image array corresponding pixel points gray value to be encrypted
AES computings, then can obtain encrypted image.
, it is necessary to first carry out the training of deep learning model before whole encryption and decrypting process, deep learning model is used
CNN, because iris image acquiring difficulty causes the problem of sample is less greatly, it is necessary to carry out CNN parameter initializations.In CNN training
During due to iris sample it is fewer, to ensure the high security of encryption, the generation of key must be using overall iris, it is impossible to
Iris is subjected to the problem of image block is extracted to make up sample deficiency.Carried in view of the above-mentioned problems, carrying out image block to iris first
Take, train SAE models, CNN parameters are initialized using SAE, CNN model trainings are carried out using view picture iris image.It is so designed that
The problem of iris sample is few is not only solved, and has guarantee to the security of image encryption.
Brief description of the drawings
Iris image encryptions of the Fig. 1 based on convolutional Neural net network and decryption flow;
Fig. 2 CASIA iris database image patterns;
Fig. 3 iris images are standardized;
Fig. 4 iris image ZCA albefactions;
Fig. 5 iris image RS codes are encoded;
Fig. 6 is trained and test set accuracy.
Embodiment
The present invention is done below in conjunction with the accompanying drawings and further explained.
A kind of iris image encryption method based on convolutional neural networks, ciphering process comprises the following steps:
Step 1, the iris image in iris database is pre-processed respectively, iris image collection X is drawn, to X
Image block extraction is carried out, iris image block data set X' is drawn.Iris preprocessing is comprised the following steps:
(1) iris image rim detection carries out Iris Location, realizes iris segmentation;
(2) iris image extracted is normalized;
(3) image after normalized is subjected to ZCA whitening processings.
Step 2, iris image block data set X' is used to train four layers of SAE models (Stacked Autoencoder, stack
Own coding model), the parameter matrix W={ W trained(1),W(2),W(3), W(4), for initializing CNN models
(Convolutional Neural Network, convolutional neural networks) convolution layer parameter K={ K(1),K(2)And full articulamentum ginseng
Number W '={ W '(3),W′(4), CNN model trainings are then carried out using iris image collection X.
In CNN models, if the no restrictive condition of initiation parameter can cause gradient disappear or amplitude diffusion show
As, therefore increase initiation parameter interval restrictive condition.CNN convolutional layer propagated forward linearization calculations are:
In formula, z represents that Internet node is exported, ωiRepresent ith feature figure correspondence weight matrix, xiRepresent i-th of spy
Figure is levied, m represents characteristic pattern number.Assuming that it is that 0 standard deviation is being uniformly distributed for σ to meet average, then every layer of variance formula is:
In formula, n represents the nodal point number of each layer of network, nk-1The node layer number of kth -1 is represented,Represent -1 layer of input of kth
Expectation,The expectation of -1 layer of weight matrix of kth is represented,The expectation of -2 layers of input of kth is represented,Represent the 1st layer of input
Expectation,Represent the expectation of i-th layer of weight matrix, niThe i-th node layer number is represented, k represents kth layer network structure sheaf.
It can be seen that from formula, with the increase of the neutral net number of plies, ifThen variance can be increasing, production
Raw parameter diffusion;IfThen variance can be less and less, produces gradient and disappears.Therefore, it is also desirable toNowIt can then draw
In back-propagation process, have:
It can similarly draw:
In formula, Loss represents object function,Kth i-th of input feature vector figure of layer is represented,Represent the 1st layer of jth input
Characteristic pattern,J-th of input feature vector figure correspondence weight matrix of kth+1 layer is represented,Represent+1 layer of weight matrix correspondence of kth
Expect,Represent+1 layer of ith feature figure of kth.
In order to prevent parameter is overflowed from being disappeared with gradient, it is necessary to meet:
Comparing can find that variance now generates contradiction, in order to solve this case, calculate variance and use formula:
For Uniform-distributed Data x ∈ [- a, a], a represents data x modulus value, and variance formula is:
By can be calculated:
Therefore, the parameter that SAE is trained is limited inIn interval, nkRepresent kth layer
SAE nodes, nk+1The node layer number of kth+1 is represented, so as to avoid the generation of above mentioned problem.
By the parameter matrix W={ W trained(1),W(2),W(3),W(4)Interval normalization is carried out, it is used as convolution kernel K in CNN
={ K(1),K(2)And full articulamentum weights W '={ W '(3),W′(4), computational methods are:
nkRepresent kth layer SAE nodes, nk+1Represent the node layer number of kth+1.
Step 3, encryption side's collection iris image, inputs the CNN models trained after pretreatment, realizes to characteristic vector V1
Extraction, characteristic vector V1Dimension according to use resume image be adjusted.
Step 4, using RS codes (Reed-solomon codes, RS code) to characteristic vector V1Encoded, draw encryption
Key Vk1With RS error correcting codes;
Step 5, encryption key V is utilizedk1AES encryption computing is carried out with image array corresponding pixel points gray value to be encrypted
(Advanced Encryption Standard, Advanced Encryption Standard), draws encrypted image, encrypted image is ciphertext, then complete
Into whole ciphering process.
Decryption is that, in order to reduce original image information, decrypting process is the inverse operation of ciphering process in general, but this
The algorithm that invention is proposed not is complete inverse transformation.Decrypting process comprises the following steps:
Step 1, iris image acquiring is carried out to decryption side, the CNN models trained is input to after pretreatment, realized and extract
Iris feature vector V2。
Step 2, due to V2And V1The difference of numerical value in some dimensions is there may exist, using RS error correcting codes to characteristic vector
V2Error correction is carried out, decruption key V is obtainedk2。
Step 3, decruption key V is utilizedk2AES inverse operations are carried out with encrypted image matrix corresponding pixel points gray value, are drawn
Decrypted image, then complete whole decrypting process.
In order to improve the confidence level and predictability of algorithm, experiment iris data collection uses the public version of CASIA iris databases,
Data set is divided into training set and test set, original iris image sample is as shown in Figure 2.The iris image of acquired original includes
The unrelated interruptions factor such as face, eyelashes, carries out Iris Location, segmentation, normalization and ZCA whitening pretreatments, iris to it first
Image preprocessing process is as shown in Figure 1.
Pretreated iris image, eliminates the disturbing factors such as part face, eyelashes, eyelid, only includes iris portion,
Feature identification degree is improved in order to reduce image redundancy, using to iris image ZCA whitening processings, such as Fig. 4.Then RS is carried out
Code coding, the purpose that RS codes coding is carried out to iris image is when preventing from carrying out characteristic vector pickup in decrypting process, to extract
Iris feature vector with encrypting when the iris feature vector that extracts it is inconsistent, be so conducive to encryption key and decruption key
Successful match, improves the probability of successful decryption, reduces operation complexity.
CNN is used to the deep learning model that iris carries out feature learning, data set includes iris image 400 altogether, altogether
10 classes, each 40 images of class, wherein 300 as training sample, 100 are used as test sample.SAE structures are using five layers of net
Network, CNN structures are using five layer networks layer, and convolution kernel size is 5 × 5, and down-sampled step-length is 2 × 2, the first convolutional layer generation feature
Figure number is 6, and the second convolutional layer generation characteristic pattern number is 12.Feature extraction result is shown and setting characteristic vector contrast, net
After network training, training accuracy and testing precision are as shown in Figure 6.
During encryption, after completing to CNN training, an eye image is gathered, the life of key is encrypted
Into encryption key and image to be encrypted are carried out into AES computings, ciphering process is realized.Complete above based on convolutional neural networks
Iris image ciphering process.
Described above is only the preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, under the premise without departing from the principles of the invention, some improvements and modifications can also be made, these improvements and modifications also should
It is considered as protection scope of the present invention.
Claims (3)
1. a kind of iris image encryption method based on convolutional neural networks, it is characterised in that:Ciphering process comprises the following steps:
Step 1, the iris image in iris database is pre-processed respectively, draws iris image collection X, X is carried out
Image block is extracted, and draws iris image block data set X';
Step 2, the iris image block data set X' is used to train four layers of SAE models, by study per layer parameter W={ W(1),W(2),W(3), W(4), for initializing CNN model convolution layer parameter K={ K(1),K(2)And full connection layer parameter W '={ W '(3),W
′(4), CNN model trainings are then carried out using the iris image collection X;
Step 3, encryption side's collection iris image, inputs the CNN models trained after pretreatment, realizes to characteristic vector V1Carry
Take, characteristic vector V1Dimension according to use resume image be adjusted;
Step 4, using RS codes to characteristic vector V1Encoded, draw encryption key Vk1With RS error correcting codes;
Step 5, encryption key V is utilizedk1AES computings are carried out with image array corresponding pixel points gray value to be encrypted, encryption is drawn
Image, then complete whole ciphering process;
Decrypting process comprises the following steps:
Step 1, iris image acquiring is carried out to decryption side, the CNN models trained is input to after pretreatment, realized and extract iris
Characteristic vector V2;
Step 2, using RS error correcting codes to characteristic vector V2Error correction is carried out, decruption key V is obtainedk2;
Step 3, decruption key V is utilizedk2AES inverse operations are carried out with the encrypted image matrix corresponding pixel points gray value, are drawn
Decrypted image, then complete whole decrypting process.
2. the iris image encryption method according to claim 1 based on convolutional neural networks, it is characterised in that:The step
In rapid 2, in CNN models, increase initiation parameter interval restrictive condition:The SAE parameter W trained are limited inInterval is interior, then convolution kernel K={ K in CNN(1),K(2)And full articulamentum weights W '={ W '(3),
W′(4)Computational methods be:
<mrow>
<msup>
<mi>K</mi>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</msup>
<mo>=</mo>
<mfrac>
<msup>
<mi>W</mi>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
<mrow>
<mo>(</mo>
<msup>
<mi>W</mi>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</msup>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>&times;</mo>
<msqrt>
<mfrac>
<mn>6</mn>
<mrow>
<msup>
<mi>n</mi>
<mn>1</mn>
</msup>
<mo>+</mo>
<msup>
<mi>n</mi>
<mn>2</mn>
</msup>
</mrow>
</mfrac>
</msqrt>
</mrow>
<mrow>
<msup>
<mi>K</mi>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</msup>
<mo>=</mo>
<mfrac>
<msup>
<mi>W</mi>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
<mrow>
<mo>(</mo>
<msup>
<mi>W</mi>
<mrow>
<mo>(</mo>
<mn>2</mn>
<mo>)</mo>
</mrow>
</msup>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>&times;</mo>
<msqrt>
<mfrac>
<mn>6</mn>
<mrow>
<msup>
<mi>n</mi>
<mn>2</mn>
</msup>
<mo>+</mo>
<msup>
<mi>n</mi>
<mn>3</mn>
</msup>
</mrow>
</mfrac>
</msqrt>
</mrow>
<mrow>
<msup>
<mi>W</mi>
<mrow>
<mo>&prime;</mo>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
</mrow>
</msup>
<mo>=</mo>
<mfrac>
<msup>
<mi>W</mi>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
<mrow>
<mo>(</mo>
<msup>
<mi>W</mi>
<mrow>
<mo>(</mo>
<mn>3</mn>
<mo>)</mo>
</mrow>
</msup>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>&times;</mo>
<msqrt>
<mfrac>
<mn>6</mn>
<mrow>
<msup>
<mi>n</mi>
<mn>3</mn>
</msup>
<mo>+</mo>
<msup>
<mi>n</mi>
<mn>4</mn>
</msup>
</mrow>
</mfrac>
</msqrt>
</mrow>
<mrow>
<msup>
<mi>W</mi>
<mrow>
<mo>&prime;</mo>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</mrow>
</msup>
<mo>=</mo>
<mfrac>
<msup>
<mi>W</mi>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</msup>
<mrow>
<mi>m</mi>
<mi>a</mi>
<mi>x</mi>
<mrow>
<mo>(</mo>
<msup>
<mi>W</mi>
<mrow>
<mo>(</mo>
<mn>4</mn>
<mo>)</mo>
</mrow>
</msup>
<mo>)</mo>
</mrow>
</mrow>
</mfrac>
<mo>&times;</mo>
<msqrt>
<mfrac>
<mn>6</mn>
<mrow>
<msup>
<mi>n</mi>
<mn>4</mn>
</msup>
<mo>+</mo>
<msup>
<mi>n</mi>
<mn>5</mn>
</msup>
</mrow>
</mfrac>
</msqrt>
</mrow>
nkRepresent kth layer SAE nodes, nk+1Represent the node layer number of kth+1.
3. the iris image encryption method according to claim 1 or 2 based on convolutional neural networks, it is characterised in that:It is right
Iris preprocessing comprises the following steps:
(1) iris image I0Rim detection carries out Iris Location, realizes iris segmentation I1;
(2) the iris image I extracted1I is normalized2;
(3) by I2Carry out ZCA whitening processings I3。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710498746.3A CN107330395B (en) | 2017-06-27 | 2017-06-27 | A kind of iris image encryption method based on convolutional neural networks |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710498746.3A CN107330395B (en) | 2017-06-27 | 2017-06-27 | A kind of iris image encryption method based on convolutional neural networks |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107330395A true CN107330395A (en) | 2017-11-07 |
CN107330395B CN107330395B (en) | 2018-11-09 |
Family
ID=60197495
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710498746.3A Active CN107330395B (en) | 2017-06-27 | 2017-06-27 | A kind of iris image encryption method based on convolutional neural networks |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107330395B (en) |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108011712A (en) * | 2017-11-13 | 2018-05-08 | 佛山科学技术学院 | A kind of Mobile medical system private data communication means |
CN109389094A (en) * | 2018-10-23 | 2019-02-26 | 北京无线电计量测试研究所 | A kind of stable iris feature extraction and matching process |
CN109995520A (en) * | 2019-03-06 | 2019-07-09 | 西北大学 | Cipher key transmission methods, image processing platform based on depth convolutional neural networks |
CN110059589A (en) * | 2019-03-21 | 2019-07-26 | 昆山杜克大学 | The dividing method of iris region in a kind of iris image based on Mask R-CNN neural network |
CN110211016A (en) * | 2018-02-28 | 2019-09-06 | 佛山科学技术学院 | A kind of watermark embedding method based on convolution feature |
CN110378138A (en) * | 2019-07-22 | 2019-10-25 | 上海鹰瞳医疗科技有限公司 | Data encryption, decryption method and neural network training method and equipment |
CN110427804A (en) * | 2019-06-18 | 2019-11-08 | 中山大学 | A kind of iris auth method based on secondary migration study |
CN110505047A (en) * | 2019-08-14 | 2019-11-26 | 江苏海洋大学 | A kind of double-encryption algorithm for iris feature protection |
CN110659452A (en) * | 2018-06-29 | 2020-01-07 | 国际商业机器公司 | Artificial intelligence supported cyber data hiding and directed task execution |
WO2020011016A1 (en) * | 2018-07-09 | 2020-01-16 | 奇酷互联网络科技(深圳)有限公司 | Iris information leakage prevention method and apparatus, and electronic device |
CN111382455A (en) * | 2020-03-18 | 2020-07-07 | 北京丁牛科技有限公司 | File protection method and device |
CN111401145A (en) * | 2020-02-26 | 2020-07-10 | 三峡大学 | Visible light iris recognition method based on deep learning and DS evidence theory |
CN111464304A (en) * | 2019-01-18 | 2020-07-28 | 江苏实达迪美数据处理有限公司 | Hybrid encryption method and system for controlling system network security |
CN111538969A (en) * | 2020-03-30 | 2020-08-14 | 北京万里红科技股份有限公司 | Document encryption method, document decryption device, electronic equipment and medium |
CN111639351A (en) * | 2020-05-20 | 2020-09-08 | 燕山大学 | Battery tracing management coding encryption and decryption method based on self-encoder and Henon mapping |
CN111654368A (en) * | 2020-06-03 | 2020-09-11 | 电子科技大学 | Key generation method for generating countermeasure network based on deep learning |
CN112037211A (en) * | 2020-09-04 | 2020-12-04 | 中国空气动力研究与发展中心超高速空气动力研究所 | Damage characteristic identification method for dynamically monitoring small space debris impact event |
CN112395635A (en) * | 2021-01-18 | 2021-02-23 | 北京灵汐科技有限公司 | Image processing method, device, secret key generating method, device, training method and device, and computer readable medium |
CN112395636A (en) * | 2021-01-19 | 2021-02-23 | 国网江西省电力有限公司信息通信分公司 | Power grid data encryption model training method, system, storage medium and equipment |
CN112668472A (en) * | 2020-12-28 | 2021-04-16 | 中国科学院自动化研究所 | Iris image feature extraction method, system and device based on federal learning |
CN112749663A (en) * | 2021-01-15 | 2021-05-04 | 金陵科技学院 | Agricultural fruit maturity detection system based on Internet of things and CCNN model |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2006025129A1 (en) * | 2004-08-30 | 2006-03-09 | Toyama-Prefecture | Personal authentication system |
CN1996378A (en) * | 2005-12-31 | 2007-07-11 | 北京华旗数码影像技术研究院有限责任公司 | Embedded device for detecting iris watermark information |
CN103917727A (en) * | 2011-11-08 | 2014-07-09 | 虹膜技术公司 | Locking apparatus with enhanced security using iris image |
CN104239815A (en) * | 2014-09-19 | 2014-12-24 | 西安凯虹电子科技有限公司 | Electronic document encryption and decryption method and method based on iris identification |
CN104992100A (en) * | 2015-07-15 | 2015-10-21 | 西安凯虹电子科技有限公司 | Iris dynamic encryption and decryption system and method for electronic document flowing |
US20150312034A1 (en) * | 2014-04-29 | 2015-10-29 | Altek Corporation | Method for image encryption and decryption incorporating physiological features and image capture device thereof |
CN105354501A (en) * | 2015-10-28 | 2016-02-24 | 广东欧珀移动通信有限公司 | Photo processing method and processing system |
CN106095144A (en) * | 2016-07-29 | 2016-11-09 | 石家庄蜗牛科技有限公司 | The mouse of a kind of multi-enciphering and authentication method thereof |
CN106096526A (en) * | 2016-06-06 | 2016-11-09 | 联想(北京)有限公司 | A kind of iris identification method and iris authentication system |
CN106559424A (en) * | 2016-11-16 | 2017-04-05 | 杭州释码云信息技术有限公司 | A kind of iris image encryption method, equipment and system |
-
2017
- 2017-06-27 CN CN201710498746.3A patent/CN107330395B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2006025129A1 (en) * | 2004-08-30 | 2006-03-09 | Toyama-Prefecture | Personal authentication system |
CN1996378A (en) * | 2005-12-31 | 2007-07-11 | 北京华旗数码影像技术研究院有限责任公司 | Embedded device for detecting iris watermark information |
CN103917727A (en) * | 2011-11-08 | 2014-07-09 | 虹膜技术公司 | Locking apparatus with enhanced security using iris image |
US20150312034A1 (en) * | 2014-04-29 | 2015-10-29 | Altek Corporation | Method for image encryption and decryption incorporating physiological features and image capture device thereof |
CN104239815A (en) * | 2014-09-19 | 2014-12-24 | 西安凯虹电子科技有限公司 | Electronic document encryption and decryption method and method based on iris identification |
CN104992100A (en) * | 2015-07-15 | 2015-10-21 | 西安凯虹电子科技有限公司 | Iris dynamic encryption and decryption system and method for electronic document flowing |
CN105354501A (en) * | 2015-10-28 | 2016-02-24 | 广东欧珀移动通信有限公司 | Photo processing method and processing system |
CN106096526A (en) * | 2016-06-06 | 2016-11-09 | 联想(北京)有限公司 | A kind of iris identification method and iris authentication system |
CN106095144A (en) * | 2016-07-29 | 2016-11-09 | 石家庄蜗牛科技有限公司 | The mouse of a kind of multi-enciphering and authentication method thereof |
CN106559424A (en) * | 2016-11-16 | 2017-04-05 | 杭州释码云信息技术有限公司 | A kind of iris image encryption method, equipment and system |
Non-Patent Citations (3)
Title |
---|
KHALED LOUKHAOUKHA 等: "An efficient image encryption algorithm based on blocks permutation and Rubik"s cube principle for iris images", 《2013 8TH INTERNATIONAL ON SYSTEM,SIGNAL PROCESSING AND THEIR APPLICATION》 * |
柴晓冬: "基于虹膜生物特征信息的图像加密方法", 《计算机应用》 * |
解瑞云 等: "基于虹膜特征的密钥生成和AES算法的图像加密", 《河南师范大学学报(自然科学版)》 * |
Cited By (31)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108011712A (en) * | 2017-11-13 | 2018-05-08 | 佛山科学技术学院 | A kind of Mobile medical system private data communication means |
CN110211016A (en) * | 2018-02-28 | 2019-09-06 | 佛山科学技术学院 | A kind of watermark embedding method based on convolution feature |
CN110659452B (en) * | 2018-06-29 | 2023-04-21 | 国际商业机器公司 | Method for hiding and orienting task execution of cyber data supported by artificial intelligence |
CN110659452A (en) * | 2018-06-29 | 2020-01-07 | 国际商业机器公司 | Artificial intelligence supported cyber data hiding and directed task execution |
WO2020011016A1 (en) * | 2018-07-09 | 2020-01-16 | 奇酷互联网络科技(深圳)有限公司 | Iris information leakage prevention method and apparatus, and electronic device |
CN109389094A (en) * | 2018-10-23 | 2019-02-26 | 北京无线电计量测试研究所 | A kind of stable iris feature extraction and matching process |
CN109389094B (en) * | 2018-10-23 | 2021-04-16 | 北京无线电计量测试研究所 | Stable iris feature extraction and matching method |
CN111464304B (en) * | 2019-01-18 | 2021-04-20 | 江苏实达迪美数据处理有限公司 | Hybrid encryption method and system for controlling system network security |
CN111464304A (en) * | 2019-01-18 | 2020-07-28 | 江苏实达迪美数据处理有限公司 | Hybrid encryption method and system for controlling system network security |
CN109995520A (en) * | 2019-03-06 | 2019-07-09 | 西北大学 | Cipher key transmission methods, image processing platform based on depth convolutional neural networks |
CN110059589A (en) * | 2019-03-21 | 2019-07-26 | 昆山杜克大学 | The dividing method of iris region in a kind of iris image based on Mask R-CNN neural network |
CN110427804B (en) * | 2019-06-18 | 2022-12-09 | 中山大学 | Iris identity verification method based on secondary transfer learning |
CN110427804A (en) * | 2019-06-18 | 2019-11-08 | 中山大学 | A kind of iris auth method based on secondary migration study |
CN110378138A (en) * | 2019-07-22 | 2019-10-25 | 上海鹰瞳医疗科技有限公司 | Data encryption, decryption method and neural network training method and equipment |
CN110505047A (en) * | 2019-08-14 | 2019-11-26 | 江苏海洋大学 | A kind of double-encryption algorithm for iris feature protection |
CN111401145A (en) * | 2020-02-26 | 2020-07-10 | 三峡大学 | Visible light iris recognition method based on deep learning and DS evidence theory |
CN111401145B (en) * | 2020-02-26 | 2022-05-03 | 三峡大学 | Visible light iris recognition method based on deep learning and DS evidence theory |
CN111382455A (en) * | 2020-03-18 | 2020-07-07 | 北京丁牛科技有限公司 | File protection method and device |
CN111538969A (en) * | 2020-03-30 | 2020-08-14 | 北京万里红科技股份有限公司 | Document encryption method, document decryption device, electronic equipment and medium |
CN111639351B (en) * | 2020-05-20 | 2022-03-15 | 燕山大学 | Battery tracing management coding encryption and decryption method based on self-encoder and Henon mapping |
CN111639351A (en) * | 2020-05-20 | 2020-09-08 | 燕山大学 | Battery tracing management coding encryption and decryption method based on self-encoder and Henon mapping |
CN111654368A (en) * | 2020-06-03 | 2020-09-11 | 电子科技大学 | Key generation method for generating countermeasure network based on deep learning |
CN112037211A (en) * | 2020-09-04 | 2020-12-04 | 中国空气动力研究与发展中心超高速空气动力研究所 | Damage characteristic identification method for dynamically monitoring small space debris impact event |
CN112037211B (en) * | 2020-09-04 | 2022-03-25 | 中国空气动力研究与发展中心超高速空气动力研究所 | Damage characteristic identification method for dynamically monitoring small space debris impact event |
CN112668472A (en) * | 2020-12-28 | 2021-04-16 | 中国科学院自动化研究所 | Iris image feature extraction method, system and device based on federal learning |
CN112749663A (en) * | 2021-01-15 | 2021-05-04 | 金陵科技学院 | Agricultural fruit maturity detection system based on Internet of things and CCNN model |
CN112749663B (en) * | 2021-01-15 | 2023-07-07 | 金陵科技学院 | Agricultural fruit maturity detection system based on Internet of things and CCNN model |
CN112395635B (en) * | 2021-01-18 | 2021-05-04 | 北京灵汐科技有限公司 | Image processing method, device, secret key generating method, device, training method and device, and computer readable medium |
CN112395635A (en) * | 2021-01-18 | 2021-02-23 | 北京灵汐科技有限公司 | Image processing method, device, secret key generating method, device, training method and device, and computer readable medium |
CN112395636B (en) * | 2021-01-19 | 2021-07-30 | 国网江西省电力有限公司信息通信分公司 | Power grid data encryption model training method, system, storage medium and equipment |
CN112395636A (en) * | 2021-01-19 | 2021-02-23 | 国网江西省电力有限公司信息通信分公司 | Power grid data encryption model training method, system, storage medium and equipment |
Also Published As
Publication number | Publication date |
---|---|
CN107330395B (en) | 2018-11-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107330395B (en) | A kind of iris image encryption method based on convolutional neural networks | |
Yuan et al. | Fingerprint liveness detection using an improved CNN with image scale equalization | |
Galbally et al. | Iris image reconstruction from binary templates: An efficient probabilistic approach based on genetic algorithms | |
Gad et al. | Iris recognition using multi-algorithmic approaches for cognitive internet of things (CIoT) framework | |
Jagadeesan et al. | Secured cryptographic key generation from multimodal biometrics: feature level fusion of fingerprint and iris | |
Mazumdar et al. | RETINA BASED BIOMETRIC AUTHENTICATION SYSTEM: A REVIEW. | |
CN106709418B (en) | Face identification method and identification device based on scene photograph and certificate photo | |
Das et al. | A new method for sclera vessel recognition using OLBP | |
Srivastava et al. | Fingerprints, Iris and DNA features based multimodal systems: a review | |
Wright et al. | One-shot-learning for visual lip-based biometric authentication | |
CN111539386A (en) | Identity authentication system integrating fingerprint and face living body detection | |
Sujana et al. | An effective CNN based feature extraction approach for iris recognition system | |
Sathish et al. | Multi-algorithmic iris recognition | |
US11762969B1 (en) | Systems and methods for facilitating biometric recognition | |
CN115294638A (en) | Iris identification system deployment method based on FPGA, iris identification method and system | |
Kalsoom et al. | Iris Recognition: Existing Methods and Open Issues | |
Bassi et al. | Human biometric identification through brain print | |
Safie et al. | Practical Consideration in using Pre-trained Convolutional Neural Network (CNN) for Finger Vein Biometric. | |
Sobabe et al. | Biometric system vulnerabilities: a typology of metadata | |
Chirchi et al. | Modified circular fuzzy segmentor and local circular encoder to iris segmentation and recognition | |
Zibran | Eye based authentication: Iris and retina recognition | |
Eldin et al. | Cancellable Multi-Biometric Feature Veins Template Generation Based on SHA-3 Hashing. | |
Sujana et al. | Multi-modal Biometric System for Face and Fingerprint using Convolutional Neural Network | |
Cherry et al. | Photoplethysmography Biometric Authentication using Convolutional Neural Network | |
Karthikeyan et al. | An Advanced Convolutional Based Fusing by Score Level for Multi-Modality Biometric Authentication. |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |