CN110633650A - Convolutional neural network face recognition method and device based on privacy protection - Google Patents
Convolutional neural network face recognition method and device based on privacy protection Download PDFInfo
- Publication number
- CN110633650A CN110633650A CN201910780037.3A CN201910780037A CN110633650A CN 110633650 A CN110633650 A CN 110633650A CN 201910780037 A CN201910780037 A CN 201910780037A CN 110633650 A CN110633650 A CN 110633650A
- Authority
- CN
- China
- Prior art keywords
- face
- convolutional neural
- face recognition
- recognition
- complex matrix
- 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.)
- Pending
Links
Images
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
- G06F21/60—Protecting data
- G06F21/602—Providing cryptographic facilities or services
-
- 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
- G06F21/60—Protecting data
- G06F21/62—Protecting access to data via a platform, e.g. using keys or access control rules
- G06F21/6218—Protecting access to data via a platform, e.g. using keys or access control rules to a system of files or objects, e.g. local or distributed file system or database
- G06F21/6245—Protecting personal data, e.g. for financial or medical purposes
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Bioethics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computer Security & Cryptography (AREA)
- Computer Hardware Design (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
- Collating Specific Patterns (AREA)
Abstract
The invention discloses a convolutional neural network face recognition method and a convolutional neural network face recognition device based on privacy protection, wherein the method comprises the following steps: acquiring a plurality of face images, and preprocessing the plurality of face images; calculating local variance of the preprocessed multiple face images; constructing a first complex matrix by taking a plurality of pre-processed human face images as real-part components and local variances of the plurality of pre-processed human face images as imaginary-part components; performing double random phase encryption of a transform domain on the first complex matrix to generate a second complex matrix; and extracting a first component and a second component from the second complex matrix, respectively inputting the first component and the second component into the convolutional neural networks for training, connecting the two trained convolutional neural networks to generate a face recognition encryption network, and recognizing the face image through the face recognition encryption network. The method has higher safety and can meet the face recognition of different requirements.
Description
Technical Field
The invention relates to the technical field of face recognition, in particular to a convolutional neural network face recognition method and device based on privacy protection.
Background
The face recognition mainly depends on face information of people, and is more and more widely applied in real life, for example, mobile phone unlocking, face brushing payment and the like, and face recognition software is introduced in sequence by amazon, google corporation, microsoft corporation and the like.
The biological information has uniqueness as an important biological feature, and a lot of small secrets such as age estimation, gender identification, blood relationship discrimination and the like based on the face image are hidden in the biological information. Generally, our photos are used for training a face recognition system without permission, and many users are infringed of the right of awareness and privacy without knowing the photos. The san Francisco, Sammervier City of the United states, this year, in turn, prohibits the use of facial recognition software by police and municipalities. The existing face recognition algorithm and system do not fully consider the protection of the privacy information of the face.
Disclosure of Invention
The present invention is directed to solving, at least to some extent, one of the technical problems in the related art.
Therefore, an object of the present invention is to provide a convolutional neural network face recognition method based on privacy protection, which has high security, and when registered face information is stolen or lost, the face information can be re-registered through a newly generated double random phase mask, so that face recognition of different requirements can be satisfied.
The invention also aims to provide a convolutional neural network face recognition device based on privacy protection.
In order to achieve the above object, an embodiment of the present invention provides a convolutional neural network face recognition method based on privacy protection, including:
acquiring a plurality of face images, and preprocessing the plurality of face images;
calculating local variance of the preprocessed multiple face images;
constructing a first complex matrix by taking a plurality of pre-processed human face images as real-part components and local variances of the plurality of pre-processed human face images as imaginary-part components;
performing transform domain double random phase encryption on the first complex matrix to generate a second complex matrix;
extracting a first component and a second component from the second complex matrix, respectively inputting the first component and the second component into two convolutional neural networks for training, connecting the two trained convolutional neural networks to generate a face recognition encryption network, and recognizing a face image through the face recognition encryption network.
According to the convolutional neural network face recognition method based on privacy protection, when the face depth features are extracted, the structural information of the image is jointly used, meanwhile, the privacy of the face information is fully considered, the face image is encrypted through double random phase coding and then input into a convolutional network for training and recognition, and the method has high safety. When the registered face information is stolen or lost, the face information can be re-registered through the newly generated double random phase mask, and the face identification with different requirements can be met.
In addition, the convolutional neural network face recognition method based on privacy protection according to the above embodiment of the present invention may further have the following additional technical features:
further, in an embodiment of the present invention, after generating the face recognition encryption network, the method further includes:
and verifying the face recognition encryption network.
Further, in an embodiment of the present invention, the verifying the face recognition encryption network includes:
acquiring a plurality of face images to form a verification set;
inputting the verification set into the face recognition encryption network for recognition to generate a recognition result of each face image;
and comparing the recognition result of each human face image with the label corresponding to each human face image, and calculating the recognition rate of the human face recognition encryption network according to the comparison result.
Further, in an embodiment of the present invention, the preprocessing the plurality of face images includes:
and aligning the plurality of face images according to the eye coordinates in the face images, and carrying out normalization processing.
Further, in an embodiment of the present invention, the calculation formula of the local variance is:
wherein, IpIs a neighborhood of a pixel point of the face image with the size of W1×W2,L=W1W2Representing the total number of pixel points in the neighborhood,the mean gray value of the neighborhood is represented, and n is the number of the human face images.
Further, in one embodiment of the present invention, the first complex matrix fi c(x, y) is:
fi c(x,y)=fi(x,y)+fi V(x,y)i
wherein f isi(x, y) are a plurality of face images, fi VAnd (x, y) is the local variance of the plurality of face images, i is 1,2, …, n.
Further, in an embodiment of the present invention, the performing transform-domain double random phase encryption on the first complex matrix to generate a second complex matrix includes:
performing double random phase encryption of a Gyrator transform domain on the first complex matrix to generate the second complex matrix g (x, y), which specifically comprises:
r(x,y)=exp[j2πr0(x,y)]
H(ε,η)=exp[j2πH0(ε,η)]
P(ε,η)=Gα{f(x,y)r(x,y)}
g(x,y)=Gβ{P(ε,η)H(ε,η)}
wherein r (x, y) is a phase mask of a spatial domain, a phase mask of a H (epsilon, eta) frequency domain, (x, y) is a spatial domain coordinate, (epsilon, eta) is a frequency domain coordinate, r is a phase mask of a H (epsilon, eta) frequency domain, and0(x,y)、H0(ε, η) is N × M [0,1 ]]And uniformly distributing random number matrixes, wherein N multiplied by M is the size of the human face image, alpha and beta are parameters of Gyrator transformation, and a second complex matrix g (x, y) is the result of double random phase encoding.
In order to achieve the above object, another embodiment of the present invention provides a convolutional neural network face recognition device based on privacy protection, including:
the preprocessing module is used for acquiring a plurality of face images and preprocessing the face images;
the calculation module is used for calculating the local variance of the preprocessed multiple face images;
the construction module is used for constructing a first complex matrix by taking a plurality of human face images before preprocessing as real-part components and taking local variances of the plurality of human face images after preprocessing as imaginary-part components;
the encryption module is used for carrying out double random phase encryption of a transform domain on the first complex matrix to generate a second complex matrix;
and the recognition module is used for extracting a first component and a second component from the second complex matrix, inputting the first component and the second component into the two convolutional neural networks respectively for training, connecting the two trained convolutional neural networks to generate a face recognition encryption network, and recognizing the face image through the face recognition encryption network.
The convolutional neural network face recognition device based on privacy protection of the embodiment of the invention not only jointly uses the structural information of the image when extracting the face depth features, but also fully considers the privacy of the face information, encrypts the face image through double random phase coding and then inputs the face image into a convolutional network for training and recognition, thereby having higher security. When the registered face information is stolen or lost, the face information can be re-registered through the newly generated double random phase mask, and the face identification with different requirements can be met.
In addition, the convolutional neural network face recognition device based on privacy protection according to the above embodiment of the present invention may further have the following additional technical features:
further, in an embodiment of the present invention, the method further includes: a verification module;
and the verification module is used for verifying the face recognition encryption network.
Further, in an embodiment of the present invention, the verification module includes: the device comprises an acquisition unit, an input unit and a comparison unit;
the acquisition unit is used for acquiring a plurality of face images to form a verification set;
the input unit is used for inputting the verification set into the face recognition encryption network for recognition to generate a recognition result of each face image;
and the comparison unit is used for comparing the identification result of each human face image with the label corresponding to each human face image and calculating the identification rate of the human face identification encryption network according to the comparison result.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the present invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
FIG. 1 is a flow chart of a convolutional neural network face recognition method based on privacy protection according to an embodiment of the present invention;
FIG. 2 is a flow chart of a convolutional neural network face recognition method based on privacy protection according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of a face image according to an embodiment of the invention;
FIG. 4 is a schematic view of a local variance map according to one embodiment of the present invention;
FIG. 5 is a diagram illustrating the results of dual random phase encoding according to one embodiment of the present invention;
FIG. 6 is a graph illustrating the recognition rate and the loss curve according to an embodiment of the present invention;
fig. 7 is a schematic structural diagram of a convolutional neural network face recognition device based on privacy protection according to an embodiment of the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention.
The following describes a convolutional neural network face recognition method and device based on privacy protection according to an embodiment of the present invention with reference to the accompanying drawings.
First, a convolutional neural network face recognition method based on privacy protection proposed according to an embodiment of the present invention will be described with reference to the accompanying drawings.
Fig. 1 is a flowchart of a convolutional neural network face recognition method based on privacy protection according to an embodiment of the present invention.
As shown in fig. 1 and fig. 2, the convolutional neural network face recognition method based on privacy protection includes the following steps:
and S101, acquiring a plurality of face images and preprocessing the face images.
Further, as a possible implementation manner, the plurality of face images are aligned according to the eye coordinates in the face image, and normalized.
Specifically, the acquired multiple face images are set to f1(x,y)、f2(x,y)、……、fn(x, y), the dimensions of which are each N M, and (x, y) represents spatial domain coordinates.
And step S102, calculating local variance of the plurality of preprocessed face images.
Specifically, after preprocessing a plurality of face images, f of each face image is calculatedi(x, y) local variance map fi V(x, y). Let neighborhood I of pixel pointpSize W1×W2Then, the local variance calculation formula is:
wherein, IpIs a neighborhood of a pixel point of the face image with the size of W1×W2,L=W1W2Representing the total number of pixel points in the neighborhood,the mean gray value of the neighborhood is represented, and n is the number of the human face images.
It should be noted that, the local variance of the brightness of the pixels at the image boundary is obtained by adopting a symmetric filling method.
Step S103, a plurality of face images before preprocessing are used as real part components, and local variances of the plurality of face images after preprocessing are used as imaginary part components, so that a first complex matrix is constructed.
Constructing a complex matrix f by using the original gray image as a real part component and the local variance map as an imaginary part componenti c(x, y), specifically expressed as:
fi c(x,y)=fi(x,y)+fi V(x,y)i
wherein f isi(x, y) are a plurality of face images, fi VAnd (x, y) is the local variance of the plurality of face images, i is 1,2, …, n.
Step S104, the first complex matrix is encrypted by double random phases in a transform domain to generate a second complex matrix.
As a possible implementation manner, taking the Gyrator transform domain as an example, performing double random phase encryption of the Gyrator transform domain on the first complex matrix to generate a second complex matrix g (x, y), specifically:
r(x,y)=exp[j2πr0(x,y)]
H(ε,η)=exp[j2πH0(ε,η)]
P(ε,η)=Gα{f(x,y)r(x,y)}
g(x,y)=Gβ{P(ε,η)H(ε,η)}
wherein r (x, y) is a phase mask of a spatial domain, a phase mask of a H (epsilon, eta) frequency domain, (x, y) is a spatial domain coordinate, (epsilon, eta) is a frequency domain coordinate, r is a phase mask of a H (epsilon, eta) frequency domain, and0(x,y)、H0(ε, η) is N × M [0,1 ]]And uniformly distributing random number matrixes, wherein N multiplied by M is the size of the human face image, alpha and beta are parameters of Gyrator transformation, and a second complex matrix g (x, y) is the result of double random phase encoding.
In particular, the depth features are extracted by combining the structural information of the images, so that the method has stronger identification performance. The face image is encrypted by adopting double random phase coding, so that the privacy information of a user can be protected, and the security is higher.
Step S105, extracting a first component and a second component from the second complex matrix, respectively inputting the first component and the second component into two Convolutional Neural Networks (CNN) for training, connecting the two trained Convolutional Neural Networks to generate a face recognition encryption network, and recognizing the face image through the face recognition encryption network.
And extracting a first component and a second component of the second complex matrix g (x, y), and inputting the first component and the second component into the CNN of the first branch and the CNN of the second branch respectively for training.
It should be noted that the two components herein include two cases: the first component is a real component, and the second component is an imaginary component; the second is that the first component and the second component are respectively an amplitude component and a phase component. The Convolutional Neural Networks (CNN) of the two branches comprise a plurality of convolutional layers and pooling layers, and finally, the convolutional neural networks are spliced together to obtain a full-connection layer.
It can be understood that when the relevant information of the user is stolen or leaked, the registered biometric template can be replaced by a new phase mask, and the method has revocable property
Further, after the face recognition encryption network is generated, the face recognition encryption network can be verified, and the recognition rate of the face recognition encryption network can be verified.
As a possible implementation manner, the authentication of the face recognition encryption network includes:
acquiring a plurality of face images to form a verification set;
inputting the verification set into a face recognition encryption network for recognition to generate a recognition result of each face image;
and comparing the recognition result of each human face image with the label corresponding to each human face image, and calculating the recognition rate of the human face recognition encryption network according to the comparison result.
Specifically, the verification set is input into a face recognition encryption network to obtain a recognition result of each face image, the recognition result of each face image is compared with a label corresponding to each face image, and the ratio obtained by dividing the sum of the number of correctly recognized results and the sum of the number of samples of the verification set is used as the recognition rate.
For example, the face images in the verification set include five face images with labels of 1,2, and 2, the five face images are sequentially input to the trained face recognition encryption network to obtain recognition results corresponding to the five face images, and if the obtained recognition results are 1,2, and 2, after comparison, the number of correct recognition is four, and if the number of verification lumped samples is five, the recognition rate is 80%.
The convolutional neural network face recognition method is explained by a specific embodiment.
Experiments were performed using an ORL face dataset with a face image size of 112 x 112, and 70% of the dataset was randomly selected for training. Fig. 3 shows 3 face images, and fig. 4 shows a local variance map obtained when the neighborhood matrix is 5 × 5; fig. 5 shows the result of the double random phase encoding, where fig. 5(a) and fig. 5(b) are the real component and the imaginary component, respectively, and fig. 5(c) and fig. 5(d) are the amplitude component and the phase component, respectively, and it can be seen that the information of the face image is hidden in both cases, which plays a role in protecting the user information. Fig. 6 shows the corresponding recognition rate and loss curve change, which converge after a certain number of iterations, fig. 6(a) shows the results of the real component and the imaginary component, and fig. 6(b) shows the results of the magnitude component and the phase component. After 10 tests, the average recognition rates of the two cases are respectively 94.9% and 93.1%.
According to the convolutional neural network face recognition method based on privacy protection provided by the embodiment of the invention, when the face depth features are extracted, the structural information of the image is jointly used, meanwhile, the privacy of the face information is fully considered, the face image is encrypted through double random phase coding and then input into a convolutional network for training and recognition, and the method has higher safety. When the registered face information is stolen or lost, the face information can be re-registered through the newly generated double random phase mask, and the face identification with different requirements can be met.
Next, a convolutional neural network face recognition apparatus based on privacy protection proposed according to an embodiment of the present invention is described with reference to the drawings.
Fig. 7 is a schematic structural diagram of a convolutional neural network face recognition device based on privacy protection according to an embodiment of the present invention.
As shown in fig. 7, the convolutional neural network face recognition device based on privacy protection includes: a preprocessing module 100, a computing module 200, a construction module 300, an encryption module 400, and an identification module 500.
The preprocessing module 100 is configured to acquire a plurality of face images and preprocess the plurality of face images.
And the calculating module 200 is used for calculating the local variance of the preprocessed multiple face images.
A constructing module 300, configured to construct a first complex matrix by using the plurality of pre-processed face images as real components and using the local variance of the plurality of pre-processed face images as imaginary components.
The encryption module 400 is configured to perform transform-domain dual random phase encryption on the first complex matrix to generate a second complex matrix.
The recognition module 500 is configured to extract a first component and a second component from the second complex matrix, input the first component and the second component into the two convolutional neural networks respectively for training, connect the two trained convolutional neural networks to generate a face recognition encryption network, and recognize the face image through the face recognition encryption network.
Further, in an embodiment of the present invention, the method further includes: a verification module 600;
and the verification module 600 is configured to verify the face recognition encryption network.
Further, in an embodiment of the present invention, the verification module further includes: the device comprises an acquisition unit, an input unit and a comparison unit;
and the acquisition unit is used for acquiring a plurality of face images to form a verification set.
And the input unit is used for inputting the verification set into a face recognition encryption network for recognition to generate a recognition result of each face image.
And the comparison unit is used for comparing the identification result of each human face image with the label corresponding to each human face image and calculating the identification rate of the human face identification encryption network according to the comparison result.
Further, in an embodiment of the present invention, the preprocessing the plurality of face images includes:
and aligning the plurality of face images according to the eye coordinates in the face images, and carrying out normalization processing.
Further, in one embodiment of the present invention, the calculation formula of the local variance is:
wherein, IpIs a neighborhood of a pixel point of the face image with the size of W1×W2,L=W1W2Representing the total number of pixel points in the neighborhood,the mean gray value of the neighborhood is represented, and n is the number of the human face images.
Further, in one embodiment of the present invention, the first complex matrix fi c(x, y) is:
fi c(x,y)=fi(x,y)+fi V(x,y)i
wherein f isi(x, y) are a plurality of face images, fi VAnd (x, y) is the local variance of the plurality of face images, i is 1,2, …, n.
Further, in an embodiment of the present invention, the encryption module is further configured to perform double random phase encryption of the Gyrator transform domain on the first complex matrix to generate a second complex matrix g (x, y), specifically:
r(x,y)=exp[j2πr0(x,y)]
H(ε,η)=exp[j2πH0(ε,η)]
P(ε,η)=Gα{f(x,y)r(x,y)}
g(x,y)=Gβ{P(ε,η)H(ε,η)}
wherein r (x, y) is a phase mask of a spatial domain, a phase mask of a H (epsilon, eta) frequency domain, (x, y) is a spatial domain coordinate, (epsilon, eta) is a frequency domain coordinate, r is a phase mask of a H (epsilon, eta) frequency domain, and0(x,y)、H0(ε, η) is N × M [0,1 ]]And uniformly distributing random number matrixes, wherein N multiplied by M is the size of the human face image, alpha and beta are parameters of Gyrator transformation, and a second complex matrix g (x, y) is the result of double random phase encoding.
It should be noted that the foregoing explanation of the embodiment of the convolutional neural network face recognition method based on privacy protection is also applicable to the apparatus of the embodiment, and is not repeated here.
According to the convolutional neural network face recognition device based on privacy protection provided by the embodiment of the invention, when the face depth features are extracted, the structural information of the image is jointly used, meanwhile, the privacy of the face information is fully considered, the face image is encrypted through double random phase coding and then input into a convolutional network for training and recognition, and the security is higher. When the registered face information is stolen or lost, the face information can be re-registered through the newly generated double random phase mask, and the face identification with different requirements can be met.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "a plurality" means at least two, e.g., two, three, etc., unless specifically limited otherwise.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (10)
1. A convolutional neural network face recognition method based on privacy protection is characterized by comprising the following steps:
acquiring a plurality of face images, and preprocessing the plurality of face images;
calculating local variance of the preprocessed multiple face images;
constructing a first complex matrix by taking a plurality of pre-processed human face images as real-part components and local variances of the plurality of pre-processed human face images as imaginary-part components;
performing transform domain double random phase encryption on the first complex matrix to generate a second complex matrix;
extracting a first component and a second component from the second complex matrix, respectively inputting the first component and the second component into two convolutional neural networks for training, connecting the two trained convolutional neural networks to generate a face recognition encryption network, and recognizing a face image through the face recognition encryption network.
2. The convolutional neural network face recognition method based on privacy protection as claimed in claim 1, further comprising after generating the face recognition encryption network:
and verifying the face recognition encryption network.
3. The convolutional neural network face recognition method based on privacy protection as claimed in claim 2, wherein the verifying the face recognition encryption network comprises:
acquiring a plurality of face images to form a verification set;
inputting the verification set into the face recognition encryption network for recognition to generate a recognition result of each face image;
and comparing the recognition result of each human face image with the label corresponding to each human face image, and calculating the recognition rate of the human face recognition encryption network according to the comparison result.
4. The method for convolutional neural network face recognition based on privacy protection as claimed in claim 1, wherein the preprocessing of the plurality of face images comprises:
and aligning the plurality of face images according to the eye coordinates in the face images, and carrying out normalization processing.
5. The privacy protection based convolutional neural network face recognition method of claim 1, wherein the calculation formula of the local variance is as follows:
6. The privacy protection based convolutional neural network face recognition method of claim 1, wherein the first complex matrix fi c(x, y) is:
fi c(x,y)=fi(x,y)+fi V(x,y)i
wherein f isi(x, y) are a plurality of face images, fi VAnd (x, y) is the local variance of the plurality of face images, i is 1,2, …, n.
7. The privacy protection based convolutional neural network face recognition method of claim 1, wherein the performing transform domain double random phase encryption on the first complex matrix generates a second complex matrix, comprising:
performing double random phase encryption of a Gyrator transform domain on the first complex matrix to generate the second complex matrix g (x, y), which specifically comprises:
r(x,y)=exp[j2πr0(x,y)]
H(ε,η)=exp[j2πH0(ε,η)]
P(ε,η)=Gα{f(x,y)r(x,y)}
g(x,y)=Gβ{P(ε,η)H(ε,η)}
wherein r (x, y) is a phase mask of a spatial domain, a phase mask of a H (epsilon, eta) frequency domain, (x, y) is a spatial domain coordinate, (epsilon, eta) is a frequency domain coordinate, r is a phase mask of a H (epsilon, eta) frequency domain, and0(x,y)、H0(ε, η) is N × M [0,1 ]]And uniformly distributing random number matrixes, wherein N multiplied by M is the size of the human face image, alpha and beta are parameters of Gyrator transformation, and a second complex matrix g (x, y) is the result of double random phase encoding.
8. A convolutional neural network face recognition device based on privacy protection, comprising:
the preprocessing module is used for acquiring a plurality of face images and preprocessing the face images;
the calculation module is used for calculating the local variance of the preprocessed multiple face images;
the construction module is used for constructing a first complex matrix by taking a plurality of human face images before preprocessing as real-part components and taking local variances of the plurality of human face images after preprocessing as imaginary-part components;
the encryption module is used for carrying out double random phase encryption of a transform domain on the first complex matrix to generate a second complex matrix;
and the recognition module is used for extracting a first component and a second component from the second complex matrix, inputting the first component and the second component into the two convolutional neural networks respectively for training, connecting the two trained convolutional neural networks to generate a face recognition encryption network, and recognizing the face image through the face recognition encryption network.
9. The privacy protection based convolutional neural network face recognition device of claim 8, further comprising: a verification module;
and the verification module is used for verifying the face recognition encryption network.
10. The convolutional neural network face recognition device based on privacy protection as claimed in claim 9, wherein the verification module comprises: the device comprises an acquisition unit, an input unit and a comparison unit;
the acquisition unit is used for acquiring a plurality of face images to form a verification set;
the input unit is used for inputting the verification set into the face recognition encryption network for recognition to generate a recognition result of each face image;
and the comparison unit is used for comparing the identification result of each human face image with the label corresponding to each human face image and calculating the identification rate of the human face identification encryption network according to the comparison result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910780037.3A CN110633650A (en) | 2019-08-22 | 2019-08-22 | Convolutional neural network face recognition method and device based on privacy protection |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910780037.3A CN110633650A (en) | 2019-08-22 | 2019-08-22 | Convolutional neural network face recognition method and device based on privacy protection |
Publications (1)
Publication Number | Publication Date |
---|---|
CN110633650A true CN110633650A (en) | 2019-12-31 |
Family
ID=68970524
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910780037.3A Pending CN110633650A (en) | 2019-08-22 | 2019-08-22 | Convolutional neural network face recognition method and device based on privacy protection |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110633650A (en) |
Cited By (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111259427A (en) * | 2020-01-21 | 2020-06-09 | 北京安德医智科技有限公司 | Image processing method and device based on neural network and storage medium |
CN111310734A (en) * | 2020-03-19 | 2020-06-19 | 支付宝(杭州)信息技术有限公司 | Face recognition method and device for protecting user privacy |
CN111368795A (en) * | 2020-03-19 | 2020-07-03 | 支付宝(杭州)信息技术有限公司 | Face feature extraction method, device and equipment |
CN111539008A (en) * | 2020-05-22 | 2020-08-14 | 支付宝(杭州)信息技术有限公司 | Image processing method and device for protecting privacy |
CN111538968A (en) * | 2020-05-27 | 2020-08-14 | 支付宝(杭州)信息技术有限公司 | Identity verification method, device and equipment based on privacy protection |
CN111553320A (en) * | 2020-05-14 | 2020-08-18 | 支付宝(杭州)信息技术有限公司 | Feature extraction method for protecting personal data privacy, model training method and hardware |
CN111783803A (en) * | 2020-08-14 | 2020-10-16 | 支付宝(杭州)信息技术有限公司 | Image processing method and device for realizing privacy protection |
CN112257741A (en) * | 2020-09-07 | 2021-01-22 | 北京航空航天大学杭州创新研究院 | Method for detecting generative anti-false picture based on complex neural network |
CN112364372A (en) * | 2020-10-27 | 2021-02-12 | 重庆大学 | Privacy protection method with supervision matrix completion |
CN112949576A (en) * | 2021-03-29 | 2021-06-11 | 北京京东方技术开发有限公司 | Attitude estimation method, attitude estimation device, attitude estimation equipment and storage medium |
WO2022152153A1 (en) * | 2021-01-18 | 2022-07-21 | 北京灵汐科技有限公司 | Image processing method and device, key generation method and device, training method, and computer readable medium |
WO2023001674A3 (en) * | 2021-07-20 | 2023-03-09 | Sony Group Corporation | Camera, method and image processing method |
WO2024007095A1 (en) * | 2022-07-04 | 2024-01-11 | 嘉兴尚坤科技有限公司 | Secure encryption method and system for face data of door access control system |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106485201A (en) * | 2016-09-09 | 2017-03-08 | 首都师范大学 | The color face recognition method of supercomplex encrypted domain |
CN107368819A (en) * | 2017-08-02 | 2017-11-21 | 首都师范大学 | Face identification method and system |
CN108132274A (en) * | 2017-12-21 | 2018-06-08 | 厦门大学 | Echo-planar imaging is without reference scan pattern distortion antidote under non-uniform magnetic field |
CN108776790A (en) * | 2018-06-06 | 2018-11-09 | 海南大学 | Face encryption recognition methods based on neural network under cloud environment |
CN108875905A (en) * | 2018-04-09 | 2018-11-23 | 华中科技大学 | A kind of visibility function Direct Inverse Method of Atmosphere and humidity profiles |
CN109350061A (en) * | 2018-11-21 | 2019-02-19 | 成都信息工程大学 | MR imaging method based on depth convolutional neural networks |
CN109902766A (en) * | 2019-03-25 | 2019-06-18 | 首都师范大学 | A kind of biological feather recognition method and device |
CN109977807A (en) * | 2019-03-11 | 2019-07-05 | 首都师范大学 | Skin detection guard method and system based on complex matrix |
-
2019
- 2019-08-22 CN CN201910780037.3A patent/CN110633650A/en active Pending
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106485201A (en) * | 2016-09-09 | 2017-03-08 | 首都师范大学 | The color face recognition method of supercomplex encrypted domain |
CN107368819A (en) * | 2017-08-02 | 2017-11-21 | 首都师范大学 | Face identification method and system |
CN108132274A (en) * | 2017-12-21 | 2018-06-08 | 厦门大学 | Echo-planar imaging is without reference scan pattern distortion antidote under non-uniform magnetic field |
CN108875905A (en) * | 2018-04-09 | 2018-11-23 | 华中科技大学 | A kind of visibility function Direct Inverse Method of Atmosphere and humidity profiles |
CN108776790A (en) * | 2018-06-06 | 2018-11-09 | 海南大学 | Face encryption recognition methods based on neural network under cloud environment |
CN109350061A (en) * | 2018-11-21 | 2019-02-19 | 成都信息工程大学 | MR imaging method based on depth convolutional neural networks |
CN109977807A (en) * | 2019-03-11 | 2019-07-05 | 首都师范大学 | Skin detection guard method and system based on complex matrix |
CN109902766A (en) * | 2019-03-25 | 2019-06-18 | 首都师范大学 | A kind of biological feather recognition method and device |
Non-Patent Citations (4)
Title |
---|
RANDA F. SOLIMAN ET AL.: "A double random phase encoding approach for cancelable iris recognition", 《PART OF SPRINGER NATURE 2018》 * |
WEI SHEN ET AL.: "A Face Privacy Protection Algorithm Based on Block Scrambling and Deep Learning", 《INTERNATIONAL CONFERENCE ON CLOUD COMPUTING AND SECURITY》 * |
章坚武 等: "卷积神经网络的人脸隐私保护识别", 《中国图象图形学报》 * |
邵珠宏: "基于四元数变换的彩色图像处理算法研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111259427A (en) * | 2020-01-21 | 2020-06-09 | 北京安德医智科技有限公司 | Image processing method and device based on neural network and storage medium |
CN111310734A (en) * | 2020-03-19 | 2020-06-19 | 支付宝(杭州)信息技术有限公司 | Face recognition method and device for protecting user privacy |
CN111368795A (en) * | 2020-03-19 | 2020-07-03 | 支付宝(杭州)信息技术有限公司 | Face feature extraction method, device and equipment |
CN111553320A (en) * | 2020-05-14 | 2020-08-18 | 支付宝(杭州)信息技术有限公司 | Feature extraction method for protecting personal data privacy, model training method and hardware |
CN111553320B (en) * | 2020-05-14 | 2021-12-21 | 支付宝(杭州)信息技术有限公司 | Feature extraction method for protecting personal data privacy, model training method and hardware |
CN111539008A (en) * | 2020-05-22 | 2020-08-14 | 支付宝(杭州)信息技术有限公司 | Image processing method and device for protecting privacy |
CN111539008B (en) * | 2020-05-22 | 2023-04-11 | 蚂蚁金服(杭州)网络技术有限公司 | Image processing method and device for protecting privacy |
WO2021238956A1 (en) * | 2020-05-27 | 2021-12-02 | 支付宝(杭州)信息技术有限公司 | Identity verification method, apparatus and device based on privacy protection |
CN111538968A (en) * | 2020-05-27 | 2020-08-14 | 支付宝(杭州)信息技术有限公司 | Identity verification method, device and equipment based on privacy protection |
CN111783803B (en) * | 2020-08-14 | 2022-06-28 | 支付宝(杭州)信息技术有限公司 | Image processing method and device for realizing privacy protection |
CN111783803A (en) * | 2020-08-14 | 2020-10-16 | 支付宝(杭州)信息技术有限公司 | Image processing method and device for realizing privacy protection |
CN112257741A (en) * | 2020-09-07 | 2021-01-22 | 北京航空航天大学杭州创新研究院 | Method for detecting generative anti-false picture based on complex neural network |
CN112364372A (en) * | 2020-10-27 | 2021-02-12 | 重庆大学 | Privacy protection method with supervision matrix completion |
WO2022152153A1 (en) * | 2021-01-18 | 2022-07-21 | 北京灵汐科技有限公司 | Image processing method and device, key generation method and device, training method, and computer readable medium |
CN112949576A (en) * | 2021-03-29 | 2021-06-11 | 北京京东方技术开发有限公司 | Attitude estimation method, attitude estimation device, attitude estimation equipment and storage medium |
CN112949576B (en) * | 2021-03-29 | 2024-04-23 | 北京京东方技术开发有限公司 | Attitude estimation method, apparatus, device and storage medium |
WO2023001674A3 (en) * | 2021-07-20 | 2023-03-09 | Sony Group Corporation | Camera, method and image processing method |
WO2024007095A1 (en) * | 2022-07-04 | 2024-01-11 | 嘉兴尚坤科技有限公司 | Secure encryption method and system for face data of door access control system |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110633650A (en) | Convolutional neural network face recognition method and device based on privacy protection | |
Galbally et al. | Iris image reconstruction from binary templates: An efficient probabilistic approach based on genetic algorithms | |
Tulyakov et al. | Symmetric hash functions for secure fingerprint biometric systems | |
Bhatnagar et al. | Biometrics inspired watermarking based on a fractional dual tree complex wavelet transform | |
Li et al. | Research on iris image encryption based on deep learning | |
CN110610144B (en) | Expression recognition method and system for privacy protection | |
Wojtowicz et al. | Digital images authentication scheme based on bimodal biometric watermarking in an independent domain | |
CN104091303A (en) | Robust image hashing method and device based on Radon transformation and invariant features | |
Barni et al. | Iris deidentification with high visual realism for privacy protection on websites and social networks | |
Koptyra et al. | Multiply information coding and hiding using fuzzy vault | |
CN107231240A (en) | A kind of higher dual identity recognition methods of security | |
Belguechi et al. | Comparative study on texture features for fingerprint recognition: application to the biohashing template protection scheme | |
CN110516425B (en) | Secret sharing method and system | |
Kamal et al. | A symmetric bio-hash function based on fingerprint minutiae and principal curves approach | |
Natgunanathan et al. | An overview of protection of privacy in multibiometrics | |
Luong et al. | Reconstructing a fragmented face from a cryptographic identification protocol | |
Muhammed et al. | A secure fingerprint template generation mechanism using visual secret sharing with inverse halftoning | |
Islam et al. | Technology review: image enhancement, feature extraction and template protection of a fingerprint authentication system | |
Ito et al. | Cancelable Face Recognition Using Deep Steganography | |
Sinduja et al. | Sheltered iris attestation by means of visual cryptography (sia-vc) | |
Koteswari et al. | vc of IRIS Images for ATM Banking | |
Othman et al. | Fingerprint+ Iris= IrisPrint | |
Fernandez et al. | Protection of online images against theft using robust multimodal biometric watermarking and T-norms | |
Al-Ani et al. | Efficient watermarking based an robust biometric features | |
RASHEED et al. | A New Card Authentication Schema Based on Embed Fingerprint in Image Watermarking and Encryption |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20191231 |