CN117539681A - Sensitive data hiding and recovering method - Google Patents
Sensitive data hiding and recovering method Download PDFInfo
- Publication number
- CN117539681A CN117539681A CN202311348830.9A CN202311348830A CN117539681A CN 117539681 A CN117539681 A CN 117539681A CN 202311348830 A CN202311348830 A CN 202311348830A CN 117539681 A CN117539681 A CN 117539681A
- Authority
- CN
- China
- Prior art keywords
- user
- integer
- vector
- data
- human body
- 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
- 238000000034 method Methods 0.000 title claims abstract description 41
- 239000013598 vector Substances 0.000 claims abstract description 298
- 238000013507 mapping Methods 0.000 claims abstract description 249
- 238000012549 training Methods 0.000 claims abstract description 70
- 238000004364 calculation method Methods 0.000 claims abstract description 30
- 238000011084 recovery Methods 0.000 claims abstract description 23
- 238000012795 verification Methods 0.000 claims description 31
- 238000006243 chemical reaction Methods 0.000 claims description 26
- 238000012545 processing Methods 0.000 claims description 20
- 238000013506 data mapping Methods 0.000 claims description 3
- 238000013500 data storage Methods 0.000 claims 1
- 230000035945 sensitivity Effects 0.000 abstract 1
- 238000007781 pre-processing Methods 0.000 description 7
- 238000012360 testing method Methods 0.000 description 5
- 238000012937 correction Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 230000009466 transformation Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 1
- 230000002441 reversible effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/14—Error detection or correction of the data by redundancy in operation
- G06F11/1402—Saving, restoring, recovering or retrying
- G06F11/1446—Point-in-time backing up or restoration of persistent data
- G06F11/1458—Management of the backup or restore process
- G06F11/1469—Backup restoration techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
- G06F11/14—Error detection or correction of the data by redundancy in operation
- G06F11/1402—Saving, restoring, recovering or retrying
- G06F11/1446—Point-in-time backing up or restoration of persistent data
- G06F11/1448—Management of the data involved in backup or backup restore
-
- 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
Abstract
The hiding and recovering method of the sensitive data comprises the following steps: when the data is hidden, generating an integer vector S by using user sensitive data, calculating to obtain a hash value H of the integer vector S, training a mapping model by a model training program or system, and mapping the human body biological characteristics of the user into the integer vector S, wherein the mapping model is a calculation model capable of mapping the human body biological characteristics of the user into vectors which are originally irrelevant to the human body biological characteristics; mapping model M obtained by training A The hash value H is stored in the integer vector S for training or stored separately; data recovery using human body biological characteristics of user and mapping model M A Restoring the integer vector S, checking whether the restored integer vector is correct or not by using the stored hash value H or the restored hash value H, and restoring the user sensitivity by using the restored integer vector S if the restored integer vector is correctSensory data; the mapping model M obtained by training the method A Hiding the user sensitive data.
Description
Technical Field
The invention belongs to the technical field of information security, and particularly relates to a method for hiding and recovering sensitive data.
Background
One way to secure sensitive and private information (e.g., passwords, secrets, etc.) stored in a mobile phone or a computer is to secure the sensitive and private information by means of a security mechanism provided by an operating system of the mobile phone or the computer, such as a human body biological feature such as a fingerprint. However, the user's mobile phone, computer may not have such a security protection mechanism based on the biometric characteristics of the human body such as a fingerprint, or a hacker may break or bypass such a security protection mechanism of the operating system, stealing sensitive, private information (sensitive data) of the user from the user's mobile phone, computer, etc. computing device.
Disclosure of Invention
The invention aims to provide a solution to the problem of security protection of sensitive data stored in a personal computing device (such as a mobile phone, a computer and the like) by a user.
Aiming at the purpose of the invention, the technical scheme provided by the invention is a hiding and recovering method of sensitive data, and is concretely as follows (scheme one).
The method comprises two parts of data hiding and data recovering;
data hiding:
a mapping model (such as a neural network, a convolutional neural network, a deep learning model and other models capable of achieving the purpose) capable of mapping the human body biological characteristics to a vector which is not related to the human body biological characteristics through training is selected or provided; the vector includes one or more numbers and is represented or expressed by a number (it is critical what the vector itself is, what is the form, but it is critical that the vector ultimately corresponds to an ordered set of numbers);
The client of the user A generates an integer vector S by using user sensitive data (such as a password, a secret and the like), wherein the integer vector S comprises one or more integers (the numbers in the vector are integers, namely, the vector corresponds to a group of integers or an integer group), and a hash value H of the integer vector S is obtained by calculation; alternatively, the user A client generates an integer vector S using user sensitive data 0 Calculating to obtain S 0 And then uses the integer vector S 0 And hash value H to generate an integer vector S;
the user A client trains the mapping model through a model training program or system, whichThe input sample data for training comprises a group of human body biological characteristic data (such as human face, fingerprint and the like) of the user A, the sample data for training is usually subjected to preprocessing such as denoising, correction, alignment, transformation, feature extraction and the like, the preprocessing belongs to a part of a mapping model or does not belong to or partially belongs to the mapping model, the integer vector S is used as a corresponding input quantity in model training, namely, the output quantity of the human body biological characteristic of the user A is used as the output quantity of the human body biological characteristic of the user A in model training, namely, the integer vector S is marked as the corresponding output of the human body biological characteristic of the user A in the model training process (namely, supervised learning is carried out, and the structure and parameters of the mapping model are selected and obtained through model training, so that the error of mapping the human body biological characteristic sample data of the user A to the integer vector S is minimum, and is smaller and is better as the integer vector S is used as a single output or each component or each integer is used as a single output, and is related to the selected model); the structure and parameters of a mapping model are determined and obtained through model training, and the human body biological characteristics of the user A are mapped into integer vectors S; the structure and parameters of the mapping model obtained by training form a mapping model M of the user A A The mapping model hides the sensitive data of the user A (different users have different mapping model structures and/or parameters, and the mapping model data can be stored as non-sensitive data of the users);
the user a client is a computing device used by a user and/or a program running therein (the computing device is a device comprising hardware and software programs);
the model training program or system is a program or system specially used for mapping model training, and the system comprises a software system or a hardware system or a system combining software and hardware;
mapping model M of user A calculated by integer vector S A Data (i.e. representing the mapping model M A Data of structure and parameters) stored on user a computing device or application or other storage mapping model M A The location of the data (e.g., a specialized service platform) is used when user a needs to use; if the integer vector S does not contain the hash value H, the mapping model M is stored at the same time A Location or position of dataOther positions store the hash value H for the user A to use when needed;
data recovery:
the client of the user A obtains the human body biological characteristics of the user A and utilizes the mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 (human body biometric data for mapping is usually subjected to preprocessing such as denoising, correction, alignment, transformation, feature extraction, etc., and the preprocessing is part of the mapping model or not or part of the mapping model), and the mapped vector T is obtained 1 Each number of which is converted into an integer to obtain an integer vector S 1 Integer vector S 1 Each integer in (2) and T before conversion 1 The absolute value of the difference between the corresponding numbers in (a) does not exceed a threshold b, wherein the threshold b is less than 0.5; if T is utilized 1 Cannot be converted to an integer vector S 1 So that S 1 Each integer in (2) and T before conversion 1 If the absolute value of the difference between the corresponding numbers does not exceed the threshold b, the client of the user A re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping human body biometric of user A to vector T 1 The vector T obtained by mapping is re-mapped according to the threshold b 1 Is converted into an integer until an integer vector S is obtained that meets the threshold requirement 1 Alternatively, the integer vector S satisfying the threshold value requirement cannot be obtained after the mapping calculation is performed again for a predetermined number of times 1 Then go into error processing;
then, if the integer vector S does not include the hash value H (data), the user a client calculates the integer vector S 1 Hash value H of (a) 1 Comparison of H 1 And H are the same, if they are the same, the integer vector S is used 1 Recovering sensitive data of the user A; if H 1 If the user A client is different from H, the user A client re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping human body biometric of user A to vector T 1 Reuse of T 1 Converting to obtain integer vector S meeting threshold requirement 1 Until H 1 Identical to H, or remapping meterH after a prescribed number of times 1 If the error is still different from H, the error processing is carried out;
if the integer vector S contains a hash value H (data), the user A client receives the integer vector S 1 Data H corresponding to the hash value H is separated 1 By H 1 Validating integer vector S 1 Accuracy of (calculate S 1 Corresponding to integer vector S 0 Comparing the hash value of the portion of (2) with H 1 If the human body generation characteristics of the user A are the same, the user A passes the verification, otherwise, the user A does not pass the verification, and if the human body generation characteristics of the user A do not pass the verification, the user A client side re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping human body biometric of user A to vector T 1 Reuse of T 1 Converting to obtain integer vector S meeting threshold requirement 1 And (5) re-performing correctness verification until verification is passed, or performing remapping calculation until verification is not passed after reaching a specified number of times, and switching to error processing.
For the hiding and recovering method of the sensitive data, the integer vector S is selected and mapped to the result T 1 Another way of handling is as follows (scheme two):
when the data is hidden, the user A client generates an integer vector S by using the sensitive data, and calculates a hash value H of the integer vector S, or the user A client generates an integer vector S by using the sensitive data 0 Calculating to obtain S 0 And then uses the integer vector S 0 And hash value H to generate an integer vector S; then, each integer in the integer vector S is multiplied by a positive integer L (i.e., the absolute value is enlarged by a factor of L) (for binary numbers, l=2 may be taken k I.e., the low k bits are all 0), an integer vector T is obtained; the user A client trains a mapping model through a model training program or system, and maps the human body biological characteristics of the user A into integer vectors T, wherein sample data for training comprises a group of human body biological characteristic data of the user A; training to obtain a mapping model M hiding sensitive data of user A A ；
When data is recovered, the client of the user A acquires the human body biological characteristics of the user ACharacterization, using a mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 Vector T 1 Each number of (a) is converted into an integer to obtain an integer vector S 1 Integer vector S 1 The value obtained by multiplying each integer in (a) by the integer L is equal to T 1 The absolute value of the difference between the corresponding pre-conversion numbers in (a) does not exceed a threshold b, which is less than L/2; if T is utilized 1 Cannot be converted to an integer vector S 1 So that S 1 The value obtained by multiplying each integer in (a) by the integer L is equal to T 1 If the absolute value of the difference between the corresponding numbers before conversion does not exceed the threshold b, the user A client re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 The vector T obtained by mapping is re-mapped according to the threshold b 1 Is converted into an integer until an integer vector S is obtained that meets the threshold requirement 1 Alternatively, the integer vector S satisfying the threshold b requirement cannot be obtained after the mapping calculation is performed again a predetermined number of times 1 Then go into error processing;
then, if the integer vector S does not include the hash value H (data), the user a client calculates the integer vector S 1 Hash value H of (a) 1 Comparison of H 1 And H are the same, if they are the same, the integer vector S is used 1 Recovering the user sensitive data; if H 1 If the user A client is different from H, the user A client re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 Using vector T 1 And threshold b are reconverted to an integer vector S 1 Until H 1 Same as H, or after the remapping calculation reaches a prescribed number of times H 1 If the error is still different from H, the error processing is carried out;
if the integer vector S contains a hash value H (data), the user A client receives the integer vector S 1 Data H corresponding to the hash value H is separated 1 By H 1 Validating integer vector S 1 Accuracy of (meter)Calculation S 1 Corresponding integer vector S 0 Comparing the hash value of the portion of (2) with H 1 If the human body generation characteristics of the user A are the same, the user A passes the verification, otherwise, the user A does not pass the verification, and if the human body generation characteristics of the user A do not pass the verification, the user A client side re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping human body biometric of user A to vector T 1 Reuse of T 1 Converting to obtain integer vector S meeting threshold requirement 1 And (5) re-performing correctness verification until verification is passed, or performing remapping calculation until verification is not passed after reaching a specified number of times, and switching to error processing.
For the hiding and recovering method of the sensitive data, the integer vector S is selected and mapped to the result T 1 Another way of handling is as follows (scheme three):
when the data is hidden, the client side of the user A generates an m-value integer vector S by using the user sensitive data, and calculates a hash value H of the integer vector S, wherein the value of each integer in the m-value integer vector S is one of the agreed m integer values, and m is more than or equal to 2 (for example, the m integer values are {0,1}, { -1,0,1}, { -100,20,230,500,1000 }); alternatively, user A client generates an m-valued integer vector S using user sensitive data 0 Calculating to obtain S 0 And then uses the integer vector S 0 And hash value H to generate an m-valued integer vector S; then, the user A client trains a mapping model through a model training program or system, and maps the human body biological characteristics of the user A into integer vectors S, wherein sample data for training comprises a group of human body biological characteristic data of the user A; training to obtain a mapping model M hiding sensitive data of user A A ；
When data is recovered, a user A client acquires human body biological characteristics of a user A and utilizes a mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 Vector T 1 Each of which is converted into one of m integer values to obtain an m-valued integer vector S 1 M-valued integer vector S 1 Each integer of (2) is equal to T 1 Corresponding to (a)The absolute value of the difference between the numbers before conversion of (a) does not exceed a threshold value corresponding to the integer obtained by conversion; if T is utilized 1 Cannot be converted to an m-valued integer vector S 1 So that S 1 Each integer of (2) is equal to T 1 The absolute value of the difference between the corresponding numbers before conversion does not exceed the threshold corresponding to the integer obtained by conversion, the client of the user A re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 The vector T obtained by mapping is re-mapped according to the threshold value 1 Is converted into one integer of m integer values until an integer vector S meeting the threshold requirement is obtained 1 Alternatively, the integer vector S satisfying the threshold b requirement cannot be obtained after the mapping calculation is performed again a predetermined number of times 1 Then go into error processing;
then, if the integer vector S does not include the hash value H (data), the user a client calculates the integer vector S 1 Hash value H of (a) 1 Comparison of H 1 And H are the same, if they are the same, the integer vector S is used 1 Recovering the user sensitive data; if H 1 If the user A client is different from H, the user A client re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 Reuse vector T 1 Converting to obtain integer vector S meeting threshold requirement 1 Until H 1 Same as H, or after the remapping calculation reaches a prescribed number of times H 1 If the error is still different from H, the error processing is carried out;
if the integer vector S contains a hash value H (data), the user A client receives the integer vector S 1 Data H corresponding to the hash value H is separated 1 By H 1 Validating integer vector S 1 Accuracy of (calculate S 1 Corresponding to integer vector S 0 Comparing the hash value of the portion of (2) with H 1 If the verification is the same, the verification is passed, otherwise, the verification is not passed), if the verification is not passed, the client side of the user A re-acquires the human body generation characteristics of the user A,reuse of mapping model M A Mapping human body biometric of user A to vector T 1 Reuse of T 1 Converting to obtain integer vector S meeting threshold requirement 1 Carrying out correctness verification again until verification is passed, or carrying out remapping calculation until verification is still not passed after reaching the specified times, and transferring to error processing;
The threshold value corresponding to the m integer values is as follows:
if the agreed m integer values are in order from small to large: i 1 ,…,I m Integer I j The threshold value of (2) is a positive number L j And U j Respectively representing the corresponding lower threshold value and upper threshold value, j=1, …, m; the thresholds for the adjacent two integers are: u (U) j +L j+1 <I j+1 -I j J=1, …, m-1 (adjacent two integers, the sum of the upper threshold of the small integer and the lower threshold of the large integer is not greater than the absolute value of the difference between the two numbers);
vector T 1 The way in which one of the numbers t is converted to one of the m integer values is as follows:
if t is in the interval [ I ] j -L j ,I j +U j ]In, t is converted into I j ，j＝1,…,m；
If t is not at any one [ I ] j -L j ,I j +U j ]In the interval, j=1, …, m, then t cannot be converted.
For the hiding and recovering method of the sensitive data, the mapping model M of the user A A In addition to using the human body biometric sample data of the user a, the mapping model may be trained using other user or human-generated or automatically generated human body biometric sample data that is mapped to an integer vector different from the integer vector S (a fixed one of the different integer vectors or a randomly selected integer vector different from the integer vector S of the user a); the human body biological characteristic sample data generated by human body or automatically is generated by using a program and has the characteristics similar to human body biological characteristics Data of characteristics but not of real human biological characteristics (e.g. program synthesized fingerprints, faces, sounds, etc.); other users used by different users or human body biological feature sample data generated manually or automatically are the same or different.
For the above method for hiding and recovering sensitive data, if the sensitive data of the user is too many, because of limitation of calculation complexity and calculation amount, hiding and recovering all sensitive data cannot be realized through mapping from the human body biological characteristics of the user to one integer vector, or the user has sensitive data with different purposes, hiding and recovering sensitive data with different purposes cannot be realized through mapping from the human body biological characteristics of the user to one integer vector, the sensitive data of the user can be divided into a plurality of integer vectors, and mapping from the human body biological characteristics of the user to a plurality of integer vectors can be realized through model training and building a plurality of mapping models respectively, but the efficiency is lower, and the following method can be adopted for this purpose:
mapping model M at user A A In addition to the human body biometric sample data, the input data may further include control data, where the control data is used to control the output result of the mapping model (the same human body biometric input data, different control data correspond to different output results).
If the input data includes control data, and the mapping model M of the user A A In the training process of the system, besides the human body biological characteristic sample data of the user A, other human body biological characteristic sample data generated by other users or people or generated automatically are used for training the mapping model, so that integer vectors corresponding to the human body biological characteristic sample data mapping of other users or people or generated automatically in the training process are changed by changing the control data or are changed by changing the uncontrolled data.
If the input data includes control data, the application modes of the control data include (not all possible application modes):
when the data is hidden, the user A client generates a plurality of integer vectors by using the user sensitive data, and a mapping model is formedWhen training, a plurality of integer vectors are used as mapping output data of the human body biological characteristics of the user A, and different control data inputs correspond to different integer vectors for output; calculating a hash value for a plurality of integer vectors, or calculating a hash value for each integer vector separately; if a hash value is calculated for a plurality of integer vectors, the human body biological characteristic data of the user A is utilized to pass through the mapping model M during data recovery A After all integer vectors are calculated and converted, comparing whether the hash value calculated before data hiding is the same with the hash value of all integer vectors obtained through mapping and conversion; if a hash value is calculated for each integer vector, the human body biological characteristic data of the user A is used to pass through the mapping model M during data recovery A And calculating and converting to obtain each integer vector, and comparing whether the hash value calculated before hiding the data is the same as the hash value of the integer vector obtained by mapping and converting.
The control data is typically not secret data.
For the above method for hiding and recovering sensitive data, the input data of the mapping model may further include user identification data of the user a, where the user identification data of the user a is converted from user identification information of the user a, and the user identification information of the user a is objective information (objectively existing information, such as birthdays, addresses, graduation times, certificate numbers, etc. of the user a or family members and friends thereof) known by the user a; during the training of the mapping model and the data recovery, the user A inputs the user identification information through the user A client, and the user A client converts the user identification information into the user identification data for the training of the mapping model and the data recovery.
It can be seen from the above description that, based on the method of the present invention, the sensitive data of the user is hidden in the structure and parameters of the mapping model from the human body biological feature to the integer vector by means of the training of the mapping model from the human body biological feature to the integer vector, and when the sensitive data of the user needs to be used, the sensitive data of the user can be recovered by using the human body biological feature and the mapping model of the user, and the method of protecting the sensitive data of the user by using the human body biological feature is independent of the security protection mechanism of the operating system of the computing device, and the problem that an attacker breaks and bypasses the security protection mechanism of the operating system of the computing device does not exist.
It should be noted that the sensitive data of the present invention is originally unrelated to the human body biological characteristics of the user and the mapping model, the sensitive data exists prior to the mapping model, the sensitive data is not generated by the human body biological characteristics of the user and the mapping model, the present invention only hides the sensitive data of the user in the structure and the parameters of the mapping model through the training of the mapping model, and recovers the sensitive data of the user through the human body biological characteristics.
Detailed Description
The following describes specific embodiments of the present invention. The following is merely illustrative of possible embodiments of the invention, and is not meant to be a limitation on the scope of the invention.
The invention is not limited to the type of human body biological characteristics, and can distinguish human body biological characteristics of different people, such as fingerprint, face, iris, voice and the like. Because the human body fingerprint features are relatively fixed, the acquisition is convenient (most of the current portable and mobile computing devices have fingerprint reading functions), and the processing is relatively simple, the method is the preferred human body biological feature in implementation. The human body biological characteristics used for mapping model training and data recovery are usually subjected to preprocessing, such as denoising, correction, alignment, transformation, feature extraction and the like, and various methods and schemes for helping or effectively and stably extracting the feature information of the human body biological characteristics of the user can be adopted in the preprocessing process, so that the invention is not limited. Preprocessing a part of the mapping model or not or part of the mapping model, and if the part of the mapping model or not belongs to the mapping model, inputting the data into the mapping model as (part of) preprocessed data.
In the implementation of the present invention, there is no relation between what the vector itself is and what the form is, the vector eventually corresponds to an ordered set of numbers including real numbers and complex numbers, but one complex number is represented by a pair of ordered real numbers, so the vector eventually corresponds to an ordered set of real numbers. How the vector is used in the mapping model, and also in combination with the form of the output (data) of the mapping model.
The invention aims at that the user sensitive data is digital information, the digital information is represented by binary numbers in a computer, the sensitive data, no matter what content is what, is what in structural form, can be regarded as a byte string or a group of byte strings, and one or more bytes in the byte strings can be regarded as integers, therefore, the simplest mode (certainly not the only mode) for obtaining the integer vector S from the user sensitive data is to directly decompose the byte string corresponding to the sensitive data into a plurality of integers, or decompose the byte string obtained after the sensitive data is subjected to reversible processing into a plurality of integers, and the number of the integers is related to the number of bytes corresponding to the integers in a computing system and a computing program used for training and computing a mapping model, if the number of the bytes corresponding to the integers in the computing system and the computing program used for training and computing the mapping model is small, the number of the integers contained in the integer vector is more, otherwise, the integer vector is less.
The present invention is not particularly limited to the mapping model, and the human body biological feature is mapped into a vector of an inclusion number which is originally irrelevant to the human body biological feature by training, for example, a neural network, a convolutional neural network, a deep learning model, etc. In the practice of the present invention, the practitioner selects the appropriate mapping model according to needs, requirements (e.g., computational efficiency, accuracy). In the model training process, the human body biological characteristics of the user A are used as input quantity, the integer vector S is used as output quantity corresponding to the input quantity in the model training, namely in the model training process, the integer vector S is marked as output corresponding to the human body biological characteristics of the user A, namely the model training is supervised learning; in the model training process, the structure and parameters of the mapping model are obtained through selection and calculation, so that the error of mapping the human body biological feature sample data of the user A to the integer vector S through the mapping model is as minimum as possible, and the smaller the error is, the closer the error is to 0 the better the error is. In particular implementations, the integer vector S may be provided as a single output or as separate outputs for each component or integer thereof, depending on the mapping model chosen. The present invention is not particularly limited to the mapping model, as long as the human body biological feature can be mapped into a vector which is originally irrelevant to the human body biological feature through training, and therefore, in the implementation, the mapping model can be a single or single type mapping model, or a combination of a plurality of or a plurality of types of mapping models.
The selection of the threshold b is related to the mismapping rate and the missed mapping rate (corresponding to the misjudgment rate and the missed judgment rate in the human body biological feature recognition and reflecting the index of the mapping accuracy), the threshold b is small, the mismapping rate is low (the error rate of mapping the human body biological features of other users to the integer vector S is low), but the missed mapping rate is high (the error rate of mapping the human body biological features of the user A to the integer vector S is high); conversely, the threshold b is large, the false mapping rate is high (the false rate of mapping the human body biometric feature of the other user to the integer vector S is high), but the missed mapping rate is low (the false rate of mapping the human body biometric feature of the user a to the integer vector S is low). One basic requirement for choosing the threshold b is: to train a mapping model M A The user human body biological characteristic sample data is used as a mapping model M obtained by training A And an integer vector S can be derived from the mapping result based on the threshold b. The value of b is generally about 0.1 to 0.3, and the threshold value b is determined according to the index requirement and by combining with the actual test in the specific implementation.
For scheme one, at the time of data recovery, vector T obtained by mapping human body biological characteristics of user is used 1 One way of converting each number of (c) into an integer is as follows:
taking the decimal part of the absolute value of the number to be converted; (1) If the decimal part does not exceed b, removing the decimal part of the number to be converted to obtain a converted integer; (2) If the decimal part is not less than 1-b, adding 1 to the absolute value of the integer part of the number to be converted, and adding the original positive and negative signs of the number to be converted to obtain a converted integer; (3) If the conversion in (1) or (2) is not possible, the conversion cannot be completed.
For scheme two, the selection of the integer L is related to the error mapping rate and the miss mapping rate, the L is small, the error mapping rate is low (the error rate of mapping human body biological features of other users to the integer vector S is low), but the miss mapping rate is high (the error rate of mapping human body biological features of the user A to the integer vector S is high); conversely, L is large, the error mapping rate is high (the error rate of mapping human body biological features of other users to the integer vector S is high), but the miss mapping rate is low (the error rate of mapping human body biological features of user a to the integer vector S is low); l can be generally 10-100 or more, b is generally about 0.1-0.3L, and the integer L, b is determined according to the index requirement and the actual test in the specific implementation.
For scheme II, during data recovery, vector T obtained by mapping human body biological characteristics of user is used 1 One way of converting each number of (c) into an integer is as follows:
taking an integer part of the absolute value of the number to be converted, dividing the integer part of the absolute value by L, and obtaining quotient and remainder; (1) If the remainder is not more than b, adding the original positive and negative signs of the number to be converted to the quotient to obtain a converted integer; (2) If the remainder is not less than L-b, adding 1 to the quotient and adding the original positive and negative signs of the number to be converted to obtain a converted integer; (3) If the conversion in (1) or (2) is not possible, the conversion cannot be completed.
For scheme two, the selection of the integer L is related to the error mapping rate and the miss mapping rate, the L is small, the error mapping rate is low (the error rate of mapping human body biological features of other users to the integer vector S is low), but the miss mapping rate is high (the error rate of mapping human body biological features of the user A to the integer vector S is high); conversely, L is large, the error mapping rate is high (the error rate of mapping human body biological features of other users to the integer vector S is high), but the miss mapping rate is low (the error rate of mapping human body biological features of user a to the integer vector S is low); l can be generally 10-100 or more, b is generally about 0.1-0.3L, and the integer L, b is determined according to the index requirement and the actual test in the specific implementation. The selection of the magnification factor L and the threshold value b also has a basic requirement that: to train a mapping model M A The user human body biological characteristic sample data is used as a mapping model obtained by trainingM is a kind of A According to L, b, the integer vector S can be derived from the mapping result.
The third main scheme aims at the situation that the byte number of the sensitive data is relatively small, such as the security protection of passwords, or the sensitive data is divided into a plurality of m-value integer vectors in combination with control data, and then the m-value integer vectors are mapped and associated with the human body biological characteristics of a user.
As described above, for scheme three, the threshold values corresponding to the m integer values are as follows:
if the agreed m integer values are in order from small to large: i 1 ,…,I m Integer I j The threshold value of (2) is a positive number L j And U j Respectively representing a lower threshold and an upper threshold, j=1, …, m; u (U) j +L j+1 <I j+1 -I j (adjacent two integers, the sum of the upper threshold of the small integer and the lower threshold of the large integer is not greater than the absolute value of the difference between the two).
For two adjacent integers I j And I j+1 In general, U can be taken j ＝L j+1 ＝(0.1～0.3)(I j+1 -I j ) For I 1 L can be taken 1 ＝U 1 For I m U can be taken out m ＝L m But this is not absolute and need not be the case in this way.
Regarding m and I 1 ,…,I m Firstly, according to the actual needs, for example, the user sensitive data can be just regarded as a vector, and the value of each component of the vector is m different integers, and secondly, the mapping accuracy needs are satisfied. In particular, the integer I is determined according to the index requirement and in combination with the actual test 1 ,…,I m And their upper and lower thresholds. Integer I 1 ,…,I m And the selection of the upper and lower threshold values of the two, wherein the selection has the same basic requirement: to train a mapping model M A The user human body biological characteristic sample data is used as a mapping model M obtained by training A According to the integer I 1 ,…,I m And the selection of their upper and lower thresholds can result in an integer vector S from the mapping result.
Whether the scheme one, the scheme two or the scheme three is selected in the specific implementation depends on the mapping effect and the calculation efficiency of the adopted mapping model, and an implementer can select one scheme by testing and comparing the mapping effect and the calculation efficiency of the three schemes in the specific implementation.
In the implementation of the invention, the mapping model M for user A A In addition to using the human body biometric sample data of user a, the mapping model may be trained using other user (e.g., sample data from a common dataset) or human-generated or automatically generated human body biometric sample data, which is mapped to an integer vector different from the integer vector S, e.g., a fixed integer vector, or a randomly selected integer vector different from the integer vector selected by the current user a, e.g., there is a well-known fixed integer vector to which the human body biometric of the other user is mapped when the mapping model of user a is trained, or for each other user when the mapping model of user a is trained.
In the implementation of the invention, if the sensitive data of the user is too much, because of the limitation of the calculation complexity and the calculation amount, the hiding and recovering of all the data can not be realized through the mapping from the human body biological characteristics of the user to one integer vector, or the sensitive data of the user with different purposes can not be realized through the mapping from the human body biological characteristics of the user to one integer vector, the sensitive data of the user can be divided into a plurality of integer vectors, and the mapping from the human body biological characteristics of the user to the different integer vectors can be realized through model training and establishing a plurality of mapping models respectively, but the implementation efficiency is lower, and the following implementation modes can be adopted:
mapping model M at user A A In the training process of (1), the input data comprises control data besides human body biological characteristic sample data, wherein the control data is used for controlling the output result of the mapping model (the same human body biological characteristic input data and different control data correspond to different output results);
when the data is hidden, the client side of the user A generates a plurality of integer vectors S by using the user sensitive data, and when the mapping model is trained, the plurality of integer vectors are used as human body biological feature mapping output data of the user A, and different control data are input and correspond to different integer vectors to be output; calculating a hash value for a plurality of integer vectors, or calculating a hash value for each integer vector separately; if a hash value is calculated for a plurality of integer vectors, the human body biological characteristic data of the user A is utilized to pass through the mapping model M during data recovery A After all integer vectors are calculated and converted, comparing whether the hash value calculated before data hiding is the same with the hash value of all integer vectors obtained through mapping and conversion; if a hash value is calculated for each integer vector, the human body biological characteristic data of the user A is used to pass through the mapping model M during data recovery A And calculating and converting to obtain each integer vector, and comparing whether the hash value calculated before hiding the data is the same as the hash value of the integer vector obtained by mapping and converting.
The application scenarios of the above application modes of the control data include: (1) The user has sensitive data with different purposes, which respectively correspond to different integer vectors, different outputs can be correspondingly realized under the condition of the same human body biological characteristics through the control data, namely, the sensitive data (integer vectors) with different purposes of the user are hidden by utilizing a mapping model, namely, the different control data correspond to the sensitive data with different purposes, and the corresponding sensitive data of the user can be hidden and restored through the control data, and at the moment, each integer vector corresponds to a hash value; (2) The user sensitive data is too much, an integer vector is generated for mapping model training, the calculation amount is large, at this time, a plurality of integer vectors can be generated for mapping model training aiming at the user sensitive data, the control data is used for controlling the mapping model to correspond to different outputs, namely the integer vectors under the condition of the same human body biological characteristics, and at this time, all the integer vectors correspond to one hash value or each integer vector corresponds to one hash value.
The type of control data is specifically what, depending on the mapping model used, for example, if the human body biological feature is a human face, the control data may be special text or special pictures additionally input, or additional data (such as integers, etc.) set by human, and the control data is specifically what is determined by the mapping model used by the implementer in combination.
In addition to the above application modes and application scenes, other application modes and application scenes are also possible for the control data, for example, the control data is a random number playing a role in disturbance, or different control data corresponds to different applications, etc.
In an embodiment, if the input data includes control data, and the mapping model M of the user A A In the training process of the system, besides the human body biological characteristic sample data of the user A, other human body biological characteristic sample data generated by other users or people or generated automatically are used for training the mapping model, so that integer vectors corresponding to the human body biological characteristic sample data mapping of other users or people or generated automatically in the training process are changed by changing the control data or are changed by changing the uncontrolled data.
In a specific implementation, the input data of the mapping model may further include user identification data of the user a, where the user identification data of the user a is converted from user identification information of the user a, and the user identification information of the user a is objective information known by the user a, for example, birthdays, addresses, graduation times, certificate numbers, and the like of the user a or family members and friends thereof; during the training of the mapping model and the data recovery, the user A inputs the user identification information through the user A client, and the user A client converts the user identification information into the user identification data for the training of the mapping model and the data recovery. The user identification information input by the user through the user a client is a digital quantity, and thus can be conveniently converted into one or more integers input as a mapping model.
Other specific technical implementations not described are known to those skilled in the relevant art and are self-evident to those skilled in the art.
Claims (9)
1. A method for hiding and recovering sensitive data is characterized in that:
the method comprises two parts of data hiding and data recovering;
data hiding:
a mapping model capable of mapping the human body biological characteristics to a vector which is originally irrelevant to the human body biological characteristics through training is selected or selected; the vector comprises one or more numbers and is represented or expressed by a number;
the client of the user A generates an integer vector S by using the user sensitive data, wherein the integer vector S comprises one or more integers, and a hash value H of the integer vector S is obtained by calculation; alternatively, the user A client generates an integer vector S using user sensitive data 0 Calculating to obtain S 0 And then uses the integer vector S 0 And hash value H to generate an integer vector S;
the user A client trains the mapping model through a model training program or system, wherein input sample data for training comprises a group of human body biological characteristic data of the user A, and an integer vector S is used as a corresponding input quantity in model training, namely, the output quantity of the human body biological characteristic of the user A, namely, in the model training process, the integer vector S is marked as the output corresponding to the human body biological characteristic of the user A; the structure and parameters of a mapping model are determined and obtained through model training, and the human body biological characteristics of the user A are mapped into integer vectors S; the structure and parameters of the mapping model obtained by training form a mapping model M of the user A A The mapping model conceals sensitive data of the user A;
the user A client is a computing device used by a user and/or a program running in the computing device;
the model training program or system is a program or system specially used for mapping model training, and the system comprises a software system or a hardware system or a system combining software and hardware;
mapping model M of user A calculated by integer vector S A Data storage in user A computing device or application system or other storage mapping model M A The location of the data is used when user a needs to use; if the integer vector S does not contain the hash value H, the mapping model M is stored at the same time A The position of the data or other positions store a hash value H for use when needed by a user A;
data recovery:
the client of the user A obtains the human body biological characteristics of the user A and utilizes the mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 Mapping the obtained vector T 1 Each number of which is converted into an integer to obtain an integer vector S 1 Integer vector S 1 Each integer in (2) and T before conversion 1 The absolute value of the difference between the corresponding numbers in (a) does not exceed a threshold b, wherein the threshold b is less than 0.5; if T is utilized 1 Cannot be converted to an integer vector S 1 So that S 1 Each integer in (2) and T before conversion 1 If the absolute value of the difference between the corresponding numbers does not exceed the threshold b, the client of the user A re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping human body biometric of user A to vector T 1 The vector T obtained by mapping is re-mapped according to the threshold b 1 Is converted into an integer until an integer vector S is obtained that meets the threshold requirement 1 Alternatively, the integer vector S satisfying the threshold value requirement cannot be obtained after the mapping calculation is performed again for a predetermined number of times 1 Then go into error processing;
then, if the integer vector S does not include the hash value H, the user a client calculates the integer vector S 1 Hash value H of (a) 1 Comparison of H 1 And H are the same, if they are the same, the integer vector S is used 1 Recovering sensitive data of the user A; if H 1 If the user A client is different from H, the user A client re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping human body biometric of user A to a vectorT 1 Reuse of T 1 Converting to obtain integer vector S meeting threshold requirement 1 Until H 1 Same as H, or after the remapping calculation reaches a prescribed number of times H 1 If the error is still different from H, the error processing is carried out;
If the integer vector S contains the hash value H, the client of the user A obtains the integer vector S 1 Data H corresponding to the hash value H is separated 1 By H 1 Validating integer vector S 1 If the verification is not passed, the user A client re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping human body biometric of user A to vector T 1 Reuse of T 1 Converting to obtain integer vector S meeting threshold requirement 1 And (5) re-performing correctness verification until verification is passed, or performing remapping calculation until verification is not passed after reaching a specified number of times, and switching to error processing.
2. The method for hiding and recovering sensitive data according to claim 1, wherein:
selection of integer vector S and mapping result T 1 Another way of handling is as follows:
when the data is hidden, the user A client generates an integer vector S by using the sensitive data, and calculates a hash value H of the integer vector S, or the user A client generates an integer vector S by using the sensitive data 0 Calculating to obtain S 0 And then uses the integer vector S 0 And hash value H to generate an integer vector S; then, multiplying each integer in the integer vector S by a positive integer L to obtain an integer vector T; the user A client trains a mapping model through a model training program or system, and maps the human body biological characteristics of the user A into integer vectors T, wherein sample data for training comprises a group of human body biological characteristic data of the user A; training to obtain a mapping model M hiding sensitive data of user A A ；
When the data is recovered, the client of the user A acquires the human body biological characteristics of the user A, thereby facilitating the recovery of the dataUsing a mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 Vector T 1 Each number of (a) is converted into an integer to obtain an integer vector S 1 Integer vector S 1 The value obtained by multiplying each integer in (a) by the integer L is equal to T 1 The absolute value of the difference between the corresponding pre-conversion numbers in (a) does not exceed a threshold b, which is less than L/2; if T is utilized 1 Cannot be converted to an integer vector S 1 So that S 1 The value obtained by multiplying each integer in (a) by the integer L is equal to T 1 If the absolute value of the difference between the corresponding numbers before conversion does not exceed the threshold b, the user A client re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 The vector T obtained by mapping is re-mapped according to the threshold b 1 Is converted into an integer until an integer vector S is obtained that meets the threshold requirement 1 Alternatively, the integer vector S satisfying the threshold b requirement cannot be obtained after the mapping calculation is performed again a predetermined number of times 1 Then go into error processing;
then, if the integer vector S does not include the hash value H, the user a client calculates the integer vector S 1 Hash value H of (a) 1 Comparison of H 1 And H are the same, if they are the same, the integer vector S is used 1 Recovering the user sensitive data; if H 1 If the user A client is different from H, the user A client re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 Using vector T 1 And threshold b are reconverted to an integer vector S 1 Until H 1 Same as H, or after the remapping calculation reaches a prescribed number of times H 1 If the error is still different from H, the error processing is carried out;
if the integer vector S contains the hash value H, the client of the user A obtains the integer vector S 1 Data H corresponding to the hash value H is separated 1 By H 1 Validating integer vector S 1 If the verification is not passed, thenThe user A client terminal re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping human body biometric of user A to vector T 1 Reuse of T 1 Converting to obtain integer vector S meeting threshold requirement 1 And (5) re-performing correctness verification until verification is passed, or performing remapping calculation until verification is not passed after reaching a specified number of times, and switching to error processing.
3. The method for hiding and recovering sensitive data according to claim 1, wherein:
The integer vector S is selected and mapped to the result T 1 Another way of handling is as follows:
when the data is hidden, a client side of a user A generates an m-value integer vector S by using user sensitive data, a hash value H of the integer vector S is obtained by calculation, and the value of each integer in the m-value integer vector S is one of the agreed m integer values, wherein m is more than or equal to 2; alternatively, user A client generates an m-valued integer vector S using user sensitive data 0 Calculating to obtain S 0 And then uses the integer vector S 0 And hash value H to generate an m-valued integer vector S; then, the user A client trains a mapping model through a model training program or system, and maps the human body biological characteristics of the user A into integer vectors S, wherein sample data for training comprises a group of human body biological characteristic data of the user A; training to obtain a mapping model M hiding sensitive data of user A A ；
When data is recovered, a user A client acquires human body biological characteristics of a user A and utilizes a mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 Vector T 1 Each of which is converted into one of m integer values to obtain an m-valued integer vector S 1 M-valued integer vector S 1 Each integer of (2) is equal to T 1 The absolute value of the difference between the corresponding numbers before conversion does not exceed the threshold value corresponding to the integer obtained by conversion; if T is utilized 1 Cannot be converted to an m-valued integer vector S 1 So that S 1 Each integer of (2) is equal to T 1 The absolute value of the difference between the corresponding numbers before conversion does not exceed the threshold corresponding to the integer obtained by conversion, the client of the user A re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 The vector T obtained by mapping is re-mapped according to the threshold value 1 Is converted into one integer of m integer values until an integer vector S meeting the threshold requirement is obtained 1 Alternatively, the integer vector S satisfying the threshold b requirement cannot be obtained after the mapping calculation is performed again a predetermined number of times 1 Then go into error processing;
then, if the integer vector S does not include the hash value H, the user a client calculates the integer vector S 1 Hash value H of (a) 1 Comparison of H 1 And H are the same, if they are the same, the integer vector S is used 1 Recovering the user sensitive data; if H 1 If the user A client is different from H, the user A client re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping the human body biological characteristics of the user A to obtain a vector T 1 Reuse vector T 1 Converting to obtain integer vector S meeting threshold requirement 1 Until H 1 Same as H, or after the remapping calculation reaches a prescribed number of times H 1 If the error is still different from H, the error processing is carried out;
if the integer vector S contains the hash value H, the client of the user A obtains the integer vector S 1 Data H corresponding to the hash value H is separated 1 By H 1 Validating integer vector S 1 If the verification is not passed, the user A client re-acquires the human body generation characteristics of the user A and re-utilizes the mapping model M A Mapping human body biometric of user A to vector T 1 Reuse of T 1 Converting to obtain integer vector S meeting threshold requirement 1 Carrying out correctness verification again until verification is passed, or carrying out remapping calculation until verification is still not passed after reaching the specified times, and transferring to error processing;
the threshold value corresponding to the m integer values is as follows:
if the agreed m integer values are in order from small to large: i 1 ,…,I m Integer I j The threshold value of (2) is a positive number L j And U j Respectively representing the corresponding lower threshold value and upper threshold value, j=1, …, m; the thresholds for the adjacent two integers are: u (U) j +L j+1 <I j+1 -I j ，j＝1,…,m-1；
Vector T 1 The way in which one of the numbers t is converted to one of the m integer values is as follows:
if t is in the interval [ I ] j -L j ,I j +U j ]In, t is converted into I j ，j＝1,…,m；
If t is not at any one [ I ] j -L j ,I j +U j ]In the interval, j=1, …, m, then t cannot be converted.
4. A method of hiding and recovering sensitive data according to any one of claims 1-3, characterized by:
mapping model M at user A A In the training process of (2), besides the human body biological characteristic sample data of the user A, the mapping model is trained by using human body biological characteristic sample data generated by other users or artificially or automatically, and the human body biological characteristic sample data generated by other users or artificially or automatically is mapped into integer vectors different from the integer vector S; the human body biological characteristic sample data generated manually or automatically is data which is generated by using a program and has characteristics similar to human body biological characteristics but is not real human body biological characteristics; other users used by different users or human body biological feature sample data generated manually or automatically are the same or different.
5. A method of hiding and recovering sensitive data according to any one of claims 1-3, characterized by:
Mapping model M at user A A Input during training of (a)The data comprises control data besides human body biological characteristic sample data, wherein the control data is used for controlling the output result of the mapping model;
if the input data includes control data, and the mapping model M of the user A A In the training process of the system, besides the human body biological characteristic sample data of the user A, other human body biological characteristic sample data generated by other users or people or generated automatically are used for training the mapping model, so that integer vectors corresponding to the human body biological characteristic sample data mapping of other users or people or generated automatically in the training process are changed by changing the control data or are changed by changing the uncontrolled data.
6. The method for hiding and recovering sensitive data according to claim 5, wherein:
if the input data includes control data, the application mode of the control data includes:
when the data is hidden, the client side of the user A generates a plurality of integer vectors by using the user sensitive data, the integer vectors are used as mapping output data of the human body biological characteristics of the user A when the mapping model is trained, and different control data are input and correspond to different integer vectors to be output; calculating a hash value for a plurality of integer vectors, or calculating a hash value for each integer vector separately; if a hash value is calculated for a plurality of integer vectors, the human body biological characteristic data of the user A is utilized to pass through the mapping model M during data recovery A After all integer vectors are calculated and converted, comparing whether the hash value calculated before data hiding is the same with the hash value of all integer vectors obtained through mapping and conversion; if a hash value is calculated for each integer vector, the human body biological characteristic data of the user A is used to pass through the mapping model M during data recovery A And calculating and converting to obtain each integer vector, and comparing whether the hash value calculated before hiding the data is the same as the hash value of the integer vector obtained by mapping and converting.
7. A method of hiding and recovering sensitive data according to any one of claims 1-3, characterized by:
the input data of the mapping model also comprises user identification data of a user A, wherein the user identification data of the user A is obtained by conversion of user identification information of the user A, and the user identification information of the user A is objective information known by the user A; during the training of the mapping model and the data recovery, the user A inputs the user identification information through the user A client, and the user A client converts the user identification information into the user identification data for the training of the mapping model and the data recovery.
8. The method for hiding and recovering sensitive data as claimed in claim 4, wherein:
The input data of the mapping model also comprises user identification data of a user A, wherein the user identification data of the user A is obtained by conversion of user identification information of the user A, and the user identification information of the user A is objective information known by the user A; during the training of the mapping model and the data recovery, the user A inputs the user identification information through the user A client, and the user A client converts the user identification information into the user identification data for the training of the mapping model and the data recovery.
9. The method for hiding and recovering sensitive data as claimed in claim 5, wherein:
the input data of the mapping model also comprises user identification data of a user A, wherein the user identification data of the user A is obtained by conversion of user identification information of the user A, and the user identification information of the user A is objective information known by the user A; during the training of the mapping model and the data recovery, the user A inputs the user identification information through the user A client, and the user A client converts the user identification information into the user identification data for the training of the mapping model and the data recovery.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311348830.9A CN117539681A (en) | 2023-10-18 | 2023-10-18 | Sensitive data hiding and recovering method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311348830.9A CN117539681A (en) | 2023-10-18 | 2023-10-18 | Sensitive data hiding and recovering method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117539681A true CN117539681A (en) | 2024-02-09 |
Family
ID=89786974
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311348830.9A Pending CN117539681A (en) | 2023-10-18 | 2023-10-18 | Sensitive data hiding and recovering method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117539681A (en) |
-
2023
- 2023-10-18 CN CN202311348830.9A patent/CN117539681A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| RU2697638C1 (en) | Method and device for autonomous authentication of a person | |
| CN109587372B (en) | Invisible image steganography based on generation of countermeasure network | |
| CN104823203B (en) | Biometric templates safety and key generate | |
| CN106503655B (en) | A kind of electric endorsement method and sign test method based on face recognition technology | |
| CN106599872A (en) | Method and equipment for verifying living face images | |
| JP6524899B2 (en) | Secret data matching device, secret data matching program, and secret data matching method | |
| CN103929425B (en) | A kind of identity registration, identity authentication method, equipment and system | |
| CN111339897B (en) | Living body identification method, living body identification device, computer device, and storage medium | |
| CN112560753A (en) | Face recognition method, device and equipment based on feature fusion and storage medium | |
| Martínez et al. | Secure crypto-biometric system for cloud computing | |
| Sandhya et al. | Cancelable fingerprint cryptosystem using multiple spiral curves and fuzzy commitment scheme | |
| CN111598051A (en) | Face verification method, device and equipment and readable storage medium | |
| CN115795406A (en) | Reversible face anonymization processing system | |
| KR101077975B1 (en) | Method of generating fuzzy vault based on biometric information and verifying user's indentification using fuzzy vault | |
| TWI416922B (en) | Authentication system utilizing image authentication code and method thereof | |
| Yang et al. | A Delaunay triangle group based fuzzy vault with cancellability | |
| RU2007120788A (en) | METHOD FOR FORMING ELECTRONIC DOCUMENTS AND DEVICE FOR ITS IMPLEMENTATION | |
| CN104751042A (en) | Credibility detection method based on password hash and biometric feature recognition | |
| CN116886315A (en) | Authentication method based on biological characteristics and zero knowledge proof under web3.0 | |
| JP2019045969A (en) | Biological image processing apparatus, biological image processing method and biological image processing program | |
| CN117539681A (en) | Sensitive data hiding and recovering method | |
| CN115150150A (en) | Information recording method based on block chain technology | |
| KR100919486B1 (en) | Method for aligning concealed fingerprint data using partial geometric hashing, Method for authenticating fingerprint data using partial geometric hashing, Apparatus and System thereof | |
| CN111931148A (en) | Image processing method and device and electronic equipment | |
| Wang et al. | Template Protection based on Chaotic Map and DNA Encoding for Multimodal Biometrics at Feature Level Fusion. |
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 |