Two-way fingerprint location filtering and data security alignment method and system with hash commitment

Abstract

This invention discloses a data security alignment method and system for bidirectional fingerprint location filtering and hash commitment, relating to the fields of data privacy protection and federated learning. The method involves receiving raw data, calculating an ID hash value based on a preset hash algorithm, splitting the ID hash value to obtain a hash prefix and a hash suffix, and pre-calculating a hash commitment value and a random salt based on the hash suffix. A fingerprint is generated based on the hash prefix, and the hash prefix and fingerprint are input into a pre-established adaptive fingerprint location filter to identify potential shared hash prefixes. These potential shared hash prefixes are verified based on the hash commitment value and random salt to obtain the true shared hash prefix. Based on the true shared hash prefix, corresponding sample features are extracted from the raw data to obtain a shared sample feature set. This shared sample feature set is homomorphically encrypted to obtain an encrypted sample feature set. The encrypted sample feature sets are then merged using additive homomorphism to obtain the intersection result, thus achieving data privacy protection and alignment.

Description

Technical Field

[0001] This invention relates to the fields of data privacy protection and federated learning technology, specifically a data security alignment method and system based on bidirectional fingerprint location filtering and hash commitment. Background Technology

[0002] In the field of vertical federated learning, cross-institutional participants need to collaboratively build machine learning models while protecting data privacy. Each participant's data only contains partial feature dimensions of the same entity. Identifying shared data without disclosing the original data and conducting joint analysis based on this shared data is a core challenge in privacy protection. However, traditional privacy set intersection methods have revealed multiple technical limitations in practical applications, indicating shortcomings in privacy protection mechanisms.

[0003] Traditional methods that directly transmit raw data or simple hash values ​​are prone to leaking sensitive information during transmission, violating data privacy compliance requirements and potentially leading to user privacy breaches or commercial data misuse, failing to meet the core privacy requirements of vertical federated learning. Most traditional methods rely on a single filter for initial entity screening. While probabilistic mapping reduces storage overhead, the inherent nature of hash collisions often introduces a large number of redundant candidate entities, increasing the burden of subsequent encryption computation and feature alignment. Furthermore, the incorporation of false positive entities can reduce the feature consistency of model training, affecting model convergence speed and prediction accuracy. In scenarios reliant on model output, low-precision models can lead to inefficient customer outreach, making it difficult to guarantee the privacy and usability of the federated learning process. Improving accuracy requires significant new resources for feature selection, collision detection, and alignment optimization, increasing the risks of model training and business inference, weakening the credibility of cross-institutional data joint modeling, and failing to meet the security and accuracy requirements of practical engineering applications. Summary of the Invention

[0004] To address the shortcomings mentioned in the background section, the present invention aims to provide a data security alignment method and system for bidirectional fingerprint location filtering and hash commitment.

[0005] Firstly, the objective of this invention can be achieved through the following technical solution: a data security alignment method for bidirectional fingerprint location filtering and hash commitment, the method comprising the following steps: The system receives raw data, calculates the ID hash value based on a preset hash algorithm, splits the ID hash value to obtain a hash prefix and a hash suffix, and pre-calculates the hash commitment value and a random salt based on the hash suffix; wherein, the raw data includes the raw data of participant A and the raw data of participant B; Fingerprints are generated based on hash prefix extraction. The hash prefix and fingerprint are input into a pre-established adaptive fingerprint location filter to filter out potential shared hash prefixes. The potential shared hash prefixes are verified based on hash commitment values ​​and random salts to obtain the true shared hash prefixes. Based on the real common hash prefix, the corresponding sample features are extracted from the original data to obtain a common sample feature set. The common sample feature set is homomorphically encrypted to obtain an encrypted sample feature set. The encrypted sample feature set is then merged using additive homomorphism to obtain the intersection result, thus achieving data privacy protection and alignment.

[0006] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the participants A and B constructing their respective ID-information mapping tables and generating sample feature encoding mapping tables based on the original data of participant A and the original data of participant B, as follows: Participant A and Participant B each construct their own ID-information mapping data. Participant A's ID-information data is represented as {( , )}, ,in This represents the set of sample IDs for participant A. This represents the set of sample features of participant A. This represents the number of samples from participant A. Indicates the first Each sample ID, Indicates the first Each sample feature; the ID-information data of participant B is represented as {( , )}, ,in This represents the set of sample IDs for participant B. This represents the set of sample features of participant B. This indicates the number of samples from participant B. Indicates the first [unclear] of participant B Each sample ID, Indicates the first [unclear] of participant B Individual sample features; For the data of participant A, a sample feature encoding mapping table is generated using the following formula: in, This represents a mapping table from participant A's data information to numbers. This represents the data information corresponding to a single ID. This represents the set of data information corresponding to all IDs in participant A. This indicates a deduplication operation. A digital mapping representing the data information of participant A; For the data of participant B, a sample feature encoding mapping table is generated using the following formula: in, This represents a mapping table from participant B's data information to numbers. This represents the data information corresponding to a single ID. This represents the set of data information corresponding to all IDs in participant B. This indicates a deduplication operation. A digital mapping representing the data information of participant B.

[0007] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: when splitting the ID hash value to obtain the hash prefix and hash suffix, the first 16 bytes are truncated as the hash prefix and the last 16 bytes as the hash suffix, as follows: Participant A and Participant B respectively establish the correspondence between ID hash values ​​and data information, and calculate the hash value of their respective IDs using the SHA256 hash function. The formula is as follows: in, This represents the mapping between the hash prefix of participant A and the complete hash, and the sample characteristics. This represents the mapping between the hash prefix of participant B and the complete hash and sample characteristics; Indicates the first in participant A The hash prefix corresponding to each ID, Indicates the first in participant B The hash prefix corresponding to each ID, Indicates the first in participant A The hash suffix corresponding to each ID Indicates the first in participant B The hash suffix corresponding to each ID This represents the SHA-256 hash function, which outputs a hash value of length 256 bits. The process of pre-compiling the hash commitment value and random salt based on the hash suffix is ​​as follows: in, Indicates the first in participant A The commitment value corresponding to each ID, Indicates the first in participant B The commitment value corresponding to each ID, Indicates the first in participant A The random salt corresponding to each ID The first participant in Participant B The random salt corresponding to each ID is used to associate and store the hash prefixes of participant A and participant B with the hash commitment value and the random salt.

[0008] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: inputting the hash prefix and fingerprint into a pre-established adaptive fingerprint location filter, and performing preliminary verification by participant A and participant B using the adaptive fingerprint location filter constructed by participant B and the adaptive fingerprint location filter constructed by participant A, respectively, to filter out potential common ID hash prefixes.

[0009] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the participant A extracts a fingerprint from its own ID hash prefix and inserts it into an adaptive fingerprint location filter, as follows: Participant A sets the adaptive fingerprint location filter capacity based on its total number of IDs, and initializes the number of buckets for the adaptive fingerprint location filter using the following formula: in, The number of buckets for the adaptive fingerprint location filter constructed for participant A. The initial bucket size for the adaptive fingerprint location filter constructed for participant A. The filter capacity of the adaptive fingerprint location filter constructed for participant A; Participant A constructs an adaptive fingerprint location filter. This filter uses a fingerprint function to re-hash the hash prefix of all participant A's IDs, truncating a fixed length as the fingerprint. Bytes 4-11 are used as the compact identifier of the element. The formula is as follows: in, This represents the ID hash fingerprint of participant A; For each participant A's ID hash value, generate two candidate positions using SHA-256 hashing. and The first 4 bytes of the hash value are used to calculate the position of the first bucket. The position of the second bucket is generated by XORing the hash of the fingerprint with the position of the first bucket. The formula is as follows: in, for The first bucket position in the adaptive fingerprint location filter constructed by participant A. for At the second bucket position of the adaptive fingerprint location filter constructed by participant A, The number of buckets for the adaptive fingerprint location filter constructed for participant A; For two candidate buckets and Check for empty slots. If there are enough empty slots in the bucket, insert the fingerprint directly and return success. If the bucket's load factor is greater than the load factor threshold, attempt to expand the capacity using the following formula: in, This represents the bucket's load factor. This indicates that slots in the bucket are already in use. If the bucket's load factor is greater than the load factor threshold, the bucket size will be doubled, but will not exceed the maximum bucket size. If both candidate buckets are full and cannot be expanded, a kick-out mechanism is initiated. A bucket is randomly selected, one of the fingerprints is randomly replaced, the new fingerprint is stored, and the position of the kicked-out old fingerprint is remapped. This replacement process is repeated until insertion is successful or the maximum number of replacements is reached. The formula is: in, Indicates the new bucket position. Indicates the current position of the bucket. Indicates an old fingerprint; After constructing the adaptive fingerprint location filter, participant A encapsulates the query logic of the adaptive fingerprint location filter into an interface and sends it to participant B through a secure channel; Participant B extracts a fingerprint from its own ID hash prefix and inserts it into an adaptive fingerprint location filter, as follows: Participant B sets the adaptive fingerprint location filter capacity based on its total number of IDs, and initializes the number of buckets for the adaptive fingerprint location filter using the following formula: in, The number of buckets for the adaptive fingerprint location filter constructed for participant B. The initial bucket size for the adaptive fingerprint location filter constructed for participant B. The filter capacity of the adaptive fingerprint location filter constructed for participant B; Participant B constructs an adaptive fingerprint location filter. This filter uses a fingerprint function to re-hash the hash prefix of all of Participant B's IDs, then truncates the hash to a fixed length as the fingerprint, and uses bytes 4-11 as the compact identifier of the element. The formula is as follows: in, This represents the ID hash fingerprint of participant B; For the ID hash value of each participant B, two candidate positions are generated using SHB-256 hashing. and The first 4 bytes of the hash value are used to calculate the position of the first bucket. The position of the second bucket is generated by XORing the hash of the fingerprint with the position of the first bucket. The formula is as follows: in, for The first bucket position in the adaptive fingerprint location filter constructed by participant B. for At the second bucket position of the adaptive fingerprint location filter constructed by participant B, The number of buckets for the adaptive fingerprint location filter built for participant B; For two candidate buckets and Check for empty slots. If there are enough empty slots in the bucket, insert the fingerprint directly and return success. If the bucket's load factor is greater than the load factor threshold, attempt to expand the capacity using the following formula: in, This represents the bucket's load factor. This indicates that slots in the bucket are already in use. If the bucket's load factor is greater than the load factor threshold, the bucket size will be doubled, but will not exceed the maximum bucket size. If both candidate buckets are full and cannot be expanded, a kick-out mechanism is initiated. A bucket is randomly selected, one of the fingerprints is randomly replaced, the new fingerprint is stored, and the position of the kicked-out old fingerprint is remapped. This replacement process is repeated until insertion is successful or the maximum number of replacements is reached. The specific formula is: in, Indicates the new bucket position. Indicates the current position of the bucket. Indicates an old fingerprint; After constructing the adaptive fingerprint location filter, participant B encapsulates the query logic of the adaptive fingerprint location filter into an interface and sends it to participant A through a secure channel.

[0010] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the process of preliminary verification by participant A and participant B respectively using the adaptive fingerprint location filter constructed by participant B and the adaptive fingerprint location filter constructed by participant A, as follows: Participant A verifies whether its own ID hash prefix exists in the adaptive fingerprint location filter constructed by Participant B. Participant A uses the adaptive fingerprint location filter to calculate the fingerprint of its ID hash prefix, and takes the 4th to 11th bytes of the hashed ID hash prefix as the compact identifier of the element, using the following formula: in, This represents the ID hash prefix fingerprint of participant A; Participant A uses its own ID hash prefix fingerprint to calculate its two candidate bucket positions in the filter according to Participant B's rules. and The formula is: in, for The first bucket position, for The position of the second bucket; Verify the existence of a fingerprint; if participant A has constructed an adaptive fingerprint location filter... or Fingerprints found in the bucket This indicates that the fingerprint conforms to the fingerprint storage rules of participant B and is a fingerprint inserted by participant B. It is a shared hash prefix, meaning the adaptive fingerprint location filter determines the existence of this hash value. The formula is: in, express The verification results of the adaptive fingerprint location filter This represents the adaptive fingerprint location filter constructed by participant B. bucket, This represents the adaptive fingerprint location filter constructed by participant B. Bucket; if the adaptive fingerprint location filter verification result is true, it is a potential common hash prefix of participant A. ; Participant B uses the adaptive fingerprint location filter built by Participant A for preliminary verification, filtering potential common ID hash prefixes, as follows: in, This represents the ID hash fingerprint of participant B; Participant B uses its own ID hash fingerprint to calculate its two candidate bucket positions in the filter according to Participant A's rules. and The formula is: in, for The first bucket position, for The position of the second bucket; Verify the existence of a fingerprint; if participant A has constructed an adaptive fingerprint location filter... or Fingerprints found in the bucket This indicates that the fingerprint conforms to the fingerprint storage rules of participant A, and is a fingerprint inserted by participant A. It is a shared hash prefix, meaning the adaptive fingerprint location filter determines the existence of this hash value. The specific formula is: in, express The verification results of the adaptive fingerprint location filter This represents the adaptive fingerprint location filter constructed by participant A. bucket, This represents the adaptive fingerprint location filter constructed by participant A. If the bucket and adaptive fingerprint location filter verification result is true, then it is a potential common hash prefix of participant B. .

[0011] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the process of verifying the potential common hash prefix based on the hash commitment value and the random salt to obtain the true common hash prefix, including: Based on the verification results of the adaptive fingerprint location filter, participants A and B extract the hash commitment values ​​corresponding to all prefixes in the potential common hash prefixes, and form a mapping, as shown in the formula: in, It is the commitment value corresponding to the potential shared hash prefix of participant A. It is the commitment value corresponding to the potential shared hash prefix of participant B. This is a mapping between the potential shared hash prefix of participant A and the corresponding commitment value. Mapping of potential shared hash prefixes of participant B to their corresponding commitment values; Participant A sends a shared hash prefix to Participant B, requesting the corresponding salt value. Participant A uses its local hash suffix, the other party's salt value, and the hash commitment value to verify whether the hash suffixes are consistent. The formula is as follows: Ultimately, participant A obtains the shared ID hash prefix. .

[0012] Participant B sends the shared hash prefix to Participant A, requesting the corresponding salt value. Participant B uses its local hash suffix, the other party's salt value, and the hash commitment value to verify whether the hash suffixes are consistent. The formula is as follows: Ultimately, participant B obtains the shared ID hash prefix. .

[0013] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the process of homomorphically encrypting the common sample feature set to obtain the encrypted sample feature set, as follows: Participants A and B use the Paillier homomorphic public key of the aggregator to encrypt the feature data corresponding to the common hash and send it to the aggregator, as follows: Participant A obtains the sample features corresponding to the common ID hash based on the correspondence between ID hash and data information, and assembles the sample features into a set. Common sample feature set Encode and encrypt bit by bit, using the following formula: in, This represents the shared ID hash prefix calculated by participant A. This represents the set of shared sample features of participant A after encryption using a homomorphic encryption public key; Participant B obtains the sample features corresponding to the common ID hash based on the correspondence between ID hash and data information, and assembles the sample features into a set. Common sample feature set Encode and encrypt bit by bit, using the following formula: in, This represents the shared ID hash prefix calculated by participant B. This represents the set of shared sample features of participant B after encryption using a homomorphic encryption public key. This indicates that participant B has a shared set of sample features; participants A and B send the encrypted shared set of sample features to the aggregation end.

[0014] In conjunction with the first aspect, in some implementations of the first aspect, the method further includes: the process by which the aggregation end merges the encrypted sample feature sets using additive homomorphism to obtain the intersection result, as follows: The aggregation end uses Paillier additive homomorphism to combine ciphertexts, shifting and adding the ciphertexts of participant A and participant B, using the following formula: in, It is the modulus in the Paillier homomorphic encryption algorithm. The ciphertext represents the set of shared sample features after combination. The intersection result is obtained by jointly calculating the encrypted data based on the additive homomorphism of homomorphic encryption.

[0015] Secondly, in order to achieve the above objectives, this invention discloses a data security alignment system for bidirectional fingerprint location filtering and hash commitment, comprising: The hash splitting module is used to receive the original data, calculate the ID hash value based on a preset hash algorithm, split the ID hash value to obtain a hash prefix and a hash suffix, and pre-calculate the hash commitment value and random salt based on the hash suffix; wherein, the original data includes the original data of participant A and the original data of participant B; The hash verification module is used to extract and generate fingerprints based on hash prefixes. The hash prefix and fingerprint are input into a pre-established adaptive fingerprint location filter to filter out potential common hash prefixes. The potential common hash prefixes are verified based on hash commitment values ​​and random salts to obtain the true common hash prefix. The encryption merging module is used to extract corresponding sample features from the original data based on the real common hash prefix, obtain a common sample feature set, perform homomorphic encryption on the common sample feature set to obtain an encrypted sample feature set, and merge the encrypted sample feature set using additive homomorphism to obtain the intersection result, thereby achieving data privacy protection and alignment.

[0016] The beneficial effects of this invention are: This invention introduces a bidirectional adaptive fingerprint location filter as the core screening structure. It optimizes space utilization by dynamically adjusting bucket size and load factor threshold, and solves the fingerprint collision problem by combining a hash commitment-based secondary verification mechanism. Simultaneously, the efficient insertion and query characteristics of the adaptive fingerprint location filter improve the speed of shared ID screening. Combined with Paillier homomorphic encryption technology, it achieves joint computation of feature data in a encrypted state. Both parties only transmit encrypted feature information, without exposing the original ID and data content, fundamentally ensuring data privacy. For large-scale data scenarios, lightweight operations such as hash truncation, batch salt value interaction, and bit-by-bit encryption are used to optimize the computation logic, reducing redundant computation and network transmission, lowering dependence on hardware resources, and making it suitable for efficient collaborative computing in distributed environments. By introducing an efficient federated data alignment mechanism, it effectively ensures the privacy and availability of the federated learning process, accelerates model convergence, and improves prediction accuracy. In scenarios that rely on model output, high-precision models do not require significant additional resources for feature iteration optimization, reducing the risks of model training and business inference, enhancing the credibility of cross-institutional data joint modeling, and fully meeting the security and accuracy requirements of practical engineering applications. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Figure 1 This is a schematic diagram of the method flow of the present invention; Figure 2 This is a schematic diagram of the workflow of the present invention; Figure 3 This is a schematic diagram illustrating the construction process of the adaptive fingerprint location filter of the present invention; Figure 4 This is a schematic diagram of the system structure of the present invention; Figure 5 This is a statistical chart showing the number of ID matches in this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1: like Figure 1As shown, the data security alignment method of bidirectional fingerprint location filtering and hash commitment includes the following steps: S101: Receive raw data, calculate the ID hash value based on a preset hash algorithm, split the ID hash value to obtain a hash prefix and a hash suffix, and pre-calculate the hash commitment value and random salt based on the hash suffix; wherein, the raw data includes the raw data of participant A and the raw data of participant B; The preprocessing process for the raw data is as follows: The preprocessing of the participants' raw data includes the key initialization stage, the ID-information mapping construction stage, and the hashing stage.

[0020] Furthermore, the aggregator generates Paillier homomorphic key pairs and distributes public keys to participants A and B, as follows: The aggregation end generates a homomorphic encrypted public-private key pair, using the following formula: in, This represents a homomorphic encryption public / private key pair. Indicates safety parameters, This represents the key generation function for the Paillier homomorphic encryption algorithm; the aggregator will use the homomorphic encryption public key. Send to participant A and participant B.

[0021] Furthermore, participants A and B construct their respective ID-information mapping tables and generate sample feature encoding mapping tables based on the original data.

[0022] Participant A and Participant B construct their respective ID-information mapping tables and generate sample feature encoding mapping tables based on the original data of Participant A and Participant B, as follows: Participant A and Participant B each construct their own ID-information mapping data. Participant A's ID-information data is represented as {( , )}, ,in This represents the set of sample IDs for participant A. This represents the set of sample features of participant A. This represents the number of samples from participant A. Indicates the first Each sample ID, Indicates the first Each sample feature; the ID-information data of participant B is represented as {( , )}, ,in This represents the set of sample IDs for participant B. This represents the set of sample features of participant B. This indicates the number of samples from participant B. Indicates the first [unclear] of participant B Each sample ID, Indicates the first [unclear] of participant B Individual sample features; For the data of participant A, a sample feature encoding mapping table is generated, and the specific formula is as follows: in, This represents a mapping table from participant A's data information to numbers. This represents the data information corresponding to a single ID. This represents the set of data information corresponding to all IDs in participant A. This indicates a deduplication operation. A digital mapping representing the data information of participant A.

[0023] For the data of participant B, a sample feature encoding mapping table is generated, and the specific formula is as follows: in, This represents a mapping table from participant B's data information to numbers. This represents the data information corresponding to a single ID. This represents the set of data information corresponding to all IDs in participant B. This indicates a deduplication operation. A digital mapping representing the data information of participant B.

[0024] When splitting the ID hash value to obtain the hash prefix and hash suffix, the first 16 bytes are taken as the hash prefix and the last 16 bytes as the hash suffix. The process is as follows: Participant A and Participant B generate their respective ID hash values, extracting the first 16 bytes as the hash prefix and the last 16 bytes as the hash suffix. Simultaneously, a mapping is constructed between the hash prefix and the complete hash, as well as the sample features. Furthermore, a hash commitment is pre-calculated based on the hash suffix, using the following formula: in, This represents the mapping between the hash prefix of participant A and the complete hash, and the sample characteristics. This represents the mapping between the hash prefix of participant B and the complete hash and sample characteristics; Indicates the first in participant A The hash prefix corresponding to each ID, Indicates the first in participant B The hash prefix corresponding to each ID, Indicates the first in participant A The hash suffix corresponding to each ID Indicates the first in participant B The hash suffix corresponding to each ID This represents the SHA-256 hash function, which outputs a hash value of length 256 bits.

[0025] The process of pre-compiling the hash commitment value and random salt based on the hash suffix is ​​as follows: Generate a commitment value and a random salt for the hash suffixes of participant A and participant B, using the following formula: in, Indicates the first in participant A The commitment value corresponding to each ID, Indicates the first in participant B The commitment value corresponding to each ID, Indicates the first in participant A The random salt corresponding to each ID The first participant in Participant B The random salt corresponding to each ID is used to associate and store the hash prefixes of participant A and participant B with their commitment values ​​and random salts.

[0026] S102: Generate fingerprints based on hash prefix extraction, input the hash prefix and fingerprint into a pre-established adaptive fingerprint location filter, filter out potential common hash prefixes, verify the potential common hash prefixes based on hash commitment value and random salt, and obtain the true common hash prefix; The hash prefix and fingerprint are input into a pre-built adaptive fingerprint location filter. Participant A and Participant B use the adaptive fingerprint location filter built by Participant B and the adaptive fingerprint location filter built by Participant A, respectively, to perform preliminary verification and filter out potential common ID hash prefixes.

[0027] Participant A extracts its own ID hash prefix into a fingerprint and inserts it into an adaptive fingerprint location filter, as follows: Participant A sets the adaptive fingerprint location filter capacity based on its total number of IDs, and initializes the number of buckets for the adaptive fingerprint location filter. The specific formula is as follows: in, The number of buckets for the adaptive fingerprint location filter constructed for participant A. The initial bucket size for the adaptive fingerprint location filter constructed for participant A. The filter capacity of the adaptive fingerprint location filter constructed for participant A; Participant A constructs an adaptive fingerprint location filter. This filter uses a fingerprint function to re-hash the hash prefix of all participant A's IDs, truncating a fixed length as the fingerprint. Bytes 4-11 are used as the compact identifier of the element. The specific formula is as follows: in, This represents the ID hash fingerprint of participant A; For each participant A's ID hash value, generate two candidate positions using SHA-256 hashing. and The first 4 bytes of the hash value are used to calculate the position of the first bucket. The position of the second bucket is generated by XORing the hash of the fingerprint with the position of the first bucket to ensure position correlation. The specific formula is as follows: in, for The first bucket position in the adaptive fingerprint location filter constructed by participant A. for At the second bucket position of the adaptive fingerprint location filter constructed by participant A, The number of buckets for the adaptive fingerprint location filter constructed for participant A; For two candidate buckets and Check for empty slots. If there are enough empty slots in the bucket, insert the fingerprint directly and return success. If the bucket's load factor is greater than the load factor threshold, attempt to expand the capacity. The specific formula is as follows: in, This represents the bucket's load factor. This indicates that slots in the bucket are in use. If the bucket's load factor is greater than the load factor threshold, the bucket size is doubled, but not exceeding the maximum bucket size.

[0028] If both candidate buckets are full and cannot be expanded, a kick-out mechanism is initiated. A bucket is randomly selected, one of the fingerprints is randomly replaced, the new fingerprint is stored, and the position of the kicked-out old fingerprint is remapped. This replacement process is repeated until insertion is successful or the maximum number of replacements is reached. The specific formula is: in, Indicates the new bucket position. Indicates the current position of the bucket. Indicates an old fingerprint; After constructing the adaptive fingerprint location filter, participant A encapsulates the query logic of the adaptive fingerprint location filter into an interface and sends it to participant B through a secure channel.

[0029] Participant B sets the adaptive fingerprint location filter capacity based on its total number of IDs, and initializes the number of buckets for the adaptive fingerprint location filter. The specific formula is as follows: in, The number of buckets for the adaptive fingerprint location filter constructed for participant B. The initial bucket size for the adaptive fingerprint location filter constructed for participant B. The filter capacity of the adaptive fingerprint location filter constructed for participant B; Participant B constructs an adaptive fingerprint location filter. This filter uses a fingerprint function to re-hash the hash prefix of all participant B's IDs, then truncates the hash to a fixed length as the fingerprint, taking bytes 4-11 as the compact identifier of the element. The specific formula is as follows: in, This represents the ID hash fingerprint of participant B; For the ID hash value of each participant B, two candidate positions are generated using SHB-256 hashing. and The first 4 bytes of the hash value are used to calculate the position of the first bucket. The position of the second bucket is generated by XORing the hash of the fingerprint with the position of the first bucket to ensure position correlation. The specific formula is as follows: in, for The first bucket position in the adaptive fingerprint location filter constructed by participant B. for At the second bucket position of the adaptive fingerprint location filter constructed by participant B, The number of buckets for the adaptive fingerprint location filter built for participant B; For two candidate buckets and Check for empty slots. If there are enough empty slots in the bucket, insert the fingerprint directly and return success. If the bucket's load factor is greater than the load factor threshold, attempt to expand the capacity. The specific formula is as follows: in, This represents the bucket's load factor. This indicates that slots in the bucket are in use. If the bucket's load factor is greater than the load factor threshold, the bucket size is doubled, but not exceeding the maximum bucket size.

[0030] If both candidate buckets are full and cannot be expanded, a kick-out mechanism is initiated. A bucket is randomly selected, one of the fingerprints is randomly replaced, the new fingerprint is stored, and the position of the kicked-out old fingerprint is remapped. This replacement process is repeated until insertion is successful or the maximum number of replacements is reached. The specific formula is: in, Indicates the new bucket position. Indicates the current position of the bucket. Indicates an old fingerprint; After constructing the adaptive fingerprint location filter, participant B encapsulates the query logic of the adaptive fingerprint location filter into an interface and sends it to participant A through a secure channel.

[0031] The process of preliminary verification by participant A and participant B using the adaptive fingerprint location filter constructed by participant B and participant A respectively is as follows: Participant A verifies whether its own ID hash prefix exists in the adaptive fingerprint location filter constructed by Participant B. First, Participant A uses the adaptive fingerprint location filter to calculate the fingerprint of its ID hash prefix. The 4th to 11th bytes of Participant A's ID hash prefix are taken as the compact identifier of the element. The specific formula is as follows: in, This represents the ID hash prefix fingerprint of participant A; Participant A uses its own ID hash prefix fingerprint to calculate its two candidate bucket positions in the filter according to Participant B's rules. and The specific formula is as follows: in, for The first bucket position, for The position of the second bucket; Verify the existence of a fingerprint; if participant A has constructed an adaptive fingerprint location filter... or Fingerprints found in the bucket This indicates that the fingerprint conforms to the fingerprint storage rules of participant B and is a fingerprint inserted by participant B. It is a shared hash prefix, meaning the adaptive fingerprint location filter determines the existence of this hash value. The specific formula is: in, express The verification results of the adaptive fingerprint location filter This represents the adaptive fingerprint location filter constructed by participant B. bucket, This represents the adaptive fingerprint location filter constructed by participant B. Bucket; if the adaptive fingerprint location filter verification result is true, it is a potential common hash prefix of participant A. ; Participant B uses the adaptive fingerprint location filter constructed by Participant A for preliminary verification to filter potential common ID hash prefixes, as follows: in, This represents the ID hash fingerprint of participant B; Participant B uses its own ID hash fingerprint to calculate its two candidate bucket positions in the filter according to Participant A's rules. and The specific formula is as follows: in, for The first bucket position, for The position of the second bucket; Verify the existence of a fingerprint; if participant A has constructed an adaptive fingerprint location filter... or Fingerprints found in the bucket This indicates that the fingerprint conforms to the fingerprint storage rules of participant A, and is a fingerprint inserted by participant A. It is a shared hash prefix, meaning the adaptive fingerprint location filter determines the existence of this hash value. The specific formula is: in, express The verification results of the adaptive fingerprint location filter This represents the adaptive fingerprint location filter constructed by participant A. bucket, This represents the adaptive fingerprint location filter constructed by participant A. Bucket. If the adaptive fingerprint location filter verification result is true, it represents the potential common hash prefix of participant B. .

[0032] The process of verifying potential shared hash prefixes based on hash commitment values ​​and random salts to obtain the true shared hash prefix includes: Based on the verification results of the adaptive fingerprint location filter, participants A and B extract the commitment values ​​corresponding to all prefixes in the potential common hash prefix and form a mapping, the specific formula of which is: in, It is the commitment value corresponding to the potential shared hash prefix of participant A. It is the commitment value corresponding to the potential shared hash prefix of participant B. This is a mapping between the potential shared hash prefix of participant A and the corresponding commitment value. Mapping of potential shared hash prefixes of participant B to their corresponding commitment values; Participant A sends a shared hash prefix to Participant B, requesting the corresponding salt value. Participant A uses its local hash suffix, the other party's salt value, and the commitment value to verify whether the hash suffixes are consistent. The specific formula is as follows: Ultimately, participant A obtains the shared ID hash prefix. .

[0033] Participant B sends the shared hash prefix to Participant A, requesting the corresponding salt value. Participant B uses its local hash suffix, the other party's salt value, and the commitment value to verify whether the hash suffixes are consistent. The specific formula is as follows: Ultimately, participant B obtains the shared ID hash prefix. .

[0034] S103: Extract corresponding sample features from the original data based on the real common hash prefix to obtain a common sample feature set. Perform homomorphic encryption on the common sample feature set to obtain an encrypted sample feature set. Merge the encrypted sample feature sets using additive homomorphism to obtain the intersection result, thereby achieving data privacy protection and alignment.

[0035] The process of homomorphically encrypting the common sample feature set to obtain the encrypted sample feature set is as follows: Participant A obtains the sample features corresponding to the common ID hash based on the correspondence between ID hash and data information, and assembles the sample features into a set. Common sample feature set Encode and encrypt bit by bit, using the following formula: in, This represents the shared ID hash prefix calculated by participant A. This represents the set of shared sample features of participant A after encryption using a homomorphic encryption public key; Participant B obtains the sample features corresponding to the common ID hash based on the correspondence between ID hash and data information, and assembles the sample features into a set. Common sample feature set Encode and encrypt bit by bit, using the following formula: in, This represents the shared ID hash prefix calculated by participant B. This represents the set of shared sample features of participant B after encryption using a homomorphic encryption public key. This indicates that participant B has a shared set of sample features; participants A and B send the encrypted shared set of sample features to the aggregation end.

[0036] The ciphertexts are merged at the aggregation end to obtain the intersection result.

[0037] Furthermore, the aggregation end utilizes Paillier's additive homomorphic combination ciphertext, performing homomorphic addition on the two ciphertexts. By shifting and adding the shared sample feature sets of participants A and B, the final ciphertext of the shared sample feature set is obtained. The specific formula is as follows: in, It is the modulus in the Paillier homomorphic encryption algorithm. This represents the encrypted set of shared sample features after combination.

[0038] Specifically, the present invention will be further illustrated below through embodiments: This embodiment provides a data privacy protection and alignment method based on bidirectional adaptive fingerprint location filter and hash commitment. To verify the beneficial effects of the present invention, a simulation experiment is conducted for scientific demonstration.

[0039] Set adaptive fingerprint location filter replacement count The initial bucket size is 2, the maximum bucket size is 8, and the load factor threshold is 0.9. Participants A and B each have 100 data entries in their datasets, some of which are shown in Tables 1 and 2 below. Table 1 Data set of Participant A Table 2 Data set of participant B By using the data security alignment method of bidirectional fingerprint location filtering and hash commitment proposed in this scheme, the intersection of privacy sets was performed, resulting in a total of 20 common data items. Some of the privacy set intersection results are shown in Table 3 below: Table 3. Results of Intersection of Privacy Sets A bidirectional adaptive fingerprint location filter is introduced as the core screening structure, combined with a hash commitment-based secondary verification mechanism. This avoids exposing the original ID or complete hash value, preventing the leakage of set features. The joint computation phase is achieved through ciphertext addition, with no interaction of original data or plaintext features throughout the process, thus achieving end-to-end privacy protection and effectively improving the security of privacy set intersection. Fingerprint extraction, location calculation, and other operations are all lightweight computations, reducing communication overhead. A comparison with existing technologies is shown in Table 4 below. Table 4 Comparison of Existing Technologies Example 2: To achieve the above objective, such as Figure 4 As shown, based on Embodiment 1, this invention discloses a data security alignment system for bidirectional fingerprint location filtering and hash commitment, comprising: The hash splitting module 11 is used to receive the original data, calculate the ID hash value based on a preset hash algorithm, split the ID hash value to obtain a hash prefix and a hash suffix, and pre-calculate the hash commitment value and random salt based on the hash suffix; wherein, the original data includes the original data of participant A and the original data of participant B; The hash verification module 12 is used to extract and generate fingerprints based on hash prefixes. It inputs the hash prefix and fingerprint into a pre-established adaptive fingerprint location filter to filter out potential common hash prefixes. It then verifies the potential common hash prefixes based on the hash commitment value and random salt to obtain the true common hash prefix. The encryption merging module 13 is used to extract corresponding sample features from the original data based on the real common hash prefix, obtain a common sample feature set, perform homomorphic encryption on the common sample feature set to obtain an encrypted sample feature set, and merge the encrypted sample feature set using additive homomorphism to obtain the intersection result, thereby achieving data privacy protection and alignment.

[0040] Based on the same inventive concept, this invention also provides a computer device, comprising: one or more processors, and a memory for storing one or more computer programs; the programs include program instructions, and the processor executes the program instructions stored in the memory. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, used to implement one or more instructions, specifically for loading and executing one or more instructions stored in a computer storage medium to implement the above-described method.

[0041] It should be further explained that, based on the same inventive concept, the present invention also provides a computer storage medium storing a computer program, which, when executed by a processor, performs the above-described method. This storage medium can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, the computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0042] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0043] The foregoing has shown and described the basic principles, main features, and advantages of this disclosure. Those skilled in the art should understand that this disclosure is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of this disclosure. Various changes and modifications can be made to this disclosure without departing from its spirit and scope, and all such changes and modifications fall within the scope of this disclosure as claimed.

Claims

1. A method of data security alignment with bidirectional fingerprint location filtering and hash commitment, characterized in that, The method includes the following steps: The system receives raw data, calculates the ID hash value based on a preset hash algorithm, splits the ID hash value to obtain a hash prefix and a hash suffix, and pre-calculates the hash commitment value and a random salt based on the hash suffix; wherein, the raw data includes the raw data of participant A and the raw data of participant B; Fingerprints are generated based on hash prefix extraction. The hash prefix and fingerprint are input into a pre-established adaptive fingerprint location filter to filter out potential shared hash prefixes. The potential shared hash prefixes are verified based on hash commitment values ​​and random salts to obtain the true shared hash prefixes. Based on the real common hash prefix, the corresponding sample features are extracted from the original data to obtain a common sample feature set. The common sample feature set is homomorphically encrypted to obtain an encrypted sample feature set. The encrypted sample feature set is then merged using additive homomorphism to obtain the intersection result, thus achieving data privacy protection and alignment.

2. The method of claim 1, wherein, The process is as follows: Participant A and Participant B construct their respective ID-information mapping tables and generate sample feature encoding mapping tables based on the original data of Participant A and Participant B: Participant A and Participant B each construct their own ID-information mapping data. Participant A's ID-information data is represented as {( , )}, ,in This represents the set of sample IDs for participant A. This represents the set of sample features of participant A. This represents the number of samples from participant A. Indicates the first Each sample ID, Indicates the first Each sample feature; the ID-information data of participant B is represented as {( , )}, ,in This represents the set of sample IDs for participant B. This represents the set of sample features of participant B. This indicates the number of samples from participant B. Indicates the first [unclear] of participant B Each sample ID, Indicates the first [unclear] of participant B Individual sample features; For the data of participant A, a sample feature encoding mapping table is generated using the following formula: in, This represents a mapping table from participant A's data information to numbers. This represents the data information corresponding to a single ID. This represents the set of data information corresponding to all IDs in participant A. This indicates a deduplication operation. A digital mapping representing the data information of participant A; For the data of participant B, a sample feature encoding mapping table is generated using the following formula: in, This represents a mapping table from participant B's data information to numbers. This represents the data information corresponding to a single ID. This represents the set of data information corresponding to all IDs in participant B. This indicates a deduplication operation. A digital mapping representing the data information of participant B.

3. The data security alignment method for bidirectional fingerprint location filtering and hash commitment according to claim 1, characterized in that, When splitting the ID hash value to obtain the hash prefix and hash suffix, the first 16 bytes are taken as the hash prefix and the last 16 bytes as the hash suffix. The process is as follows: Participant A and Participant B respectively establish the correspondence between ID hash values ​​and data information, and calculate the hash value of their respective IDs using the SHA256 hash function. The formula is as follows: in, This represents the mapping between the hash prefix of participant A and the complete hash, and the sample characteristics. This represents the mapping between the hash prefix of participant B and the complete hash and sample characteristics; Indicates the first in participant A The hash prefix corresponding to each ID, Indicates the first in participant B The hash prefix corresponding to each ID, Indicates the first in participant A The hash suffix corresponding to each ID Indicates the first in participant B The hash suffix corresponding to each ID This represents the SHA-256 hash function, which outputs a hash value of length 256 bits. The process of pre-compiling the hash commitment value and random salt based on the hash suffix is ​​as follows: in, Indicates the first in participant A The commitment value corresponding to each ID, Indicates the first in participant B The commitment value corresponding to each ID, Indicates the first in participant A The random salt corresponding to each ID The first participant in Participant B The random salt corresponding to each ID is used to associate and store the hash prefixes of participant A and participant B with the hash commitment value and the random salt.

4. The data security alignment method for bidirectional fingerprint location filtering and hash commitment according to claim 1, characterized in that, The hash prefix and fingerprint are input into a pre-established adaptive fingerprint location filter. Participant A and Participant B respectively use the adaptive fingerprint location filter built by Participant B and the adaptive fingerprint location filter built by Participant A to perform preliminary verification and filter out potential common ID hash prefixes.

5. The data security alignment method for bidirectional fingerprint location filtering and hash commitment according to claim 4, characterized in that, Participant A extracts a fingerprint from its own ID hash prefix and inserts it into an adaptive fingerprint location filter, as follows: Participant A sets the adaptive fingerprint location filter capacity based on its total number of IDs, and initializes the number of buckets for the adaptive fingerprint location filter using the following formula: in, The number of buckets for the adaptive fingerprint location filter constructed for participant A. The initial bucket size for the adaptive fingerprint location filter constructed for participant A. The filter capacity of the adaptive fingerprint location filter constructed for participant A; Participant A constructs an adaptive fingerprint location filter. This filter uses a fingerprint function to re-hash the hash prefix of all participant A's IDs, truncating a fixed length as the fingerprint. Bytes 4-11 are used as the compact identifier of the element. The formula is as follows: in, This represents the ID hash fingerprint of participant A; For each participant A's ID hash value, generate two candidate positions using SHA-256 hashing. and The first 4 bytes of the hash value are used to calculate the position of the first bucket. The position of the second bucket is generated by XORing the hash of the fingerprint with the position of the first bucket. The formula is as follows: in, for The first bucket position in the adaptive fingerprint location filter constructed by participant A. for At the second bucket position of the adaptive fingerprint location filter constructed by participant A, The number of buckets for the adaptive fingerprint location filter constructed for participant A; For two candidate buckets and Check for empty slots. If there are enough empty slots in the bucket, insert the fingerprint directly and return success. If the bucket's load factor is greater than the load factor threshold, attempt to expand the capacity using the following formula: in, This represents the bucket's load factor. This indicates that slots in the bucket are already in use. If the bucket's load factor is greater than the load factor threshold, the bucket size will be doubled, but will not exceed the maximum bucket size. If both candidate buckets are full and cannot be expanded, a kick-out mechanism is initiated. A bucket is randomly selected, one of the fingerprints is randomly replaced, the new fingerprint is stored, and the position of the kicked-out old fingerprint is remapped. This replacement process is repeated until insertion is successful or the maximum number of replacements is reached. The formula is: in, Indicates the new bucket position. Indicates the current position of the bucket. Indicates an old fingerprint; After constructing the adaptive fingerprint location filter, participant A encapsulates the query logic of the adaptive fingerprint location filter into an interface and sends it to participant B through a secure channel; Participant B extracts a fingerprint from its own ID hash prefix and inserts it into an adaptive fingerprint location filter, as follows: Participant B sets the adaptive fingerprint location filter capacity based on its total number of IDs, and initializes the number of buckets for the adaptive fingerprint location filter using the following formula: in, The number of buckets for the adaptive fingerprint location filter constructed for participant B. The initial bucket size for the adaptive fingerprint location filter constructed for participant B. The filter capacity of the adaptive fingerprint location filter constructed for participant B; Participant B constructs an adaptive fingerprint location filter. This filter uses a fingerprint function to re-hash the hash prefix of all of Participant B's IDs, then truncates the hash to a fixed length as the fingerprint, and uses bytes 4-11 as the compact identifier of the element. The formula is as follows: in, This represents the ID hash fingerprint of participant B; For the ID hash value of each participant B, two candidate positions are generated using SHB-256 hashing. and The first 4 bytes of the hash value are used to calculate the position of the first bucket. The position of the second bucket is generated by XORing the hash of the fingerprint with the position of the first bucket. The formula is as follows: in, for The first bucket position in the adaptive fingerprint location filter constructed by participant B. for At the second bucket position of the adaptive fingerprint location filter constructed by participant B, The number of buckets for the adaptive fingerprint location filter built for participant B; For two candidate buckets and Check for empty slots. If there are enough empty slots in the bucket, insert the fingerprint directly and return success. If the bucket's load factor is greater than the load factor threshold, attempt to expand the capacity using the following formula: in, This represents the bucket's load factor. This indicates that slots in the bucket are already in use. If the bucket's load factor is greater than the load factor threshold, the bucket size will be doubled, but will not exceed the maximum bucket size. If both candidate buckets are full and cannot be expanded, a kick-out mechanism is initiated. A bucket is randomly selected, one of the fingerprints is randomly replaced, the new fingerprint is stored, and the position of the kicked-out old fingerprint is remapped. This replacement process is repeated until insertion is successful or the maximum number of replacements is reached. The specific formula is: in, Indicates the new bucket position. Indicates the current position of the bucket. Indicates an old fingerprint; After constructing the adaptive fingerprint location filter, participant B encapsulates the query logic of the adaptive fingerprint location filter into an interface and sends it to participant A through a secure channel.

6. The data security alignment method for bidirectional fingerprint location filtering and hash commitment according to claim 5, characterized in that, The process of preliminary verification by participant A and participant B using the adaptive fingerprint location filter constructed by participant B and the adaptive fingerprint location filter constructed by participant A, respectively, is as follows: Participant A verifies whether its own ID hash prefix exists in the adaptive fingerprint location filter constructed by Participant B. Participant A uses the adaptive fingerprint location filter to calculate the fingerprint of its ID hash prefix, and takes the 4th to 11th bytes of the hashed ID hash prefix as the compact identifier of the element, using the following formula: in, This represents the ID hash prefix fingerprint of participant A; Participant A uses its own ID hash prefix fingerprint to calculate its two candidate bucket positions in the filter according to Participant B's rules. and The formula is: in, for The first bucket position, for The position of the second bucket; Verify the existence of a fingerprint; if participant A has constructed an adaptive fingerprint location filter... or Fingerprints found in the bucket This indicates that the fingerprint conforms to the fingerprint storage rules of participant B and is a fingerprint inserted by participant B. It is a shared hash prefix, meaning the adaptive fingerprint location filter determines the existence of this hash value. The formula is: in, express The verification results of the adaptive fingerprint location filter This represents the adaptive fingerprint location filter constructed by participant B. bucket, This represents the adaptive fingerprint location filter constructed by participant B. Bucket; if the adaptive fingerprint location filter verification result is true, it is a potential common hash prefix of participant A. ; Participant B uses the adaptive fingerprint location filter built by Participant A for preliminary verification, filtering potential common ID hash prefixes, as follows: in, This represents the ID hash fingerprint of participant B; Participant B uses its own ID hash fingerprint to calculate its two candidate bucket positions in the filter according to Participant A's rules. and The formula is: in, for The first bucket position, for The position of the second bucket; Verify the existence of a fingerprint; if participant A has constructed an adaptive fingerprint location filter... or Fingerprints found in the bucket This indicates that the fingerprint conforms to the fingerprint storage rules of participant A, and is a fingerprint inserted by participant A. It is a shared hash prefix, meaning the adaptive fingerprint location filter determines the existence of this hash value. The specific formula is: in, express The verification results of the adaptive fingerprint location filter This represents the adaptive fingerprint location filter constructed by participant A. bucket, This represents the adaptive fingerprint location filter constructed by participant A. If the bucket and adaptive fingerprint location filter verification result is true, then it is a potential common hash prefix of participant B. .

7. The data security alignment method for bidirectional fingerprint location filtering and hash commitment according to claim 6, characterized in that, The process of verifying potential shared hash prefixes based on hash commitment values ​​and random salts to obtain true shared hash prefixes includes: Based on the verification results of the adaptive fingerprint location filter, participants A and B extract the hash commitment values ​​corresponding to all prefixes in the potential common hash prefixes, and form a mapping, as shown in the formula: in, It is the commitment value corresponding to the potential shared hash prefix of participant A. It is the commitment value corresponding to the potential shared hash prefix of participant B. This is a mapping between the potential shared hash prefix of participant A and the corresponding commitment value. Mapping of potential shared hash prefixes of participant B to their corresponding commitment values; Participant A sends a shared hash prefix to Participant B, requesting the corresponding salt value. Participant A uses its local hash suffix, the other party's salt value, and the hash commitment value to verify whether the hash suffixes are consistent. The formula is as follows: Ultimately, participant A obtains the shared ID hash prefix. . Participant B sends the shared hash prefix to Participant A, requesting the corresponding salt value. Participant B uses its local hash suffix, the other party's salt value, and the hash commitment value to verify whether the hash suffixes are consistent. The formula is as follows: Ultimately, participant B obtains the shared ID hash prefix. .

8. The data security alignment method for bidirectional fingerprint location filtering and hash commitment according to claim 1, characterized in that, The process of performing homomorphic encryption on the common sample feature set to obtain the encrypted sample feature set is as follows: Participants A and B use the Paillier homomorphic public key of the aggregator to encrypt the feature data corresponding to the common hash and send it to the aggregator, as follows: Participant A obtains the sample features corresponding to the common ID hash based on the correspondence between ID hash and data information, and assembles the sample features into a set. Common sample feature set Encode and encrypt bit by bit, using the following formula: in, This represents the shared ID hash prefix calculated by participant A. This represents the set of shared sample features of participant A after encryption using a homomorphic encryption public key; Participant B obtains the sample features corresponding to the common ID hash based on the correspondence between ID hash and data information, and assembles the sample features into a set. Common sample feature set Encode and encrypt bit by bit, using the following formula: in, This represents the shared ID hash prefix calculated by participant B. This represents the set of shared sample features of participant B after encryption using a homomorphic encryption public key. This indicates that participant B has a shared set of sample features; participants A and B send the encrypted shared set of sample features to the aggregation end.

9. The data security alignment method for bidirectional fingerprint location filtering and hash commitment according to claim 8, characterized in that, The process by which the aggregation end merges the encrypted sample feature sets using additive homomorphism to obtain the intersection result is as follows: The aggregation end uses Paillier additive homomorphism to combine ciphertexts, shifting and adding the ciphertexts of participant A and participant B, using the following formula: in, It is the modulus in the Paillier homomorphic encryption algorithm. The ciphertext represents the set of shared sample features after combination. The intersection result is obtained by jointly calculating the encrypted data based on the additive homomorphism of homomorphic encryption.

10. A data security alignment system for bidirectional fingerprint location filtering and hash commitment, employing the data security alignment method for bidirectional fingerprint location filtering and hash commitment as described in any one of claims 1 to 9, characterized in that, include: The hash splitting module is used to receive the original data, calculate the ID hash value based on a preset hash algorithm, split the ID hash value to obtain a hash prefix and a hash suffix, and pre-calculate the hash commitment value and random salt based on the hash suffix; wherein, the original data includes the original data of participant A and the original data of participant B; The hash verification module is used to extract and generate fingerprints based on hash prefixes. The hash prefix and fingerprint are input into a pre-established adaptive fingerprint location filter to filter out potential common hash prefixes. The potential common hash prefixes are verified based on hash commitment values ​​and random salts to obtain the true common hash prefix. The encryption merging module is used to extract corresponding sample features from the original data based on the real common hash prefix, obtain a common sample feature set, perform homomorphic encryption on the common sample feature set to obtain an encrypted sample feature set, and merge the encrypted sample feature set using additive homomorphism to obtain the intersection result, thereby achieving data privacy protection and alignment.

Images