Identification based on encrypted biometric data
View Patent ↗A database comprising biometric data stored in encrypted form is managed by a management unit. It comprises a set of filters respectively associated with filter identifiers. A biometric data item is received at a management unit; next, said biometric data item is stored in an encrypted form at a given address in the database. Then keywords are obtained on the basis of a first set of hash functions and of the biometric data item. A subset of indexing filters is associated with each keyword by selecting, for each keyword, filters as a function of the respectively associated filter identifiers, of said keywords, and of a second set of hash functions; and the given address is associated with each of the filters of the subset of filters.
1. A management method for storing and indexing biometric data in a database in encrypted form, said database comprising a set of filters (B1, . . . , Bm) respectively associated with filter identifiers; wherein said management method comprises the following steps at a management unit:
/1/ receiving a biometric data item;
/2/ storing, at a given address in the database, said biometric data item in encrypted form;
/3/ obtaining keywords based on a first set of hash functions and on said biometric data item;
/4/ associating a subset of indexing filters with each keyword by selecting filters for each keyword as a function of the respectively associated filter identifiers, of said keywords, and of a second set of hash functions; and
/5/ associating said given address in encrypted form with each of the filters in the subset of indexing filters.
2. The management method of claim 1 , wherein step /3/ is implemented by the following steps:
obtaining respective first hash values by applying the first set of hash functions to the biometric data item;
obtaining said keywords by respectively combining said first hash values with respective identifiers of the hash functions in the first set of hash functions.
3. The management method of claim 1 , wherein step /4/ is implemented by the following steps:
obtaining second hash values by respectively applying the second set of hash functions to the keywords;
selecting each filter from the subset of indexing filters with which is associated a filter identifier which corresponds to said second hash values.
4. The management method of claim 1 , wherein a first family and a second family of hash functions are initially determined, and wherein the first set and the second set of hash functions are respectively a subpart of said first family and second family of hash functions.
5. The management method of claim 1 , wherein step /4/ is also implemented on the basis of an error limit.
6. The management method of claim 1 , wherein the first set of hash functions is of the LSH type and the second set of hash functions is of the cryptographic type.
7. The management method of claim 1 , wherein the filters in the set of filters are Bloom filters.
8. An identification method for identifying biometric data in a database managed by the management method of claim 3 , said database comprising a structure based on filters (B1, . . . , Bm); wherein said identification method comprises the following steps at an identification unit which has access to the database;
/i/ receiving a query indicating hashed keywords;
/ii/ in the database, determining filters associated with said hashed keywords;
/iii/ obtaining a list of database addresses in encrypted form which are respectively associated with said determined filters;
/iv/ deciding that the keywords have a correspondence in the database, when a same address in said list of database addresses is associated with at least a determined number of filters.
9. The identification method of claim 8 , wherein the hashed keywords are received from a transmitter, said transmitter having obtained the hashed keyword by the following steps:
capturing a biometric data item;
obtaining hash values by respectively applying a third set of hash functions to said captured biometric data item;
obtaining the keywords by respectively combining said first hash values with respective identifiers of hash functions in the third set; and
obtaining hashed keywords by applying a fourth set of hash functions to said keywords.
10. The identification method of claim 9 , wherein the first and third sets of hash functions come from a first family of hash functions and the second and fourth sets of hash functions come from a second family of hash functions.
11. A database management unit adapted to implement the management method of claim 1 .
12. An identification unit adapted to implement the identification method of claim 8 .
13. An identification system comprising the database management unit of claim 11 , a database, and an identification unit.