Method for designating a user account and associated electronic device

Abstract

An electronic device (2) comprises: - a module for extracting a code (C') by analyzing audio or video content (CT'); - a module for selecting, from a set of user accounts, at least one user account (Uj) based on at least one piece of information associated with the user account in question; and - a module for designating at least one user account (Uk) based on accusation values ​​determined respectively for the selected user accounts (Uj) according to the extracted code and a code associated with the user account in question (Uk). A method for designating at least one user account from a set of user accounts is also described. See Figure 2 for the abstract.

Description

Title of the invention: Method for designating a user account and associated electronic device Technical field of the invention

[0001] The present invention relates to the technical field of digital watermarking of audiovisual content.

[0002] It relates in particular to a method of designating a user account and an associated electronic device. State of the art

[0003] It is known to insert a code by digital tattooing within audio or video content in order to be able to subsequently determine the origin of this content.

[0004] For example, in audio or video content distribution systems, it has been proposed to associate a code with each user account and to insert this code into the content distributed to this user account.

[0005] Thus, when an unauthorized copy of the content is found, it is possible by extracting the code by analyzing this copy to trace back to the user account at the origin of the unauthorized copy.

[0006] To circumvent this protection measure, some hackers work together to detect the location of the code insertion by comparing different distributed versions of the same content. Hackers can then modify the code inserted by digital tattooing so that it is no longer possible to trace the person who actually made the unauthorized copy.

[0007] It has been proposed, for example in the article "Optimal Probabilistic Fingerprint Codes

[0008] When an unauthorized copy is discovered, an accusation function is then used which makes it possible to determine, with a certain probability of error, which codes can be linked to the code extracted from this unauthorized copy.

[0009] The required length of the codes (in number of symbols, i.e. in number of bits in the case of the binary codes proposed in the above-mentioned article) depends on the number of potential users (i.e. the number of user accounts in the example of the distribution of audio or video content), the number of potentially associated pirates and the aforementioned probability of error, and easily tends to exceed the lengths acceptable in practice for the insertion of the code into a content. Presentation of the invention

[0010] In this context, the present invention provides a method for designating at least one user account from a set of user accounts, in a audio or video content distribution system in which codes are respectively associated with the user accounts of the set, the method comprising the following steps:

[0011] - extraction of a code by analysis of audio or video content;

[0012] - selection, within the set of user accounts, of at least one account user based on at least one piece of information associated with the relevant user account;

[0013] - designation of said at least one user account on the basis of values charges determined respectively for the selected user accounts based on the extracted code and the code associated with the user account concerned.

[0014] The user account selection step allows the hackers' search to be focused on a limited number of user accounts. Thus, the length of the code (in number of symbols) necessary to obtain a given error probability will be reduced since this error probability will be targeted within a smaller population.

[0015] The information associated with the user account concerned is, for example, information dependent on the content reproduced using the user account concerned.

[0016] Said information may be an estimate of a level of interest in said audio or video content determined based on data associated with the user account concerned.

[0017] This data associated with the user account concerned may include, for example, profile data associated with the user account concerned and / or consumption data associated with the user account concerned. This consumption data associated with the user account concerned is, for example, indicative of audio or video content reproduced (e.g., viewed in the case of video content) using the user account concerned.

[0018] It may be provided, for example, that at the selection stage, the user accounts for which the estimate of the level of interest is greater than a threshold are selected.

[0019] According to another possibility, it can be provided that at the selection step, the user accounts of a subset of accounts are selected for which the sum of the estimates of the level of interest respectively associated with the user accounts of the subset is maximum. As explained in the description which follows, this subset can have a predetermined cardinal (i.e. a number of elements).

[0020] The selection step can also use a distance defined between two user accounts as the number of common contents reproduced by these two user accounts.

[0021] It may further be provided that at the designation step, the designated user accounts are the user accounts for which the determined accusation value is greater than a predetermined threshold.

[0022] According to an exemplary implementation described later, the codes comprising a predetermined number of symbols, the method may comprise a preliminary step of associating a code with each user account of the set by determining each of the symbols of the code concerned randomly from at least two possible values.

[0023] It can be provided in this context that each symbol is determined randomly in accordance with a predetermined probability distribution between the different possible values ​​and associated with the position of the symbol concerned within the code concerned.

[0024] In this context, for a given user account, the accusation value can be determined based on the extracted code, the code associated with the given user account and the probability distributions respectively associated with the different symbol positions.

[0025] The method proposed above can be implemented within an electronic device, for example such as that described below.

[0026] The invention also provides an electronic device comprising:

[0027] - a module for extracting a code by analyzing audio or video content;

[0028] - a module for selecting, within a set of user accounts, at least a user account based on at least one piece of information associated with the relevant user account;

[0029] - a module for designating at least one user account based on values charges determined respectively for the selected user accounts based on the extracted code and a code associated with the user account concerned.

[0030] Of course, the various features, variants and embodiments of the invention may be combined with each other in various combinations to the extent that they are not incompatible or mutually exclusive. Detailed description of the invention

[0031] Furthermore, various other characteristics of the invention emerge from the appended description given with reference to the drawings which illustrate non-limiting forms of embodiment of the invention and where:

[0032] [Fig.l] schematically represents a system for distributing audio or video content;

[0033] [Fig.2] represents the main elements of an electronic device according to a mode possible implementation of the invention; and

[0034] [Fig.3] is a flowchart representing the main steps of an example of a method according to the invention.

[0035] [Fig.l] represents a system for distributing audio and / or video content (for example multimedia content) comprising a distribution server S and reproduction equipment Eb E, EN.

[0036] Each reproduction equipment EB E, EN can be used by a user subscribed to the distribution system by means of a user account Ub U;, UN.

[0037] The association between a reproduction device and a user account is not fixed in time, a user being able for example to have several reproduction devices to access audio or video contents using his user account. However, for the sake of simplification, it is considered in the present description that a user subscribed to the distribution system by means of a user account U; accesses the audio or video contents by means of a reproduction device Ei.

[0038] Each reproduction equipment Eb E, EN may comprise a sound reproduction device capable of reproducing audio content received from the distribution server S.

[0039] Each reproduction equipment Eb E, EN may further comprise a display device capable of displaying video content received from the distribution server S.

[0040] Each reproduction equipment EB E, EN may further comprise in practice a communication circuit designed to receive the aforementioned audio or video content (or a coded representation of this content) from the distribution server S.

[0041] Each reproduction equipment EB E, EN may optionally comprise a decoder designed to decode the aforementioned coded representation and thus produce the audio content to be reproduced (by means of the sound reproduction device) or the video content to be displayed (by the display device).

[0042] Each reproduction equipment EB E, EN can be for example a multifunction terminal (such as a multifunction mobile telephone), a touch tablet, a television or a decoder-display screen assembly.

[0043] As will be explained in more detail below, the distribution server S is designed to insert, by digital tattooing and within an original audio or video content CO, a code Ci, Ci, CN associated with a user account Ui, Ui, UN, and to transmit the content marked by this code Ci, Ci, CN (or watermarked content CE, CE, CTN) to the equipment Eb E, EN used by this user account Ui, Ui, UN.

[0044] According to a possible variant, the code Ci, Ci, CN associated with a user account Ui, Ui, UN could be inserted (by digital tattoo) within the audio content or original video by a decoder of the reproduction equipment Eb E;, EN used by this user account Ub Ui5 UN, before reproduction, by this reproduction equipment Ei, Ej, En, of the audio or video content marked by this code Cb C;, CN (ie tattooed).

[0045] [Fig.2] represents the main elements of an electronic device 2 according to a possible embodiment of the invention.

[0046] The electronic device 2 comprises an extraction module 4, a selection module 6 and a designation module 8.

[0047] The electronic device 2 comprises for example a processor and a memory.

[0048] At least some of the aforementioned modules can in this case be implemented due to the execution, by the processor, of computer program instructions stored in the memory.

[0049] Alternatively or additionally, at least one of the aforementioned modules may be implemented using a dedicated electronic circuit, such as an application-specific integrated circuit.

[0050] The extraction module 4 is designed to extract a code C' by analyzing an audio or video content CT'.

[0051] As described in more detail below, this audio or video content CT' is distributed watermarked content (i.e. marked by a code Ci, Ci, CN associated with a user account Ui, Ui, UN, in accordance for example with the scheme described above with reference to [Fig.l]), possibly altered by an attempt to remove the marking, then copied and detected as an unauthorized copy. Due to the aforementioned potential alteration, the extracted code C' may not correspond to any of the codes Ci, Ci, CN assigned to the different user accounts Ub Ui5 UN.

[0052] The selection module 6 is designed to select certain user accounts Uj from among the set of user accounts Ub Ui5 UN. For each user account considered, the selection (or not) of this user account is here carried out according to at least one piece of information associated with this user account, for example according to at least one piece of information depending on the contents reproduced using the user account concerned.

[0053] As explained below, the selection module 6 can thus use data relating to the respective consumption histories of the different user accounts Ub Ui5 UN.

[0054] Consumption histories are, for example, reproduction histories of audio or video content (i.e., viewing or display histories in the case of video content). Such histories are commonly used, for example, by recommendation engines, and are therefore generally available within content distribution systems.

[0055] Data relating to the consumption histories of the different accounts user Ub Ui5 UN are for example stored within a storage unit 10.

[0056] The storage unit 10 is shown external to the electronic device 2 in [Fig. 2], as in the conceivable case where this storage unit 10 is included in a remote server. Alternatively, however, the storage unit 10 could be internal to the electronic device 2.

[0057] The designation module 8 is designed to calculate, for each of the user accounts Uj selected by the selection module 6, an accusation value as a function of the code extracted C' by the extraction module 4 and of the code associated with the user account Uj concerned (using for example an accusation function such as proposed in the article "Symmetric Tardos fingerprinting codes for arbitrary aphabet sizes", by B. Skoric, S. Katzenbeisser and MU Celik, Cryptology ePrint Archive, 2007 / 041, 2007, as also explained below in the context of the description of the method of [Fig.3]).

[0058] The designation module 8 is then designed to designate (among the user accounts Uj selected by the selection module 6) at least one user account Uk on the basis of the accusation values ​​respectively calculated for the selected user accounts Uj.

[0059] By obtaining the charge values ​​in accordance herein with the aforementioned article "Symmetric Tardos fingerprinting codes for arbitrary alphabet sizes", a Uk user account so designated corresponds to a user account used to make the unauthorized copy mentioned above.

[0060] [Fig. 3] is a flowchart representing the main steps of an example of a method according to the invention.

[0061] As will emerge from the following description, such a method can be implemented in the distribution system of [Fig.l] (steps E2 to E6), then in the electronic device of [Fig.2] (steps E8 to E14).

[0062] The method begins with a step E2 of associating a code C; with each user account U; of the set of user accounts Ub U;, UN. This association step is for example carried out by the distribution server S.

[0063] Each code C; is for example formed from a predetermined number M of symbols. Each code C; is for example constructed in step E2 by determining each of the symbols of the code C; concerned randomly from at least two possible values.

[0064] The codes Ci can for example be constructed (then associated respectively with the user accounts Ui) in accordance with the solution proposed in the article "Symmetric Tardos fingerprinting codes for arbitrary alphabet sizes" cited above.

[0065] For each code Ci, each symbol can thus be determined randomly in accordance with a predetermined probability distribution between the different possible values ​​and associated with the position of the symbol concerned within the C code concerned.

[0066] The method then comprises a step E4 of inserting (by digital tattooing) the code Ci (associated with the user account UO within the original content CO so as to produce marked content (or tattooed content) CT; (intended to be reproduced by means of the aforementioned user account U;).

[0067] The marking process (i.e. digital watermarking) used is for example of the type described in the article "Secure spread spectrum watermarking for multimedia", by LJ. Cox, J. Kilian, FT Leighton and T. Shamoon, in IEEE Transactions on Image Processing (vol. 6, no. 12, December 1997).

[0068] The method also comprises a step E6 of distributing (or transmitting) the content marked CT; to the reproduction equipment E; associated (possibly temporarily as already indicated) with the user account U;

[0069] In the example described here, the insertion steps E4 and distribution steps E6 are carried out by the distribution server S.

[0070] In an alternative embodiment in which it is a decoder (equipping the reproduction equipment Ei) associated with the user account U; which performs the insertion of the code Ci, the steps E4 and E6 would be replaced by a step of distribution of the original content CO by the distribution server S to the reproduction equipment E;, and a step of insertion (by digital tattoo) of the code Ci (associated with the user account UO within the original content CO by the decoder.

[0071] In all cases, the content marked Cl) can then be reproduced by means of the reproduction equipment E;, for example viewed when it is video content. As already indicated, the reproduction of the content (here of the content marked Cl)) can be a sound reproduction of this content (case of audio content) and / or a display (or viewing) of this content (case of video content).

[0072] For each user account U;, a list of the different contents distributed to a user account U; and / or reproduced (eg viewed) by means of this user account U; is stored, here in particular within the storage unit 10. Such a list is an example of consumption history.

[0073] The distributed contents CTb Cl), CTN are sometimes copied in an unauthorized manner and an unauthorized copy CT' of a distributed content CT; can thus be detected (step E8).

[0074] The unauthorized copy CT' may differ slightly from the distributed content CT; in particular because the persons making this unauthorized copy CT' generally seek- grumbling to remove the marking made.

[0075] The electronic device 2 can then be used to designate one or more user account(s) as likely responsible for the unauthorized copy CT' on the basis of this unauthorized copy CT'.

[0076] To do this, the extraction module 4 proceeds in step E10 to extract a code C' by analyzing the marked content CT' (here the unauthorized copy detected in step E8). The extraction process used is the reciprocal of the marking process (i.e. digital watermarking) used in step E4 (see the article "Secure spread spectrum watermarking for multimedia" cited above).

[0077] The selection module 6 then proceeds to a step E12 of selecting at least one user account Uj from among the set of user accounts Ub U;, UN based in particular on information associated with the user account concerned. Several user accounts Uj can thus be selected at this step.

[0078] According to a first possible embodiment, the selection module 6 uses an appetence function app(x,y) which expresses the appetence of a user account x for a content y taking into account the consumption history (or reproduction history, i.e. for example the viewing history) of the user account x. As an appetence function, a similarity measure between the content y and the contents of the consumption history of the user account x can be used. Such similarity measures are for example presented in the article "Step By Step Content-Based Recommendation System" published on the medium.com website.

[0079] The aforementioned information, on the basis of which the user account is possibly selected, may in this case be an estimation of a level of interest in the content CT' determined (thanks to the aforementioned app appetite function) as a function of data associated with the user account concerned (for example as a function of profile data associated with the user account concerned or, as already indicated, consumption data associated with the user account concerned, such as the consumption history data of the user account concerned).

[0080] The selection module 6 selects for example in step E12 the user accounts Uj for which the estimation of the level of interest app(Uj, CT') for the content CT' is greater than a threshold amin, that is to say the user accounts Uj for which:

[0081] app(Uj, CT') > amin.

[0082] Alternatively, the number ns of user accounts selected in step E12 could be predetermined; the selection module selects for example in this case (all) the user accounts of a subset of accounts having as cardinal the predetermined number ns and for which the sum of the estimates of the level of interest respectively associated with the user accounts of the subset is maximum.

[0083] In this variant, the number ns can be chosen in advance depending, for example, on the reliability sought at the designation stage, the available calculation time, the length of the codes used (this length being linked to the length of the marked content).

[0084] According to a second possible embodiment, the selection module 6 uses (in the selection step E12) a distance defined between two user accounts as the number of common contents reproduced (i.e. viewed or listened to) by these two user accounts.

[0085] The selection module 6 can then, for example, implement a partitioning technique (or "clustering" according to the commonly used English term) so as to select, from among the user accounts having reproduced the audio or video content corresponding to the content marked CT', the user accounts closest in terms of the aforementioned distance (number of contents reproduced in common).

[0086] As in the first embodiment possibility, in certain embodiments it is possible to fix in advance the number ns of user accounts selected in the selection step E12. In this case, the selected user accounts are the ns user accounts having reproduced the content corresponding to the content CT' and for which the sum of the ns(ns-l) / 2 distances two by two is minimal.

[0087] For each user account Uj selected in the selection step E12 (and only for these selected user accounts), the designation module 8 then determines an accusation value (associated with this user account Uj) as a function of the code C' extracted in the step E10 and of the code associated with the user account concerned (step El4).

[0088] In the example described here where probability distributions are used when constructing the codes C; (see above the description of step E2), the accusation value associated with a user account Uj is determined as a function of the extracted code C', the code Q associated with the user account Uj concerned and the probability distributions respectively associated with the different symbol positions, for example in accordance with what is described in the article "Symmetric Tardas finger-printing codes for arbitrary aphabet sizes" cited above.

[0089] The designation module 8 then proceeds to a step E16 of designation (among the selected user accounts Uj) of one or more user account(s) Uk on the basis of the accusation values ​​determined respectively in step E14 for the selected user accounts Uj.

[0090] The designated user account(s) are for example the user account(s) for which the accusation value determined in step E14 is greater than a predetermined threshold.

[0091] By selecting in advance (in step E12) only part of the accounts user, the designation process of steps E14 and E16 is carried out on a reduced number of user accounts and can thus, for example, be carried out with shorter codes (in number of symbols), the insertion of which into distributed content of usual size will not pose a problem.

Claims

Claims

1. Method for designating at least one user account (Uk) from a set of user accounts (Ub U;, UN), in an audio or video content distribution system in which codes (Ci, Ci, CN) are respectively associated with the user accounts (Ui, Ui, UN) of the set, the method comprising the following steps: - extraction (E10) of a code (C') by analysis of an audio or video content (CT'); - selection (E12), within the set of user accounts (Ui, Ui, UN), of at least one user account (Uj) as a function of at least one piece of information associated with the user account (Uj) concerned; - designation (E14, E16) of said at least one user account (Uk) on the basis of accusation values ​​determined respectively for the selected user accounts (Uj) according to the extracted code (C') and the code associated with the user account (Uk) concerned.

2. A method according to claim 1, wherein said information is an estimate of a level of interest in said audio or video content determined based on data associated with the relevant user account (Uj).

3. The method of claim 2, wherein the data associated with the relevant user account (Uj) comprises profile data associated with the relevant user account (Uj).

4. The method of claim 2 or 3, wherein the data associated with the relevant user account (Uj) comprises consumption data associated with the relevant user account (Uj).

5. Method according to one of claims 2 to 4, in which, in the selection step (El2), the user accounts (Uj) are selected for which the estimate of the level of interest is greater than a threshold.

6. Method according to one of claims 2 to 4, in which, in the selection step (E12), the user accounts of a subset of accounts are selected for which the sum of the estimates of the level of interest respectively associated with the user accounts of the subset is maximum.

7. Method according to one of claims 1 to 6, in which the selection step (El2) uses a distance defined between two user accounts as the number of common contents reproduced by these two user accounts.

8. Method according to one of claims 1 to 7, in which, in the designation step (E16), the designated user accounts are the user accounts for which the determined accusation value is greater than a predetermined threshold.

9. Method according to one of claims 1 to 8, in which the codes (Cb Ci, CN) comprise a predetermined number of symbols, and in which the method comprises a preliminary step (E2) of associating a code (Ci; C; ; CN) with each user account (Ui; Ui; UN) of the set by determining each of the symbols of the code concerned randomly from at least two possible values.

10. A method according to claim 9, wherein each symbol is determined randomly in accordance with a predetermined probability distribution between the different possible values ​​and associated with the position of the symbol concerned within the code concerned.

11. A method according to claim 10, wherein, for a given user account (Uj), the accusation value is determined as a function of the extracted code (C'), the code associated with the given user account and the probability distributions respectively associated with the different symbol positions.

12. Electronic device (2) comprising: - a module (4) for extracting a code (C') by analyzing audio or video content (CT'); - a module (6) for selecting, within a set of user accounts (Ub U;, UN), at least one user account (Uj) based on at least one piece of information associated with the user account (U j) concerned; - a module (8) for designating at least one user account (Uk) based on accusation values ​​determined respectively for the selected user accounts (Uj) based on the extracted code (C') and a code associated with the user account (Uk) concerned.

Images