Method for dynamically generating a text transcription of a continuously broadcast audio stream.

Abstract

Method for dynamically generating a text transcription of a continuously streamed audio stream. The invention relates to a method for dynamically generating a text transcription of a continuously streamed audio stream (AD). This method, implemented at the level of a device receiving the audio stream, comprises: - the processing (11) of the audio stream (AD) by a speech recognition system, delivering a first text transcription of said audio stream (TTB); - the provision (12) of said first text transcription (TTB) to a text transcription validation module comprising at least one human-machine interface (HMI) for validating said first text transcription; - obtaining (13), from the validation module, a second text transcription of the audio stream, called the validated text transcription (TTV). Abstract figure: Figure 1

Description

Title of the invention: Method for dynamically generating a text transcription of a continuously broadcast audio stream. Technical field

[0001] The invention lies in the field of text transcription of audio content. More particularly, the invention relates to the dynamic text transcription of audio content broadcast live. Prior art

[0002] With the ever-increasing number of audio and video contents accessible live or delayed (for example via the Internet), and the multiplication of equipment allowing them to be consumed (televisions, computers, smartphones, tablets, etc.), including in mobile situations, the need for text transcription of speech has multiplied.

[0003] The provision of such transcriptions, for example in the form of subtitles integrated within video content, or more simply a textual transcription on a screen of audio content, is indeed of undeniable interest in many situations. For example, it allows a user to read the information associated with audio or video content on an electronic terminal, even with the sound muted, which can be useful to avoid disturbing other users of a public place or public transport. It also makes it possible to make certain audio content understandable to a wider audience, for example by offering subtitles in a language different from that in which the content is broadcast.In certain situations and / or for certain content, the presence of such a transcription may even be mandatory, for example in the context of compliance with laws for equal rights and opportunities for all, so that deaf or hard of hearing people can have access to information in the same way and under the same conditions as hearing people.

[0004] The generation of text transcriptions of pre-recorded content, intended to be broadcast later, does not generally pose any particular problem: since time is not a critical factor in this case, the production of a good quality transcription, and its synchronization, if necessary, with associated video content, rely on now well-developed techniques, which can be implemented at controlled costs. However, the situation is different when the generation of such text transcriptions relates to content broadcast live (often also referred to as “live” content), since the transcription must be generated quickly, on the fly, as the audio stream is broadcast.

[0005] To achieve the implementation of such transcriptions in real time (or at least in near real time), there are currently two main solutions.

[0006] A first solution consists of using a voice recognition system (also often referred to as ASR, from the English “Automatic Speech Recognition”) to automatically obtain and restore a text transcription of the speech. Such a solution, which is based for example on the use of artificial neural networks, is still unreliable with regard to the quality of the transcription obtained, with a word error rate which is likely to remain relatively high in certain situations (for example in the case of particular diction or accent of the speaker, remarks associated with a complex or very specific lexical field, etc.).This lack of reliability is problematic, particularly when it comes to transcribing in real time speeches broadcast live on important and potentially sensitive subjects, such as, for example, certain interventions by statesmen concerning health, environmental, economic, political or military risks, etc.

[0007] A second solution consists of using the services of people specialized in live and written speech reproduction, known as typists. Such a solution makes it possible to obtain a textual transcription of better quality and more reliable than the first solution, relying in particular on human capacities to contextualize the speech, and possibly to reformulate or adapt it for greater clarity in writing. This solution is however very expensive, on the one hand because typist specialists are rare, and on the other hand because each transcription mission requires them to carry out significant preparatory work in advance, consisting in particular of appropriating the context, identifying the vocabulary likely to be used by the speaker, and configuring in advance the tools used accordingly, so as to be ready to deliver a quality service during the live broadcast.

[0008] Both of these existing solutions therefore have their own specific drawbacks. Furthermore, when they are used to transcribe the audio content of a live video stream, the ongoing display (for example in the form of subtitles) of the generated text transcription suffers from a slight time lag compared to the audio and video content, attributable to the processing time required by the machine (in the case of the first solution) or the typist (in the case of the second solution) to generate this transcription.

[0009] There is therefore a need for a solution that provides a more satisfactory compromise between transcription quality and implementation cost for generating text transcriptions associated with live broadcast content, i.e., content that is streamed as it is produced. More particularly, there is a need for a live text transcription solution that delivers results of sufficient quality to meet the majority of uses, including in particularly demanding contexts, while remaining at controlled costs. Summary of the invention

[0010] The present technique makes it possible to propose a solution aimed at remedying certain drawbacks of the prior art. According to one aspect, the present technique relates to a method for dynamically generating a text transcription of a streaming audio stream. Such a method comprises, at a device receiving said audio stream:

[0011] - processing said audio stream by a voice recognition system, delivering a first text transcription of said audio stream;

[0012] - providing said first text transcription to a validation module of text transcription comprising at least one human-machine interface for validating said first text transcription;

[0013] - obtaining, from said validation module, a second trans textual description of said audio stream, called validated textual transcription.

[0014] In this way, the present technique offers an interesting alternative to the solutions of the prior art, by proposing a hybrid solution making it possible to obtain a text transcription of better quality than that which would have been obtained via a voice recognition system used alone, and at lower costs than that which would have been obtained via the use of pure velotypy services.

[0015] In a particular embodiment, the method for dynamically generating a text transcription further comprises a translation of said validated text transcription, delivering at least a third text transcription of said audio stream in a language other than an original language associated with said audio stream, called translated text transcription.

[0016] In this way, the text transcription obtained with the present technique can be available in several languages, thus promoting understanding by a wider audience of the streaming content.

[0017] In a particular embodiment, said audio stream corresponds to an audio track of a streaming source video stream.

[0018] In this way, the present technique is particularly suitable for the on-the-fly generation of subtitles associated with video content.

[0019] According to a particular characteristic of this embodiment, the method for dynamically generating a text transcription further comprises:

[0020] - buffering said source video stream for a time delay of a predetermined duration, delivering a timed video stream;

[0021] - the on-the-fly integration of said validated textual transcription and / or of said au at least one translated text transcript within said timed video stream, delivering an enriched timed video stream.

[0022] In this way, by delaying the broadcasting of the source video stream by a time, for example, at least equal to the processing time necessary for generating a text transcription of at least part of the broadcast content, the present technique makes it possible to reduce or eliminate the time lag, generally observed with the solutions of the prior art, between the moment when an audio segment is broadcast and the moment when the associated text transcription is displayed.

[0023] In a particular embodiment, said predetermined duration is less than one minute.

[0024] In this way, the broadcasting of the video stream remains carried out in live conditions or at least in “quasi-live” conditions, the delay of the video stream being of short duration.

[0025] In a particular embodiment, said integration of the text transcription within the timed video stream comprises the on-the-fly addition of metadata to this stream, said metadata comprising said validated text transcription and / or said at least one translated text transcription.

[0026] In this way, the present technique offers a lot of flexibility to the end consumer of the content, in terms of rendering the text transcription. More particularly, the integration of the text transcription in the form of metadata offers the advantage, compared to other solutions such as for example an overlay of text directly in the video stream, of allowing the consumer of the content to choose whether or not he wishes to display said transcription, or even to choose the display language of said transcription if applicable.

[0027] According to a particular characteristic of this embodiment, said addition of metadata to said timed video stream takes into account timestamp data associated with the processing implemented by said voice recognition system.

[0028] In this way, the display of the text transcription, for example in the form of subtitles, is perfectly synchronized with the audio track of the broadcast video stream.

[0029] According to another aspect, the present technique also relates to a device for dynamically generating a text transcription of a streaming audio stream. Such a device comprises:

[0030] - means for processing said audio stream, comprising a recognition system voice birth, delivering a first text transcription of said audio stream;

[0031] - means for providing said first text transcription to a module of validation of text transcription comprising at least one human-machine interface for validating said first text transcription;

[0032] - means for obtaining, from said validation module, a second text transcription of said audio stream, called validated text transcription.

[0033] According to another aspect, the proposed technique also relates to a computer program product downloadable from a communications network and / or stored on a computer-readable medium and / or executable by a microprocessor, comprising program code instructions for executing a method for dynamically generating a text transcription of a streaming audio stream as described above, when executed on a computer.

[0034] The proposed technique also relates to a computer-readable recording medium on which is recorded a computer program comprising program code instructions for executing the steps of the method as described above, in any of its embodiments.

[0035] Such a recording medium may be any entity or device capable of storing the program. For example, the medium may comprise a storage means, such as a ROM, for example a CD ROM or a microelectronic circuit ROM, or a magnetic recording means, for example a USB key or a hard disk.

[0036] On the other hand, such a recording medium may be a transmissible medium such as an electrical or optical signal, which may be conveyed via an electrical or optical cable, by radio or by other means, so that the computer program it contains is remotely executable. The program according to the invention may in particular be downloaded over a network, for example the Internet.

[0037] The different embodiments mentioned above can be combined with each other for implementing the invention. Figures

[0038] Other characteristics and advantages of the invention will appear more clearly on reading the following description of a preferred embodiment, given as a simple illustrative and non-limiting example, and the appended drawings, among which:

[0039] [Fig-1] illustrates the general principle of a method for dynamically generating a text transcription of a streaming audio stream, in a particular embodiment of the proposed technique;

[0040] [Fig.2] schematically presents the different functional blocks of a device for dynamically generating a text transcription of a streaming video stream, in a particular embodiment of the proposed technique;

[0041] [Fig.3] schematically describes a synchronization mechanism of a trans text description with an audio or video stream to which it is associated, in a particular embodiment of the proposed technique;

[0042] [Fig.4] presents a simplified architecture of a device for dynamically generating a text transcription for implementing the proposed technique, in a particular embodiment. Detailed description of the invention

[0043] The present application makes it possible to overcome certain of the aforementioned drawbacks of the prior art solutions for generating text transcriptions relating to live broadcast content.

[0044] The proposed technique aims in particular to propose a solution offering a particularly interesting compromise between the quality of the textual transcription obtained and the cost of implementation, which the solutions of the prior art do not offer.

[0045] The present technique relating to live broadcast streams, by processing "by segment" (e.g. transmission by segment, delivery by segment, etc.) is meant in the remainder of this document a processing carried out on the fly, as and when a stream of audio or text data transmitted continuously is received, typically a processing by word or by group of words (for example by "breath group", i.e. by group of words pronounced in a single breath, without pause) carried out on the fly.

[0046] Furthermore, the term “module” can correspond to a software component as well as to a hardware component or a set of hardware and software components, a software component itself corresponding to one or more computer programs or sub-programs or more generally to any element of a program capable of implementing a function or a set of functions.

[0047] In all the figures of this document, identical elements and steps are designated by the same numerical reference.

[0048] According to a first aspect, the present technique relates to a method for dynamically generating a text transcription of a streaming audio stream. The general principle of such a method is illustrated in relation to [Fig.l], in a particular embodiment of the proposed technique. This method is implemented within a device comprising means for receiving a source data stream comprising said audio stream FA. Such a source data stream may consist solely of the audio stream FA in question, or take, for example, the form of a video stream comprising at least one audio track forming the audio stream FA.This device is typically an intermediate device in a live broadcast chain of the source data stream, making it possible in particular, for example, as described below, to enrich the source data stream received from a live broadcast server before relaying it to a plurality of readers intended to restore it to end users consuming the associated content.

[0049] In a step 11, the audio stream FA received directly or extracted from a video stream by the dynamic generation device is processed by a voice recognition system, such as for example of the “Speech to Text” type, NLP (“Natural Language Processing” in English), etc. Such a system is integrated into the generation device, or alternatively connected to this device via a communication network. More particularly, the audio stream FA is transmitted to the voice recognition system as it is received, this system delivering in return, by segment, a text data stream representative of a first text transcription of the audio stream FA, called raw text transcription TTB in the sense that it has not been subject to any verification at this stage.

[0050] In a step 12, the raw text transcription TTB is transmitted by segment, i.e. as it is generated in step 11, to a text transcription validation module of the generation device. Such a validation module provides a USR user, for example via a front-end software application, at least one human-machine interface (HMI) making it possible to consult in real time and continuously, on a display screen, the raw text transcription TTB delivered by the voice recognition system. Via this human-machine interface HMI, the USR user thus has the possibility of validating the text transcription, possibly after having modified it in substance (for example for the purposes of correction, reformulation, moderation, etc.) and / or on the form (for example to add line breaks, dashes, modify the font or color of the text to better differentiate speakers, etc.) using tools available for this purpose within the validation module.

[0051] In a step 13, the generation device thus obtains by segment, as and when they are validated in the validation module, a new text transcription of the audio stream FA, called validated text transcription TTV in the sense that it has been reviewed and validated by a user.

[0052] This provides a hybrid solution, making it possible to obtain a text transcription of better quality than that which would have been obtained via an automatic speech recognition system used alone, while being less expensive than a pure velotyping solution, insofar as the user in charge of validating the transcription does not need to have all the skills of a velotypist specialist because he already has access, thanks to the proposed technique, to a raw text transcription base which he simply needs to adapt to produce a reliable transcription of the audio stream broadcast live, all in almost real time.

[0053] According to a particular characteristic, the proposed technique also offers the possibility, for example by means of a dedicated configuration option, of deactivating the validation by a human of the text transcription generated in step 11 at the output of the voice recognition system. Such a configuration option, accessible for example within an administration interface of the dynamic generation device according to the present technique, makes it possible in particular to ensure the provision of a text transcription of the audio stream broadcast continuously, including in the case where a person normally responsible for validating the automatically generated text transcription proves to be unavailable or has to be temporarily absent. When activated, such disabling of the validation by a human of the text transcription can take different forms. In a first implementation, it can for example consist of a pure and simple deactivation of steps 12 and 13 previously described, which are then no longer implemented, at least temporarily.Alternatively, in a second implementation, steps 12 and 13 remain implemented, but the text transcription generated by the voice recognition system and then provided to the text transcription validation module is automatically validated by this module after a predetermined time (for example of the order of ten seconds, possibly configurable within the administration interface mentioned above), when no interaction of a user with the human-machine interface made available by this validation module is detected during this period of time.

[0054] Optionally, in a step 14, the validated text transcription TTV is possibly translated, by an automatic text translation system, into at least one language other than the original language associated with the audio stream FA. This automatic translation system is integrated into the generation device, or alternatively connected to this device via a communication network. Such a translation is carried out on the fly, the automatic translation system delivering per segment at least one translated text transcription TTT of the validated text transcription TTV.

[0055] The validated text transmission TTV and / or the translated text transcription(s) TTT generated on the fly can then be used for various uses, and in particular to enrich the source audio or video data stream as described below in a particular embodiment of the present technique.

[0056] We will now describe, in relation to [Fig. 2], an example of functional architecture of a device 20 for dynamically generating a text transcription of a continuously broadcast audio stream, in a particular embodiment of the proposed technique. More particularly, the device illustrated in [Fig. 2] is in particular suitable for generating subtitles associated with a live broadcast source data stream, taking the form of a video stream comprising at least one audio track. As described below, various complementary aspects of the method for dynamically generating a text transcription are highlighted through this example, in different particular embodiments of the proposed technique.

[0057] In this example of architecture, given for illustrative and non-limiting purposes, the device 20 is divided into two main functional blocks - a block 21 for managing the live stream and a block 22 for live generation of at least one text transcription - each comprising different modules described below.

[0058] The live stream management block 21 receives as input, at a stream manipulation module 211, a live FVO video stream (original video stream), for example from a video server SRV_V. The stream manipulation module 211 then, on the one hand, buffers the original FVO video stream, and, on the other hand, extracts, in the form of an audio stream FA, the audio track associated with the FVO video stream. As detailed below, the original FVO video stream is more particularly placed in the buffer memory for a time delay of a predetermined duration, typically of the order of a few seconds to a few tens of seconds, but generally less than one minute. Such a time delay makes it possible to slightly delay the video stream while the text transcription operations of the original FVO video stream are carried out.The delay time being very short, the broadcasting conditions remain similar to those of a live broadcast.

[0059] The extracted audio stream FA is then transmitted to the block 22 for live generation of at least one text transcription, where it is received at an audio management module 221.

[0060] The audio management module 221 relays this audio stream FA to a voice recognition engine 222. The audio management module 221 receives in return, by segment, a stream of textual data corresponding to the recognized text, as the audio stream is processed by the voice recognition engine 222. According to a particular characteristic, each textual segment thus recognized and transmitted to the audio management module 221 by the voice recognition engine 222 includes start and end timestamp information for the corresponding audio segment in the original video stream FVO.

[0061] The audio management module 221 transmits this text data stream (and any associated time stamp information) to a validation module 223, where it is displayed on the fly in a validation human-machine interface in order to be quickly checked, possibly modified and / or moderated and / or reformulated, then validated, by segment, by a USR user, via suitable means made available within the validation module.

[0062] The validated segments of the text data stream are returned to the audio management module 221. Optionally, as already presented in relation to [Fig.l], the audio management module 221 transmits the text data stream validated to a translation engine 224. The audio management module 221 receives in return, per segment, at least one stream of translated text data, corresponding to the validated text translated into a language different from the original language of the original video stream. Each translated segment remains associated with the start and end timestamp information of the corresponding audio segment, in its original language, in the original video stream FVO.

[0063] At this point, at the live generation block of at least one textual transcription 22, the stream of textual data validated in the original language of the original video stream, and possibly one or more streams of textual data translated into other languages, are available.

[0064] These textual data streams are transmitted to the live stream management block 21, where they are received in real time, by segment, at a stream enrichment module 212. As they are received, in a particular embodiment, the stream enrichment module 212 then integrates these textual data streams on the fly into the video stream retrieved from the buffer memory where it was placed by the stream manipulation module 211. As described previously, this video stream retrieved from the buffer memory corresponds more particularly to the original video stream delayed (or timed) by a time delay of a predetermined duration at least equal to the duration necessary to generate the validated and possibly translated textual data stream.According to a particular characteristic of the proposed technique, these textual data streams are integrated into the timed video stream, in the form of metadata corresponding to their respective language, taking into account the timestamp information available for each segment of said data streams. For example, as illustrated in relation to [Fig. 3], each time a frame or a group of video frames (video frames 1, 2, 3, 4 or 5 in [Fig. 3]) is read in the buffer memory, an associated segment (typically a word, for example one of the four words W1, W2, W3 and W4 constituting the sentence "Hello, how are you?", or a group of words) of the textual data stream obtained at the output of the GEN_TT operation for dynamic generation of the textual transcription of the original video stream FVO according to the present technique is integrated into the timed video stream, in the form of subtitle metadata.More specifically, in a SYNC synchronization operation, it is the segment whose timestamp information corresponds to that of the current frame read in the timed video stream that is inserted and associated with this frame in the timed video stream (the time delay of a duration DLY obtained thanks to the buffering of the original video stream FVO making it possible to ensure that this segment of the validated and possibly translated textual data stream is already available at the time of reading the corresponding video frame of the timed video stream). In this way, an enriched video stream FVE is obtained. featuring perfect synchronization between text transcription and video frames.

[0065] The FVE enriched video stream, which is delayed by only a few seconds compared to the original video stream, can then be broadcast in a quasi-live configuration to PLY_V video players which, if they are compatible with the standards relating to subtitling (which is the case for the majority of current players), allow the end consumer to display or not the subtitles, and, if necessary, to choose the subtitling language (provided that the text data stream translated into the chosen language is available). Thanks to the proposed technique, these subtitles also appear without any time lag when playing the FVE enriched video stream.

[0066] In other words, in summary, if we refer to the tablets labeled 1, 2, 3, 4 and A, B, C, D in [Fig.2]:

[0067] - the live stream management block 21 is responsible for: 1. Extracting an audio stream matching the audio track of the live video stream; 2. Buffer the original live video stream to delay the live broadcast; 3. Gather the generated subtitles; 4. Enrich the live video stream “on the fly” by integrating the subtitles, while ensuring synchronization with speech; 5. Deliver the enriched video stream to the broadcast with the previously configured low delay;

[0068] - the live generation block of at least one text transcription 22 is in responsible for: A. Generating speech recognition from the live audio stream, for example using an automatic speech recognition (ASR) system; B. Providing a user interface to easily enable live editing, moderation and validation of the speech recognition, in order to generate validated subtitles; C. Optionally performing a translation of the subtitles into another language; and D. Providing a stream of subtitles, possibly in multiple languages, to be integrated into the slightly time-delayed live video stream.

[0069] Thus, such an implementation according to at least one particular embodiment of the proposed technique allows:

[0070] - not to start from a “blank page” for the generation of the transcription textual reproduction of a live audio stream, benefiting from the results of initial raw and automated voice recognition processing;

[0071] - to allow time for a human being to quickly review the raw text recognized and to offer the possibility of modifying and validating it easily in real time, via a dedicated human-machine interface;

[0072] - to automatically integrate the validated text transcription obtained, for example in the form of subtitles, in the video stream broadcast almost live, with a quality of subtitling and synchronization equivalent to that which would be possible to obtain for pre-recorded video content intended to be broadcast on a delayed basis (i.e. not in direct).

[0073] As already described previously, according to another aspect, the proposed technique also relates to a device for dynamically generating a text transcription of a streaming audio stream, capable of carrying out the method presented in any of its embodiments. More particularly, such a device according to the present technique comprises:

[0074] - means for processing the audio stream, comprising a recognition system vocal, delivering a first text transcription of the audio stream, called raw text transcription;

[0075] - means for providing the raw text transcription to a va module liation comprising at least one human-machine interface for validating the raw text transcription;

[0076] - means for obtaining from the validation module, a second text transcription of the audio stream, known as validated text transcription, established via said human-machine interface from the raw text transcription.

[0077] [Fig.4] represents, in a schematic and simplified manner, the structure of such a device, in a particular embodiment. The intermediate device according to the proposed technique comprises for example a memory 41 consisting of a buffer memory M, a processing unit 42, equipped for example with a microprocessor pP, and controlled by the computer program Pg 43, implementing steps of the method for dynamically generating a text transcription of a continuously broadcast audio stream, according to at least one embodiment of the invention. To this end, the generation device also comprises at least one communication interface (for example an Ethernet communication interface), allowing it to receive and transmit data streams (video, audio, text) from and to other equipment present in the communication network.

[0078] Upon initialization, the code instructions of the computer program 43 are loaded into the buffer memory before being executed by the processor of the processing unit 42. The processing unit 42 continuously receives as input E a source data stream comprising an audio stream, coming from equipment of a live broadcasting channel, for example coming from a live broadcasting server.

[0079] The microprocessor of the processing unit 42 then carries out the steps of the method for dynamically generating a text transcription of the continuously broadcast audio stream, according to the instructions of the computer program 43. More particularly, the audio stream is processed in a first step by a voice recognition system, and the resulting raw text data stream is processed in a second step by a human operator by means of a dedicated human-machine interface delivered by the computer program 43, thus making it possible to obtain in real time or quasi-real time at output S a validated text transcription of the audio stream.

Claims

Claims

1. Method for dynamically generating a text transcription of a continuously broadcast audio stream (FA), implemented at a device receiving said audio stream, said method being characterized in that it comprises the following steps: - processing (11) said audio stream (FA) by a voice recognition system, delivering a first text transcription of said audio stream (TTB); - supplying (12) said first text transcription (TTB) to a text transcription validation module comprising at least one human-machine interface (HMI) for validating said first text transcription;- obtaining (13), from said validation module, a second text transcription of said audio stream, called validated text transcription (TTV), said validated text transcription corresponding to said first text transcription when no interaction with said validation human-machine interface is detected during a predetermined time.;

2. Method according to claim 1, characterized in that it further comprises a step of translation (14) of said validated text transcription (TTV), delivering at least a third text transcription of said audio stream in a language other than an original language associated with said audio stream, called translated text transcription (TTT).

3. Method according to claim 2, characterized in that said audio stream corresponds to an audio track of a streaming source video stream.

4. Method according to claim 3, characterized in that it further comprises the following steps: - buffering said source video stream for a time delay of a predetermined duration, delivering a timed video stream; - on-the-fly integration of said validated text transcription and / or of said at least one translated text transcription within said timed video stream, delivering an enriched timed video stream.

5. Method according to claim 4, characterized in that said predetermined duration is less than one minute.

6. Method according to claim 4, characterized in that said integration comprises the on-the-fly addition, to said timed video stream, of metadata comprising said validated text transcription and / or said at least a translated text transcription.

7. Method according to claim 6, characterized in that said addition of metadata to said timed video stream takes into account timestamp data associated with the processing implemented by said voice recognition system.

8. Device for dynamically generating a text transcription of a continuously broadcast audio stream, said device being characterized in that it comprises: - means for processing said audio stream, comprising a voice recognition system, delivering a first text transcription of said audio stream; - means for supplying said first text transcription to a text transcription validation module comprising at least one human-machine interface for validating said first text transcription; - means for obtaining, from said validation module, a second text transcription of said audio stream, called validated text transcription, said validated text transcription corresponding to said first text transcription when no interaction with said validation human-machine interface is detected during a predetermined time.

9. Computer program product downloadable from a communications network and / or stored on a computer-readable medium and / or executable by a microprocessor, characterized in that it comprises program code instructions for executing a method according to any one of claims 1 to 7, when executed by a computer.

10. A computer-readable recording medium comprising program code instructions which, when executed by a computer, cause the computer to carry out the method of any one of claims 1 to 7.

Images