Method and system for receiving a data stream and optimizing the detection of a specific position within the data stream
A method features receiving a data stream (DS) having video and audio data; detecting a specific position within the DS classified by a predefined criterion (PC) and received by user's clients matching the PC; connecting the receiver not matching the PC with a client-server system matching the PC; receiving the DS by the client-server system and forwarding it to the receiver; extracting the video data and the audio data from the DS stream; creating a screenshot out of the video data and an associated timestamp; feeding the screenshot to an AI-system trained to detect whether it belongs to a predefined category; generating audio data corresponding with the screenshot and analyzing it for detecting whether it belongs to the predefined category; and if both the screenshot and the audio data generated/analyzed belong to the predefined category, then deciding that the associated timestamp for the screenshot defines the specific position.
1 . A method for receiving a data stream by a receiver and detecting at least one specific position within the data stream, wherein
the data stream comprises at least one video data stream and at least one audio data stream;
the data stream is classified according to at least one predefined criterion; and
the data stream is intended to be received by user's clients that match the at least one predefined criterion, characterized in that the method comprises the following steps:
connecting the receiver that does not match the at least one predefined criterion with a client-server system that matches the at least one predefined criterion;
receiving the data stream by the client-server system and forwarding the data stream to the receiver;
extracting the at least one video data stream and the at least one audio data stream from the data stream;
creating at least one screenshot out of the at least one video data stream and an associated timestamp;
feeding the at least one screenshot to an Artificial Intelligence (AI) system, including where the AI-system is based on a neural network ( 16 ), that is trained to detect whether the at least one screenshot belongs to a predefined category;
generating audio data corresponding with the at least one screenshot and automatically analyzing the audio data for detecting whether it belongs to the predefined category;
if both the at least one screenshot and the audio data generated and analyzed belong to the predefined category, then deciding that the associated timestamp for the at least one screenshot defines a specific position.
2 . The method of claim 1 , characterized in that the data stream is a television signal and transmitted via cable, satellite, internet, or a combination thereof.
3 . The method of claim 1 , characterized in that the user's clients comprise a Hybrid Broadcast Broadband television device.
4 . The method of claim 1 , characterized in that the user's clients are classified according to a federal state, a province, a region, or a combination thereof.
5 . The method of claim 1 , wherein an Internet Protocol (IP) address is assigned to each of the user's clients of the client-server system, and IP addresses are used for classification, characterized in that the receiver is connected to the client-server system via a Virtual Private Network.
6 . The method of claim 1 , characterized in that the receiver is connected to more than one client-server system and receives more than one data stream, the more than one data streams being classified differently.
7 . The method of claim 1 , characterized in that the predefined category defines a type of content, including advertisements, a specific content, including a specific advertisement or both the type of content and the specific content, of the data stream.
8 . The method of claim 1 , wherein the data stream comprises first video data that belong to at least one predefined category and second video data that do not belong to the at least one predefined category, characterized in that the specific position is defined to be a start position whenever the associated timestamp that was analyzed previously to a subsequent timestamp corresponding to a subsequent specific position subsequently detected to not belong to the at least one predefined category.
9 . The method of claim 1 , wherein the data stream is transmitted to a first client for being viewed by a user, characterized in that additional information is transmitted to the first client or a second client of the user if the associated timestamp defines the specific position.
10 . The method of claim 1 , characterized in that a likelihood for the at least one screenshot to belong to the predefined category is determined, or a corresponding likelihood for the audio data or an audio fingerprint ( 9 ) to belong to the predefined category is determined and the associated timestamp for the at least one screenshot is only decided to belong to the predefined category, or a combination thereof, if
the at least one screenshot belongs to the predefined category and the corresponding likelihood determined for the audio fingerprint is above a predefined first threshold; or
the audio data and/or audio fingerprint belongs to the predefined category and the likelihood determined for at least one screenshot is above a predefined second threshold; or
the corresponding likelihood determined for the audio fingerprint is above a first threshold and the likelihood determined for the at least one screenshot is above a second threshold.
11 . The method of claim 10 , characterized in that the likelihood for the at least one screenshot to belong to the predefined category is determined by the AI-system.
12 . The method of claim 1 , characterized in that
the at least one audio data stream is converted into a converted-audio data stream having a low resolution signal, a mono audio signal, or a low resolution signal and mono audio signal; and
the converted audio data stream is used for determining whether the audio data at a current position belong to the predefined category.
13 . The method of claim 1 , characterized in that
an audio fingerprint is created from the at least one audio data stream or the audio data converted; and
the audio fingerprint is analyzed for determining whether the audio data belongs to the predefined category.
14 . The method of claim 1 , characterized in that an audio data signal, an audio fingerprint, or the audio data signal and the audio fingerprint, corresponds with at least one screenshot if the associated timestamp for the audio data signal, or the audio fingerprint, or both, matches the associated timestamp of the at least one screenshot.
15 . Server system for receiving a data stream by a receiver and for detecting at least one specific position within a data stream, wherein the data stream comprises at least one video data stream and at least one audio data stream, and the data stream is classified according to at least one predefined criterion and intended to be received by user's clients that match the at least one predefined criterion, characterized in that the server system comprises:
means for connecting the receiver that does not match the at least one predefined criterion with a client-server system that matches the at least one predefined criterion;
means for receiving the data stream by the client-server system and forwarding the data stream to the receiver;
means for extracting the at least one video data stream and the at least one audio data stream from the data stream;
means for creating at least one screenshot from the at least one video data stream and an associated timestamp;
an Artificial Intelligence (AI) system, including where the AI-system is based on a neural network, that is trained to detect whether the at least one screenshot belongs to a predefined category;
means for automatically analyzing audio data corresponding with the at least one screenshot for detecting whether it belongs to the predefined category;
means for deciding that a current timestamp that is associated with a current screenshot defines a specific position, if both the at least one screenshot and the audio data analyzed belong to the predefined category.
16 . The method of claim 2 , characterized in that the user's clients comprise a Hybrid Broadcast Broadband television device.
17 . The method of claim 2 , characterized in that the user's clients are classified according to a federal state, a province, a region, or a combination thereof.
18 . The method of claim 2 , wherein an Internet Protocol (IP) address is assigned to each of the user's clients of the client-server system, and IP addresses are used for classification, characterized in that the receiver is connected to the client-server system via a Virtual Private Network.
19 . The method of claim 2 , characterized in that the receiver is connected to more than one client-server system and receives more than one data stream, the more than one data streams being classified differently.
20 . The method of claim 2 , characterized in that the predefined category defines a type of content, including advertisements, or a specific content, including a specific advertisement, or both the type of content and the specific content, of the data stream.