Abstract: Accurate real time automatic detection of events in content of a data stream, such as a transition to a commercial block in the content of a broadcast audio/video data stream, relies on a trainable event classifier that operates on a well-balanced training set input to the classifier. The present disclosure provides a computer controlled method of operating a training tool for classifying events annotated in the content of a data stream. The training tool presents training samples comprising separators and corresponding descriptors that relate to trigger features obtained from variations in parameters of the annotated data stream, and derived features restoring relationships between various separators and corresponding descriptors.

Abstract: Method and system for generating a fingerprint representing a portion of an information signal. The method comprises decomposing a portion of the information signal into plural frequency sub bands at a decomposition level, calculating a spectral property of the signal in each of said plural frequency sub bands, comparing each spectral property against a first criterion thereby generating a comparison result, combining each comparison result for constituting the fingerprint, at least once repeating the decomposing, calculating, comparing and combining, wherein for each repetition the decomposing is performed using a decomposition level different from a previous decomposition level.

Abstract: Client terminals report an easy-to-calculate identifier such as the Internet URL or a cryptographic hash of the content to a server. The server collects and counts the reported identifiers so as to obtain preliminary statistics. By aggregating these reported identifiers into the preliminary statistics, identifiers are revealed that are likely popular content. The server selects one or more identifiers from the preliminary statistics and makes these available to at least a subset of clients. The clients that obtain these one or more identifiers then access content and compute the easy-to-calculate identifiers as usual. If the computed identifier matches one of the identifiers obtained from the server, the client will additionally extract a watermarked identifier or compute a digital fingerprint of the content in question and report this to the server. The server then uses the received identifier or fingerprint to create final statistics by aggregating the preliminary statistics.

Abstract: Client terminals report an easy-to-calculate identifier such as the Internet URL or a cryptographic hash of the content to a server. The server collects and counts the reported identifiers so as to obtain preliminary statistics. By aggregating these reported identifiers into the preliminary statistics, identifiers are revealed that are likely popular content. The server selects one or more identifiers from the preliminary statistics and makes these available to at least a subset of clients. The clients that obtain these one or more identifiers then access content and compute the easy-to-calculate identifiers as usual. If the computed identifier matches one of the identifiers obtained from the server, the client will additionally extract a watermarked identifier or compute a digital fingerprint of the content in question and report this to the server. The server then uses the received identifier or fingerprint to create final statistics by aggregating the preliminary statistics.

Abstract: An inserted segment of a video data stream is detected if no graphical object is detected. The presence of at least one active graphical object in the video data stream is detected concurrently with detecting appearance of a new graphical object in the video data stream. The most reliable graphical object from the at least one active graphical object and the new graphical object is determined and the presence of the most reliable graphical object is detected from a point in the video data stream at the new graphical object was detected to appear. An inserted segment of the video data stream is detected if no graphical object is detected.

Abstract: A first transition between a first video segment and a second video segment is detected by a first detector (105). A second transition is detected by a second detector (107). The first and second detectors (105, 107) are different. The outputs of the first and second detectors (105, 107) are compared (109). The reliability of the second method is determined by comparing the transition detected by the first detector (105) with the transition detected by the second detector (107). At least the second transition is used to determine a final transition if the second method is determined to be reliable. The second transition is not used to determine the final transition if the second method is determined to be unreliable.

Abstract: Advanced watermark embedders use psycho visual/acoustic models to minimize perception of the embedded watermark in media contents. However, it is expensive to implement such advanced watermark embedders in consumer appliances (3) such as DVD recorders. In accordance with the invention, embedding strength parameters (JND) are calculated (13) off-line at a remote location (1) and transmitted along with the media contents to the consumer appliance (3). A relatively simple embedder (33) uses the received parameters to control the strength of the watermark (36) to be embedded. The parameters may be accommodated in the transmitted signal as payload of a fragile watermark or as user data in an encrypted MPEG stream.

Abstract: Watermark detectors have a buffer in which a number of image tiles are folded and accumulated prior to computing the correlation between buffer contents and the watermark pattern being looked for. The intention of the folding and accumulation process is to average out the video content while accumulating the embedded watermark energy. This no longer appears to hold for strongly compressed video, such as DIVX, which exhibits a lot of artificial noise and undesired similarity (block patterns). As a result thereof, correlation peaks are often below the threshold. In a similar manner, the compression affects scale detection. According to this invention, only frames (or parts thereof) that are not so heavily compressed and therefore have a high probability of carrying enough watermark energy are folded and accumulated. To this end, a quality metric is calculated, the quality metric being indicative of the degree of compression of the data.

Abstract: A graphics card of a computer system is configured to detect watermarks within a video signal that includes one or more windows of a display. The frames of the video signal are analyzed to locate image areas in which the video signal changes, and a bounding box around these areas of change is defined as an area(s) of interest. Watermark detection is then performed within the area(s) of interest.

Abstract: A method and apparatus (200) for embedding a watermark in an information signal bit-stream are described. The method comprises receiving a portion of an information signal bit-stream (MPEG2). A first copy of the received portion is stored in a first buffer (220). A second copy of the received portion is watermarked, and the resulting watermarked signal stored in a second buffer (240). At predetermined intervals, a check is performed to determine if the bit-rate of the received portion has been changed by being watermarked. If the check determines the bit-rate has changed, the first copy of the received portion from the first buffer is output. Otherwise the watermarked signal from the second buffer is output.

Abstract: A method and arrangement are disclosed for processing a compressed media signal, for example, embedding a watermark in an MPEG2 video signal. The watermark, a spatial noise pattern (140), is embedded (123) by selectively discarding the smallest quantized DCT coefficients. The discarded coefficients are subsequently merged in the runs of other run/level pairs. To compensate for a too large reduction of the bit rate, some of the new run/level pairs are not variable-length encoded (124) but represented by longer code words according to further coding rule (125) providing such longer code words, for example, MPEG's “Escape coding”.

Abstract: A method and arrangement are disclosed for embedding a watermark in an MPEG compressed video stream. The watermark (a spatial noise pattern) is embedded by selectively discarding the smallest quantized DCT coefficients. The discarded coefficients are subsequently merged in the runs of the remaining coefficients. The decision whether a coefficient is discarded or not is made on the basis of a pre-calculated watermark buffer and the number of already discarded coefficients per 8×8 DCT block. The advantages of this method are (i) a very simple bit rate control system and (ii) no need for drift compensation. The algorithm can be implemented in a very efficient manner with respect to memory requirements and computational complexity.