Abstract: Methods, apparatus, systems and articles of manufacture are disclosed to select reference sub-fingerprints for comparison to query sub-fingerprints based on a determination that a query sub-fingerprint is a match with a reference sub-fingerprint, generate a count vector that stores total counts of matches between the query sub-fingerprints and different subsets of the reference sub-fingerprints, each of the different subsets being aligned to the query sub-fingerprints at a different offset from a reference point, each of the different offsets being mapped by the count vector to a different total count, calculate a maximum count among the total counts, a median of the total counts, and a difference between the maximum count and the median of the total counts, and classify the reference sub-fingerprints as a match with the query sub-fingerprints based on the difference between the maximum count in the count vector and the median.

Abstract: Methods and systems for automated video segmentation are disclosed. A sequence of video frames having video segments of contextually-related sub-sequences may be received. Each frame may be labeled according to segment and segment class. A video graph may be constructed in which each node corresponds to a different frame, and each edge connects a different pair of nodes, and is associated with a time between video frames and a similarity metric of the connected frames. An artificial neural network (ANN) may be trained to predict both labels for the nodes and clusters of the nodes corresponding to predicted membership among the segments, using the video graph as input to the ANN, and ground-truth clusters of ground-truth labeled nodes. The ANN may be further trained to predict segment classes of the predicted clusters, using the segment classes as ground truths. The trained ANN may be configured for application runtime video sequences.

Abstract: Systems and methods for monitoring of icon in an external display device are disclosed. Images of an icon displayed in a display device may be continually captured as video frames by a video camera of an icon monitoring system. While operating in a first mode, video frames may be continually analyzed to determine if the captured image matches an active template icon known to match the captured image of the icon. While the captured image matches the active template icon, operating in the first mode continues. Upon detecting a failed match to the active template icon, the system starts operating in a second to search among known template icons for a new match. Upon finding a new match, the active template icon may be updated to the new match, and operation switches back to the first mode. Times of transitions between the first and second modes may be recorded.

Abstract: Methods, apparatus, systems and articles of manufacture are disclosed for efficient media indexing. An example method disclosed herein includes means for initiating a list of hash seeds, the list of hash seeds including at least a first hash seed value and a second hash seed value among other hash seed values, means for generating to generate a first bucket distribution based on the first hash seed value and a first hash function and generate a second bucket distribution based on the second hash seed value used in combination with the first hash seed value, means for determining to determine a first entropy value of the first bucket distribution, wherein data associated with the first bucket distribution is stored in a first hash table and determine a second entropy value of the second bucket distribution.

Abstract: Methods, apparatus, systems and articles of manufacture are disclosed for efficient media indexing. An example method disclosed herein includes means for initiating a list of hash seeds, the list of hash seeds including at least a first hash seed value and a second hash seed value among other hash seed values, means for generating to generate a first bucket distribution based on the first hash seed value and a first hash function and generate a second bucket distribution based on the second hash seed value used in combination with the first hash seed value, means for determining to determine a first entropy value of the first bucket distribution, wherein data associated with the first bucket distribution is stored in a first hash table and determine a second entropy value of the second bucket distribution.

Abstract: Systems and methods for monitoring of icon in an external display device are disclosed. Images of an icon displayed in a display device may be continually captured as video frames by a video camera of an icon monitoring system. While operating in a first mode, video frames may be continually analyzed to determine if the captured image matches an active template icon known to match the captured image of the icon. While the captured image matches the active template icon, operating in the first mode continues. Upon detecting a failed match to the active template icon, the system starts operating in a second to search among known template icons for a new match. Upon finding a new match, the active template icon may be updated to the new match, and operation switches back to the first mode. Times of transitions between the first and second modes may be recorded.

Abstract: Disclosed herein are system, method, and computer program product embodiments for a computing system with a content-characteristic based trigger feature. An embodiment operates by determining a hybrid system of transmitting a fingerprint of media is being used to transmit one or more media fingerprints. A first fingerprint is generated from a first portion of media content and a second fingerprint is generated from of a second portion of the media content. It is determined that the first portion of media content and the second portion of the media content lack a threshold extent of similarity with each other. The second fingerprint is provided to a content identification server to identify the second portion of the media content. An action is performed based on an identified second portion of the media content.

Abstract: A client device accesses a video input stream from an intermediate device for display. The client device analyzes the video input stream to determine that the video input stream matches a template corresponding to a screen portion. Based on the video input stream matching the template, a video output stream is generated and caused to be presented on a display. In some example embodiments, the analysis is performed while the client device is replacing video content received from a content source via the intermediate device. For example, commercials transmitted from a national content provider to a smart TV via a set-top box may be replaced with targeted commercials. During the replacement, menus generated by the set-top box may be detected and the replacement video altered by the smart TV to include the menus.

Abstract: In one aspect, an example method includes receiving, by a computing system, media content; generating, by the computing system, a fingerprint of the received media content; determining, by the computing system, that a channel-change operation was performed; responsive to determining that the channel-change operation was performed, transmitting, by the computing system, the generated fingerprint to a content identification server to identify the received media content; and performing an action based on the identified media content.

Abstract: In one aspect, an example method includes (i) receiving, by a computing system, media content; (ii) generating, by the computing system, a fingerprint of the received media content; (iii) determining, by the computing system, that a channel-change operation was performed; (iv) responsive to determining that the channel-change operation was performed, transmitting, by the computing system, the generated fingerprint to a content identification server to identify the received media content; and (v) performing an action based on the identified media content.

Abstract: Methods and systems for automated video segmentation are disclosed. A sequence of video frames having video segments of contextually-related sub-sequences may be received. Each frame may be labeled according to segment and segment class. A video graph may be constructed in which each node corresponds to a different frame, and each edge connects a different pair of nodes, and is associated with a time between video frames and a similarity metric of the connected frames. An artificial neural network (ANN) may be trained to predict both labels for the nodes and clusters of the nodes corresponding to predicted membership among the segments, using the video graph as input to the ANN, and ground-truth clusters of ground-truth labeled nodes. The ANN may be further trained to predict segment classes of the predicted clusters, using the segment classes as ground truths. The trained ANN may be configured for application runtime video sequences.

Abstract: Systems and methods for using radar-based detection in a home environment and utilizing this detection for identifying human behavior are provided. A system, according to one implementation, includes a processing device and a memory device configured to store instructions that, when executed, enable the processing device to obtain movement data from one or more radar-based sensing devices arranged within a predefined setting. The instructions further enable the processing device to analyze the movement data to identify a human activity and to determine one or more characteristics of the human activity. Also, the processing device is configured to analyze the human activity and the one or more characteristics of the human activity to determine an identity of a person performing the human activity.

Abstract: Methods and systems for automated video segmentation are disclosed. A sequence of video frames having video segments of contextually-related sub-sequences may be received. Each frame may be labeled according to segment and segment class. A video graph may be constructed in which each node corresponds to a different frame, and each edge connects a different pair of nodes, and is associated with a time between video frames and a similarity metric of the connected frames. An artificial neural network (ANN) may be trained to predict both labels for the nodes and clusters of the nodes corresponding to predicted membership among the segments, using the video graph as input to the ANN, and ground-truth clusters of ground-truth labeled nodes. The ANN may be further trained to predict segment classes of the predicted clusters, using the segment classes as ground truths. The trained ANN may be configured for application runtime video sequences.

Abstract: In one aspect, an example method includes (i) receiving, by a computing system, media content; (ii) generating, by the computing system, a fingerprint of a portion of the received media content; (iii) determining, by the computing system, that the received media content has a predefined characteristic; (iv) responsive to determining that the received media content has the predefined characteristic, transmitting, by the computing system, the generated fingerprint to a content identification server to identify the portion of the received media content; and (v) performing an action based on the identified portion of media content.

Abstract: Methods and systems for automated video segmentation are disclosed. A sequence of video frames having video segments of contextually-related sub-sequences may be received. Each frame may be labeled according to segment and segment class. A video graph may be constructed in which each node corresponds to a different frame, and each edge connects a different pair of nodes, and is associated with a time between video frames and a similarity metric of the connected frames. An artificial neural network (ANN) may be trained to predict both labels for the nodes and clusters of the nodes corresponding to predicted membership among the segments, using the video graph as input to the ANN, and ground-truth clusters of ground-truth labeled nodes. The ANN may be further trained to predict segment classes of the predicted clusters, using the segment classes as ground truths. The trained ANN may be configured for application runtime video sequences.

Abstract: A client device accesses a video input stream from an intermediate device for display. The client device analyzes the video input stream to determine that the video input stream matches a template corresponding to a screen portion. Based on the video input stream matching the template, a video output stream is generated and caused to be presented on a display. In some example embodiments, the analysis is performed while the client device is replacing video content received from a content source via the intermediate device. For example, commercials transmitted from a national content provider to a smart TV via a set-top box may be replaced with targeted commercials. During the replacement, menus generated by the set-top box may be detected and the replacement video altered by the smart TV to include the menus.

Abstract: Methods and systems for disambiguating musical artist names are disclosed. Musical-artist-release records (MARRs) may be input to a multi-modal artificial neural network (ANN). Each MARR may be associated with a musical release of an artist, and may include a release ID and an artist ID, and release data in categories including music media content and metadata categories including sub-definitive musician name of the artist and release subcategories. All n-tuples of MARRs may be formed, and for each n-tuple, the ANN may be applied concurrently to each MARR to generate a release feature vector (RFV) that includes a set of sub-feature vectors, each characterizing a different category of release data. For each n-tuple, the ANN may be trained to cluster in a multi-dimensional RFV space RFVs of the same artist ID, and to separate RFVs of different artist IDs. The MARRs and their RFVs may be stored in a release database.

Abstract: A client device accesses a video input stream from an intermediate device for display. The client device analyzes the video input stream to determine that the video input stream matches a template indicating a change of speed. In some example embodiments, the analysis is performed while the client device is generating a video output stream that replaces video content received from a content source via the intermediate device. Based on the video input stream matching the template, the client device modifies the video output stream to show the change of speed. For example, commercials transmitted from a national content provider to a smart TV via a set-top box may be replaced with targeted commercials. During the replacement, fast-forwarding of the commercial may be detected by matching a template with the video input stream. While the video input stream is being fast-forwarded, the smart TV fast-forwards the targeted commercial.

Abstract: Methods, apparatus, systems and articles of manufacture are disclosed for efficient media indexing. An example method disclosed herein includes means for initiating a list of hash seeds, the list of hash seeds including at least a first hash seed value and a second hash seed value among other hash seed values, means for generating to generate a first bucket distribution based on the first hash seed value and a first hash function and generate a second bucket distribution based on the second hash seed value used in combination with the first hash seed value, means for determining to determine a first entropy value of the first bucket distribution, wherein data associated with the first bucket distribution is stored in a first hash table and determine a second entropy value of the second bucket distribution.

Abstract: Methods, apparatus, systems and articles of manufacture are disclosed for efficient media indexing. An example method disclosed herein includes selecting a first hash seed value based on a first entropy value calculated for a first bucket distribution resulting from use of the first hash seed value to store data in a first hash table, selecting a second hash seed value to be used in combination with the first hash seed value based on a second entropy value calculated on a second bucket distribution resulting from use of the first hash seed value in combination with the second hash seed value, and storing data in the first hash table based on the first hash seed value and a second hash table based on the second hash seed value.