Abstract: In one aspect, an example method for generating a candidate image for use as backdrop imagery for a graphical user interface is disclosed. The method includes receiving a raw image and determining an edge image from the raw image using edge detection. The method also includes identifying a candidate region of interest (ROI) in the raw image based on the candidate ROI enclosing a portion of the edge image having edge densities exceeding a threshold edge density. The method also includes manipulating the raw image relative to a backdrop imagery canvas for a graphical user interface based on a location of the candidate ROI within the raw image. The method also includes generating, based on the manipulating, a set of candidate backdrop images in which at least a portion of the candidate ROI occupies a preselected area of the backdrop imagery canvas, and storing the set of candidate backdrop images.

Abstract: In one aspect, an example method includes (i) extracting a sequence of audio features from a portion of a sequence of media content; (ii) extracting a sequence of video features from the portion of the sequence of media content; (iii) providing the sequence of audio features and the sequence of video features as an input to a transition detector neural network that is configured to classify whether or not a given input includes a transition between different content segments; (iv) obtaining from the transition detector neural network classification data corresponding to the input; (v) determining that the classification data is indicative of a transition between different content segments; and (vi) based on determining that the classification data is indicative of a transition between different content segments, outputting transition data indicating that the portion of the sequence of media content includes a transition between different content segments.

Abstract: Example systems and methods may selection of video frames using a machine learning (ML) predictor program are disclosed. The ML predictor program may generate predicted cropping boundaries for any given input image. Training raw images associated with respective sets of training master images indicative of cropping characteristics for the training raw image may be input to the ML predictor, and the ML predictor program trained to predict cropping boundaries for raw image based on expected cropping boundaries associated training master images. At runtime, the trained ML predictor program may be applied to runtime raw images in order to generate respective sets of runtime cropping boundaries corresponding to different cropped versions of the runtime raw image. The runtime raw images may be stored with information indicative of the respective sets of runtime boundaries.

Abstract: In one aspect, a method includes detecting a fingerprint match between query fingerprint data representing at least one audio segment within podcast content and reference fingerprint data representing known repetitive content within other podcast content, detecting a feature match between a set of audio features across multiple time-windows of the podcast content, and detecting a text match between at least one query text sentences from a transcript of the podcast content and reference text sentences, the reference text sentences comprising text sentences from the known repetitive content within the other podcast content. The method also includes responsive to the detections, generating sets of labels identifying potential repetitive content within the podcast content. The method also includes selecting, from the sets of labels, a consolidated set of labels identifying segments of repetitive content within the podcast content, and responsive to selecting the consolidated set of labels, performing an action.

Abstract: Example systems and methods of selection of video frames using a machine learning (ML) predictor program are disclosed. The ML predictor program may generate predicted cropping boundaries for any given input image. Training raw images associated with respective sets of training master images indicative of cropping characteristics for the training raw image may be input to the ML predictor, and the ML predictor program trained to predict cropping boundaries for raw image based on expected cropping boundaries associated training master images. At runtime, the trained ML predictor program may be applied to a sequence of video image frames to determine for each respective video image frame a respective score corresponding to a highest statistical confidence associated with one or more subsets of cropping boundaries predicted for the respective video image frame. Information indicative of the respective video image frame having the highest score may be stored or recorded.

Abstract: A machine is configured to identify a media file that, when played to a user, is likely to modify an emotional or physical state of the user to or towards a target emotional or physical state. The machine accesses play counts that quantify playbacks of media files for the user. The playbacks may be locally performed or detected by the machine from ambient sound. The machine accesses arousal scores of the media files and determines a distribution of the play counts over the arousal scores. The machine uses one or more relative maxima in the distribution in selecting a target arousal score for the user based on contextual data that describes an activity of the user. The machine selects one or more media files based on the target arousal score. The machine may then cause the selected media file to be played to the user.

Abstract: In one aspect, an example method for generating a candidate image for use as backdrop imagery for a graphical user interface is disclosed. The method includes receiving a raw image and determining an edge image from the raw image using edge detection. The method also includes identifying a candidate region of interest (ROI) in the raw image based on the candidate ROI enclosing a portion of the edge image having edge densities exceeding a threshold edge density. The method also includes manipulating the raw image relative to a backdrop imagery canvas for a graphical user interface based on a location of the candidate ROI within the raw image. The method also includes generating, based on the manipulating, a set of candidate backdrop images in which at least a portion of the candidate ROI occupies a preselected area of the backdrop imagery canvas, and storing the set of candidate backdrop images.

Abstract: An example method may include receiving, at a computing device, a digital image associated with a particular media content program, the digital image containing one or more faces of particular people associated with the particular media content program. A computer-implemented automated face recognition program may be applied to the digital image to recognize, based on at least one feature vector from a prior-determined set of feature vectors, one or more of the particular people in the digital image, together with respective geometric coordinates for each of the one or more detected faces. At least a subset of the prior-determined set of feature vectors may be associated with a respective one of the particular people. The digital image together may be stored in non-transitory computer-readable memory, together with information assigning respective identities of the recognized particular people, and associating with each respective assigned identity geometric coordinates in the digital image.

Abstract: Apparatus, systems, articles of manufacture, and methods for volume adjustment are disclosed herein. An example method includes collecting data corresponding to a volume of an audio signal as the audio signal is output through a device, when an average volume of the audio signal does not satisfy a volume threshold for a specified timespan, determining a difference between the average volume and a desired volume, and applying a gain to the audio signal to adjust the volume of the audio signal to the desired volume, the gain determined based on the difference between the average volume and the desired volume.

Abstract: In one aspect, an example method includes (i) extracting a sequence of audio features from a portion of a sequence of media content; (ii) extracting a sequence of video features from the portion of the sequence of media content; (iii) providing the sequence of audio features and the sequence of video features as an input to a transition detector neural network that is configured to classify whether or not a given input includes a transition between different content segments; (iv) obtaining from the transition detector neural network classification data corresponding to the input; (v) determining that the classification data is indicative of a transition between different content segments; and (vi) based on determining that the classification data is indicative of a transition between different content segments, outputting transition data indicating that the portion of the sequence of media content includes a transition between different content segments.

Abstract: A machine is configured to identify a media file that, when played to a user, is likely to modify an emotional or physical state of the user to or towards a target emotional or physical state. The machine accesses play counts that quantify playbacks of media files for the user. The playbacks may be locally performed or detected by the machine from ambient sound. The machine accesses arousal scores of the media files and determines a distribution of the play counts over the arousal scores. The machine uses one or more relative maxima in the distribution in selecting a target arousal score for the user based on contextual data that describes an activity of the user. The machine selects one or more media files based on the target arousal score. The machine may then cause the selected media file to be played to the user.

Abstract: Apparatus, systems, articles of manufacture, and methods for volume adjustment are disclosed herein. An example method includes collecting data corresponding to a volume of an audio signal as the audio signal is output through a device, when an average volume of the audio signal does not satisfy a volume threshold for a specified timespan, determining a difference between the average volume and a desired volume, and applying a gain to the audio signal to adjust the volume of the audio signal to the desired volume, the gain determined based on the difference between the average volume and the desired volume.

Abstract: In one aspect, an example method includes (i) retrieving, from a text index, closed captioning repetition data for a segment of a sequence of media content; (ii) generating features using the closed captioning repetition data; (iii) providing the features as input to a classification model, wherein the classification model is configured to output classification data indicative of a likelihood of the features being characteristic of a program segment; (iv) obtaining the classification data output by the classification model; (v) determining a prediction of whether the segment is a program segment using the classification data; and (vi) storing the prediction for the segment in a database.

Abstract: A clustering machine can cluster descriptive vectors in a balanced manner. The clustering machine calculates distances between pairs of descriptive vectors and generates clusters of vectors arranged in a hierarchy. The clustering machine determines centroid vectors of the clusters, such that each cluster is represented by its corresponding centroid vector. The clustering machine calculates a sum of inter-cluster vector distances between pairs of centroid vectors, as well as a sum of intra-cluster vector distances between pairs of vectors in the clusters. The clustering machine calculates multiple scores of the hierarchy by varying a scalar and calculating a separate score for each scalar. The calculation of each score is based on the two sums previously calculated for the hierarchy. The clustering machine may select or otherwise identify a balanced subset of the hierarchy by finding an extremum in the calculated scores.

Abstract: An example method may include receiving, at a computing device, a digital image associated with a particular media content program, the digital image containing one or more faces of particular people associated with the particular media content program. A computer-implemented automated face recognition program may be applied to the digital image to recognize, based on at least one feature vector from a prior-determined set of feature vectors, one or more of the particular people in the digital image, together with respective geometric coordinates for each of the one or more detected faces. At least a subset of the prior-determined set of feature vectors may be associated with a respective one of the particular people. The digital image together may be stored in non-transitory computer-readable memory, together with information assigning respective identities of the recognized particular people, and associating with each respective assigned identity geometric coordinates in the digital image.

Abstract: In one aspect, an example method includes (i) determining, by a computing system, a mean image of a set of frames of video content; (ii) extracting, by the computing system, a reference template of static content from the mean image; (iii) identifying, by the computing system, the extracted reference template of static content in a frame of the set of frames of the video content; (iv) labeling a segment within the video content as either a program segment or an advertisement segment based on the identifying of the extracted reference template of static content in the frame of the video content; and (v) generating data identifying the labeled segment.

Abstract: A method and system for computer-based generation of podcast metadata, to facilitate operations such as searching for and recommending podcasts based on the generated metadata. In an example method, a computing system obtains a text representation of a podcast episode and obtains person data defining a list of person names such as celebrity names. The computing system then correlates the person data with the text representation, to find a match between a listed person name a text string in the text representation. Further, the computing system predicts a named-entity span in the text representation and determines that the predicted named-entity span matches a location of the text string in the text representation of the podcast episode, and based on this determination, the computing system generates and outputs metadata that associates the person name with the podcast episode.

Abstract: Example systems and methods for automated generation of banner images are disclosed. A program identifier associated with a particular media program may be received by a system, and used for accessing a set of iconic digital images and corresponding metadata associated with the particular media program. The system may select a particular iconic digital image for placing a banner of text associated with the particular media program, by applying an analytical model of banner-placement criteria to the iconic digital images. The system may apply another analytical model for banner generation to the particular iconic image to determine (i) dimensions and placement of a bounding box for containing the text, (ii) segmentation of the text for display within the bounding box, and (iii) selection of font, text size, and font color for display of the text. The system may store the particular iconic digital image and banner metadata specifying the banner.

Abstract: Example systems and methods of selection of video frames using a machine learning (ML) predictor program are disclosed. The ML predictor program may generate predicted cropping boundaries for any given input image. Training raw images associated with respective sets of training master images indicative of cropping characteristics for the training raw image may be input to the ML predictor, and the ML predictor program trained to predict cropping boundaries for raw image based on expected cropping boundaries associated training master images. At runtime, the trained ML predictor program may be applied to a sequence of video image frames to determine for each respective video image frame a respective score corresponding to a highest statistical confidence associated with one or more subsets of cropping boundaries predicted for the respective video image frame. Information indicative of the respective video image frame having the highest score may be stored or recorded.

Abstract: A clustering machine can cluster descriptive vectors in a balanced manner. The clustering machine calculates distances between pairs of descriptive vectors and generates clusters of vectors arranged in a hierarchy. The clustering machine determines centroid vectors of the clusters, such that each cluster is represented by its corresponding centroid vector. The clustering machine calculates a sum of inter-cluster vector distances between pairs of centroid vectors, as well as a sum of intra-cluster vector distances between pairs of vectors in the clusters. The clustering machine calculates multiple scores of the hierarchy by varying a scalar and calculating a separate score for each scalar. The calculation of each score is based on the two sums previously calculated for the hierarchy. The clustering machine may select or otherwise identify a balanced subset of the hierarchy by finding an extremum in the calculated scores.