Abstract: Camera equipment used for capturing a live event is expensive and may require professionals to operate the equipment. Such camera equipment is not accessible for capturing events and experiences by everyday users. Suppose media content is captured by lower-cost input devices such as mobile phones with a video camera. A production platform can use computer vision techniques to produce enhanced media content from media content captured by the lower-cost input devices. In one example, the production platform can aggregate and analyze the received media content to extract information and determine blind spots. The production platform can create or retrieve content based on the extracted information and insert the content in the blind spots. The production platform can deliver enhanced media content to an end user device, the enhanced media content may have the received media content and the inserted content.

Abstract: Disclosed herein are various embodiments, for a private listening system for streaming audio and video. An example embodiment operates by receiving a first selection of a first version of audio content for output to a first audio device connected to a streaming media device. A second selection of a second version of the audio content for output to a second audio device connected to the streaming media device is received. Both the first version of the audio content is provided to the first audio device and the second version of the audio content is provided to the second audio device for simultaneous and synchronized output with the video content output to the video device.

Abstract: Disclosed herein are various embodiments for providing content searching for people with visual impairments or blindness. An embodiment operates by receiving a command to search multimedia content including both video content and audio content. One or more scene changes, including a first scene change, corresponding to the video content are determined. The search command is executed on the multimedia content. It is detected that the multimedia content has reached the first scene change responsive to the executing the search command. An audio cue t is audibly output responsive to the detection.

Abstract: In one aspect, an example method includes (i) estimating, using a skeletal detection model, a pose of an original actor for each of multiple frames of a video; (ii) obtaining, for each of a plurality of the estimated poses, a respective image of a replacement actor; (iii) obtaining replacement speech in the replacement actor's voice that corresponds to speech of the original actor in the video; (iv) generating, using the estimated poses, the images of the replacement actor, and the replacement speech, synthetic frames corresponding to the multiple frames of the video that depict the replacement actor in place of the original actor, with the synthetic frames including facial expressions for the replacement actor that temporally align with the replacement speech; and (iv) combining the synthetic frames and the replacement speech so as to obtain a synthetic video that replaces the original actor with the replacement actor.

Abstract: Disclosed herein are various embodiments for server-side scene change content stitching. An example embodiment operates by providing first content for display on a display device. A point in a timeline of the first content when second content is to be displayed is identified. A scene change in the first content closest to the point in the timeline is identified. A first frame having a distance X prior to the scene change is identified, and the first content for display. The second content is provided for display on the display device beginning at the scene change. At a conclusion of the second content, the display is paused, and seeking and decoding the first content to the scene change by decoding the first content to the identified first frame plus the distance X is performed. The decoded first content, from the scene change, is provided to the display device for display.

Abstract: Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for utilizing a content acquisition recommendation system to generating a set of candidate content assets, generate embeddings and popularity score estimates for the set of candidate content assets, aggregate the set of candidate content assets with a set of existing content assets to generate a simulation set of content assets, determine a target set of users for the simulation set of content assets, generate, for at least a portion of the target set of users and based on a trained machine learning model, a result set of recommended content assets, determining an impact of the candidate content assets located in the result set of recommended content assets and generate a proposal for an acquisition of candidate content assets.

Abstract: Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for real-time online learning for short-form content ranking. An example embodiment operates by receiving a user feedback event that includes user feedback on a short-form content item played using a media device, and publishing a message with the event to a publish-subscribe (pub/sub) topic. Retrieved with a pull operation, the user feedback event from the message is saved to a real-time data store. A recommendation engine is triggered to generate a playlist batch of short-form content recommendations based on the user feedback event. Short-form content items specified in the playlist batch are transmitted to the media device for playback. User feedback on short-form content items in an immediately preceding playlist batch during the same usage session can inform the selections made in the generated playlist batch.

Abstract: Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for modifying one or more parameters of a data streaming bitrate selection algorithm based on machine learning. An example embodiment operates by training and operating a first machine learning model to predict a sustainable network bandwidth. A second machine learning model is trained to receive the sustainable network bandwidth and predict a likelihood that this network bandwidth will not empty a data buffer of streaming data. A bitrate is selected based on the likelihood being below a threshold percentage, such as 50%.

Abstract: Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for distributing the performance of speech recognition among a remote control device and a voice platform in the cloud. In some embodiments, the remote control device operates to receive a voice input from a user. The remote control device detects a trigger word in the voice input. The remote control device then processes the voice input. The remote control device then transmits the voice input to a voice platform based on the detecting in order to determine an intent associated with the voice input.

Abstract: Dialogue state tracking for voice assistants involves correctly tracking intent and entities of a task that a user is performing. A dialogue state, having a tracked intent and one or more tracked entities, can then be used to perform the task. Building a dialogue state tracking system within a voice assistant is not trivial. In some embodiments, a dialogue state tracking system involving one or more large language models can be implemented downstream of a natural language understanding system to produce the tracked intent and the one or more tracked entities. In some embodiments, a dialogue state tracking system involving one or more large language models can be implemented upstream of a natural language understanding system to produce rephrased natural language text, which is in turn processed by the natural language understanding system to produce the tracked intent and the one or more tracked entities.

Abstract: Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for establishing a communication link between a battery-powered device and a WAP to reduce the power consumption burden of the battery-powered device. In an embodiment, a communication context of the battery-powered device and the WAP is determined, for example, by determining whether there is at least one intermediate device having a wired power source via which the battery-powered device can communicate with the WAP. A communication scheme of a plurality of different communication schemes is selected, for example, by selecting the communication scheme of the plurality of different communication schemes that imposes the lowest power consumption burden on the battery-powered device and is also supported by the communication context.

Abstract: Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for sharing video streams in a camera system network. The video streams may be modified and/or masked to obscure private video stream features from being shared. An example embodiment operates by linking a first camera system to a first user device. The first camera system detects a second camera system in its field of view, wherein the second camera system is linked to a second user device. Upon exchanging requests to share video streams, the first and second user devices may view the video streams from both camera systems. The video streams may be modified to mask and/or obscure private property and/or known persons to preserve privacy despite sharing video streams. For example, the first camera system may modify and stream a masked video stream to the second user device.

Abstract: Disclosed herein are system, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for optimizing content item placements on a content publisher application of a media device based on multi-publisher content item measurement data associated with a set of media devices. An example embodiment operates by receiving multi-publisher content item measurement data that comprises data signals for a plurality of content items from a plurality of different publishers that are presented at a plurality of different households. A content item, from among the plurality of content items, is determined to be transmitted to a media device of a particular household for presentation via a content publisher application of the media device. In response to determining that the content item is to be transmitted to the media device, the content item is transmitted to the media device for presentation via the content publisher application of the media device.

Abstract: Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for a wireless media system including a wireless speaker device and a host device forming a peer-to-peer wireless network. The host device provides audio content to the speaker device, and the speaker device can decode and process the audio content received from the host device. The speaker device can be in various states that consume different level of powers, such as a standby state, a wake state, a connected state. The speaker device can determine the host device is in an inactive state or an active state. When the speaker device determines the host device is in an inactive state, the speaker device can enter the standby state, enable the communication circuit and the audio processing circuit of the speaker device to be in a low power state to save power.

Abstract: Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for stochastic candidate selection for content recommendation. An example embodiment operates by a computer-implemented method for stochastic candidate selection for content recommendation. The method includes receiving, by at least one computer processor, a first plurality of content candidates and selecting a second plurality of content candidates from the first plurality of content candidates. The method further include ranking the second plurality of content candidates based on one or more parameters and selecting a third plurality of content candidates from the ranked second plurality of content candidates. The method can further include displaying the third plurality of content candidates using a display device.

Abstract: A method and system for controlling media-content presentation. An example method includes a computing system detecting that a media-player device has moved from a home location of the media-player device to a visited location and that the media-player device is connected with a media-presentation device at the visited location. Further, the example method includes, responsive to the detecting, the computing system causing the media-player device to output, for presentation by the media-presentation device at the visited location, media content selected by the computing system based on the media-player device having moved from the home location to the visited location. For instance, the computing system could cause the media-player device to output, for presentation by the media-presentation device, a message welcoming a user to the visited location and/or information about the visited location such as information about tourist attractions at the visited location for instance.

Abstract: Use of watermarking in a media stream as a trigger for carrying out a fingerprint-related action, such as starting to generate fingerprints of the media stream, increasing a rate of generation of fingerprints of the media stream, starting to report fingerprints of the media stream, requesting fingerprints, or engaging in a fingerprint comparison. In an implementation, a media client that is processing the media stream could detect the watermark in the media stream and could responsively carry out or cause to be carried out a fingerprint-related action. Or an entity in a media-distribution path along which the media stream flows to a media client could detect the watermark in the media stream and could responsively carry out or cause to be carried out a fingerprint-related action. Further, the watermark could be inserted into the media stream in the first place based on transport-stream carried metadata, among other possibilities.

Abstract: Device discovery on a local network can cause suspended devices to not respond or to wake up and respond, both of which may be undesirable. To enable suspended devices to respond to a device discovery request without waking up, network interface controllers (NIC) are programmed with a device discovery protocol. A first device broadcasts a discovery frame that includes information indicating that it is a discovery frame. The NIC of a second device determines whether the received frame is a discovery frame. In response to receiving a discovery frame, the NIC of the second device returns a response that includes its device identifier. The NIC may issue the response frame directly, without waking up the second device. Alternatively, the NIC causes the second device to enter a discovery state, which enables the second device to respond to a discovery packet and then return to a suspend mode.

Abstract: Disclosed herein are system, apparatus, article of manufacture, method and/or computer program product embodiments, and/or combinations and sub-combinations thereof, for ad insertion by a display device coupled to a media device via a high-definition media interface (HDMI) connection, where the media device provides media content and/or a control signal. When the media device pauses the media content, the display device can determine that a pause event has occurred and insert an ad shown on the display device. Further, some embodiments include determining the context and/or content of the media content that is paused, and determining an ad that is customized to the determined context and/or content to be displayed on the display device. In some embodiments, the display device can determine additional information from the control signal that may also be used to determine the ad to be displayed on the display device.

Abstract: A computer-implemented method is described and includes automatically creating a group profile for an account associated with a streaming service platform, wherein the group profile is informed by a plurality of individual user profiles associated with the account, wherein each of the individual user profiles is associated with a different one of a plurality of users; presenting via a user interface (UI) of the streaming service platform representations of each of the group profile and the individual user profiles; prompting one of the users to select one of the profiles to use to view content items via the streaming service platform; and, subsequent to selection by the user of the group profile, presenting via the UI a collection of content item recommendations, wherein the content item recommendations are informed by contents of the individual user profiles collectively.