Abstract: A method for reducing artifacts in streaming video implemented by a computing device involves determining whether a video quality assessment score for a current segment of a channel of a video stream with a first resolution and a first bit rate meets a quality threshold, replacing the current segment with a replacement segment with a lower resolution than the first resolution, in response to the current segment failing to meet the quality threshold, and publishing the replacement segment at the first bit rate as part of the video stream.

Abstract: A method and system for managing transcoding of data in a stream that includes identifying an input source change for the stream with a new input source type, and adding a decoder for the new input source type, the decoder configured to output for a respective encoder in a transcoder pipeline.

Abstract: A method to highlight video segments in a video stream, where the method includes receiving a video stream from a video source, identifying a highlight segment within the video stream based on a machine learning model, the highlight segment being deemed to be worthy of replay by the machine learning model, and starting and ending frames of the highlight segment being identified by applying the machine learning model to the video stream and corresponding audio data, and providing an availability indication of the highlight segment in the video stream once the starting and ending frames of the highlight segment are identified.

Abstract: A method and system corrects a content delivery infrastructure. The method of the system includes receiving a request to resolve reported issues for the content delivery infrastructure, collecting content delivery metrics for the content delivery infrastructure, executes a language transformer model on the request and the content delivery metrics to generate a set of possible resolutions with confidence ratings, and implementing an automated solution based on a resolution from the set of possible resolutions, in response to the resolution having a confidence rating above a threshold.

Abstract: A method for video quality assessment implemented by a computing device, where the method includes selecting a first set of frames from an input video source, determining features of the first set of frames for input into a first machine learning model, applying the first machine learning model to the features of the first set of frames to obtain a video quality score and a confidence score, and triggering a retraining of the first machine learning model in response to the confidence score being determined to be out of bounds.

Abstract: A method for reducing artifacts in streaming video implemented by a computing device involves determining whether a video quality assessment score for a current segment of a channel of a video stream with a first resolution and a first bit rate meets a quality threshold, replacing the current segment with a replacement segment with a lower resolution than the first resolution, in response to the current segment failing to meet the quality threshold, and publishing the replacement segment at the first bit rate as part of the video stream.

Abstract: A method of measuring video stream visual stability, the method including receiving a first set of network packets carrying data of the video stream, determining network performance metrics for a session associated with the first set of network packets, retrieving priority fault errors from a packet header of at least one network packet of the first set of the network packets, adding the priority fault errors and the network performance metrics to time series data, and applying a machine learning model to the time series data to obtain a visual stability score for the first set of network packets.

Abstract: A method for video quality assessment implemented by a computing device, where the method includes selecting a first set of frames from an input video source, determining features of the first set of frames for input into a first machine learning model, applying the first machine learning model to the features of the first set of frames to obtain a video quality score and a confidence score, and triggering a retraining of the first machine learning model in response to the confidence score being determined to be out of bounds.

Abstract: A method and system corrects a content delivery infrastructure. The method of the system includes receiving a request to resolve reported issues for the content delivery infrastructure, collecting content delivery metrics for the content delivery infrastructure, executes a language transformer model on the request and the content delivery metrics to generate a set of possible resolutions with confidence ratings, and implementing an automated solution based on a resolution from the set of possible resolutions, in response to the resolution having a confidence rating above a threshold.

Abstract: A method of measuring video stream visual stability, the method including receiving a first set of network packets carrying data of the video stream, determining network performance metrics for a session associated with the first set of network packets, retrieving priority fault errors from a packet header of at least one network packet of the first set of the network packets, adding the priority fault errors and the network performance metrics to time series data, and applying a machine learning model to the time series data to obtain a visual stability score for the first set of network packets.

Abstract: A method of measuring video stream visual stability, the method including receiving a first set of network packets carrying data of the video stream, determining network performance metrics for a session associated with the first set of network packets, retrieving priority fault errors from a packet header of at least one network packet of the first set of the network packets, adding the priority fault errors and the network performance metrics to time series data, and applying a machine learning model to the time series data to obtain a visual stability score for the first set of network packets.

Abstract: A method for generating context information for a video stream includes selecting a set of frames of a video stream, applying a first machine learning model to the set of frames to extract action information from the set of frames, applying natural language learning to the set of frames to identify dialogue associated with the set of frames, and generating context information to categorize the dialogue and action information for the set of frames.

Abstract: A method for video quality assessment implemented by a computing device, where the method includes selecting a first set of frames from an input video source, determining features of the first set of frames for input into a first machine learning model, applying the first machine learning model to the features of the first set of frames to obtain a video quality score and a confidence score, and triggering a retraining of the first machine learning model in response to the confidence score being determined to be out of bounds.

Abstract: A method is executed by an electronic device in a video streaming system including a macroblock detector. The method includes receiving a video stream from a video source, selecting a frame from the video stream, detecting edges in the frame of the video stream, generating straight lines from detected edges in the frame of the video stream, identifying squares in the straight lines in the frame with polygon approximation, and outputting an indication of a detected macroblock in the image in response to identifying at least one square with characteristics of a macroblock in the frame.

Abstract: A method and system for managing transcoding of data in a stream that includes identifying an input source change for the stream with a new input source type, and adding a decoder for the new input source type, the decoder configured to output for a respective encoder in a transcoder pipeline.

Abstract: A method is executed by an electronic device in a video streaming system including a macroblock detector. The method includes receiving a video stream from a video source, selecting a frame from the video stream, detecting edges in the frame of the video stream, generating straight lines from detected edges in the frame of the video stream, identifying squares in the straight lines in the frame with polygon approximation, and outputting an indication of a detected macroblock in the image in response to identifying at least one square with characteristics of a macroblock in the frame.

Abstract: A method for detect and mitigate anomaly in video streaming platforms is disclosed. In one embodiment, performance data from a set of workers is received at a central telemetry system (CTS), where the performance data is indicative of operational status of the set of workers. The CTS processes the performance data, including generating task-specific monitoring data based on the performance data, and it identifies whether the performance data or the task-specific monitoring data contains any anomaly. Upon an anomaly being identified, the CTS mitigates the anomaly by interacting with the set of workers.

Abstract: A method implemented in an electronic device serving as a stream distributor distributing unicast video and audio streams of a streaming platform system is disclosed. The stream distributor receives a unicast data flow containing data of a video and audio stream from a load balancer, the unicast data flow being transmitted in user datagram protocol (UDP) packets. The method includes the electronic device detecting initiation of a maintenance activity on a virtual machine or its host associated with the stream distributor. The method includes the electronic device indicating a failure mode of the stream distributor, where the failure mode indication is to cause the load balancer to switch the unicast data flow to an alternate stream distributor. The method also includes the electronic device detecting a completion of the load balancer switching the unicast data flow to the alternate stream distributor, and halting the distribution of the unicast data flow.

Abstract: A method implemented in an electronic device serving as a worker processing video and audio streams of a streaming platform system is disclosed. The method includes the electronic device requesting a task group from an orchestrator of the streaming platform system, where the orchestrator manages video and audio stream processing of the streaming platform system. The method also includes the electronic device receiving the task group allocated for the worker from the orchestrator, where the task group is allocated based at least partially on characteristics of the worker. The method also includes the electronic device initiating one or more task running processes to execute the allocated task group, and where each task running process performs at least one of processing one or more inputs corresponding to an incoming edge to the node, and producing one or more outputs corresponding to one or more outgoing edges from the node.

Abstract: A method to coordinate video stream timing is disclosed. In one embodiment, a request is received at a streaming platform coordinator (SPC) to create a plurality of synchronized media workflows for video sources. Using the plurality of video sources as input, the SPC transmits a request from the SPC to at least one video streaming platform to create the plurality of media workflows, which are assigned to a plurality of workers. The SPC causes the plurality of videos to be routed to the plurality of workers upon receiving the request to route them. The SPC then determines a set of timing offsets to synchronize the plurality of media workflows, and makes the set of timing offsets available to the plurality of workers so that the workers publish synchronized media outputs. A system to perform the method and a storage medium storing instructions to perform the method are disclosed too.