公开(公告)号：US10200428B1

公开(公告)日：2019-02-05

申请号：US15085965

申请日：2016-03-30

Applicant: Amazon Technologies, Inc.

Inventor： Jesse Jerome Rosenzweig ,  Michael Reddick Coleman ,  Dustin Encelewski ,  Stefan Christian Richter

IPC: H04L29/08 ,  H04L29/06 ,  H04L12/24 ,  H04L12/26

Abstract: Methods, systems, and computer-readable media for unicast routing of a media stream to subscribers are disclosed. A stream routing service is implemented using one or more computing devices that are provisioned from a multi-tenant provider network. The service acquires a media stream over one or more input channels from a media source. The service replicates the media stream to a plurality of subscriber components that have subscribed to the media stream. The media stream is replicated to the subscriber components using unicast transmission to individual addresses of the subscriber components.

公开(公告)号：US10225309B1

公开(公告)日：2019-03-05

申请号：US15077647

申请日：2016-03-22

Applicant: Amazon Technologies, Inc.

Inventor： Jesse Jerome Rosenzweig ,  Gregory Kenneth Truax

IPC: G06F15/16 ,  H04L29/06

Abstract: Methods, systems, and computer-readable media for monitoring of media pipeline health using tracing are disclosed. At a plurality of stages of a media delivery pipeline, tracing metadata is generated for elements of a media stream. The tracing metadata comprises a content identifier, a segment identifier, and a stage identifier. The tracing metadata is generated from the plurality of stages and sent to a metadata repository using instrumentation of components that implement the plurality of stages. A problematic stage within the plurality of stages is identified based at least in part on analysis of the tracing metadata.

公开(公告)号：US11120293B1

公开(公告)日：2021-09-14

申请号：US15823249

申请日：2017-11-27

Applicant: Amazon Technologies, Inc.

Inventor： Jesse Jerome Rosenzweig ,  Brian Lewis ,  Leah Siddall

IPC: G06K9/62 ,  H04N19/87 ,  G06T7/246 ,  H04N19/142 ,  G06T7/215 ,  G06F16/16 ,  G06F16/783 ,  G06F16/9535

Abstract: Various types of objects or occurrences can be automatically detected in input media being processed using a transcoder. The media content can be analyzed to determine various transitions, such as scene changes, which provide insight into useful locations for performing object recognition. Representative frames subsequent a transition are analyzed to determine whether they are appropriate for image analysis, using factors such as amount of motion, brightness, color, or pixel disparity within the frame. If a representative frame meets the various criteria, that frame is sent to an object recognition service for analysis. The output of the service can be a set of object tags that provide information identifying the object and its location in the media. The output tags can be encoded into the output video or stored to an associated metadata file, among other such options.