Google Inc.

Krishnan Eswaran

Vijnan Shastri

Krishna Kumar Gadepalli

Thierry Foucu

Jie Sun

Jianpeng Dong

Jiening Zhan

Wendy Tobagus

Dmitry Broyde

Shravya R Shetty

Shahar Jamshy

Daniel Shing Shun Tse

Zvika Ben-Haim

Charles Christian Krasic

Yang Xia

A computer-implemented system is described for identifying and retrieving similar radiology images to a query image. The system includes one or more fetchers receiving the query image and retrieving a set of candidate similar radiology images from a data store. One or more scorers receive the query image and the set of candidate similar radiology images and generate a similarity score between the query image and each candidate image. A pooler receives the similarity scores from the one or more scorers, ranks the candidate images, and returns a list of the candidate images reflecting the ranking. The scorers implement a modelling technique to generate the similarity score capturing a plurality of similarity attributes of the query image and the set of candidate similar radiology images and annotations associated therewith. For example, the similarity attributes could be patient, diagnostic and/or visual similarity, and the modelling techniques could be triplet loss, classification loss, regression loss and object detection loss.

Similar image search for radiology

A method includes separating, using a separation component, a plurality of input streams into a plurality of audio streams having different bitrates and a plurality of video streams having different bitrates; transcoding, using an audio transcoder, the plurality of audio streams into a plurality of transcoded audio streams having different transcode qualities; transcoding, using a video transcoder, the plurality of video streams into a plurality of transcoded video streams having different transcode qualities; segmenting, using a segmentation component, the plurality of audio streams into a plurality of equivalent audio segments based on identified segment boundaries associated with the plurality transcoded audio streams and the plurality of video streams into a plurality of equivalent video segments based on identified segment boundaries associated with the plurality of transcoded video streams; and joining, using a conjoining component, the plurality of equivalent audio segments and the plurality of video segments into a single stream.

Scalable, live transcoding with support for adaptive streaming and failover

A system and method for a live streaming platform that can redundantly process input streams in parallel ingestion pipelines is disclosed herein. Ingested input streams in the parallel pipelines can be segmented using a segmentation function that creates identical segments in each of the streams in the pipelines. If errors occur, or there are disruptions in one or more of the input streams or pipelines, the live streaming platform can ensure that segments are identifiable and different parts of the pipelines can join a main stream at different points in time without disruption.

A method including receiving an input media stream and splitting a video portion of a received portion of the input media stream into input video chunks. For each input video chunk of the video portion of the input media stream, the method includes assigning the input video chunk to one or more slave transcoders, splitting the input video chunk into video subchunks, and separately feeding each input video subchunk to each of the one or more slave transcoders. The method includes separately receiving output video subchunks from the one or more slave transcoders. Each output video subchunk respectively corresponds to one of the input video subchunks. The method also includes assembling the output video subchunks into an output video chunk corresponding to the input video chunk and assembling the output video chunks into an output media stream.

Transcoding media streams using subchunking

A method including receiving an input media stream and splitting a video portion of a received portion of the input media stream into input video chunks, while receiving a remaining portion of the input media stream. For each input video chunk of the video portion of the input media stream, the method includes assigning the input video chunk to one or more slave transcoders, splitting the input video chunk into video subchunks, and separately feeding each input video subchunk to each of the one or more slave transcoders. The method includes separately receiving output video subchunks from the one or more slave transcoders. Each output video subchunk respectively corresponds to one of the input video subchunks. The method also includes assembling the output video subchunks into an output video chunk corresponding to the input video chunk and assembling the output video chunks into an output media stream.

A method for enabling visual browsing of a plurality videos includes identifying a low frame rate version for each video from a plurality of videos using one or more computers, wherein the low frame rate version of each video from the plurality of videos is at a frame rate that provides a visually perceptible pause between successive visually distinct frames; and outputting information that defines a visual display, wherein the visual display presents a plurality of graphical representations each corresponding to a respective video from the plurality of videos, wherein each graphical representation from the plurality of graphical representations includes the low frame rate version of the respective video from the plurality of videos.

Visually browsing videos

Systems and methods for processing a live video stream are provided. An ingestion component receives segments of a live video stream in association with a request to process the live video stream. A control component generates an instruction graph for the live video stream based on predetermined criteria. The instruction graph defines a workflow for processing the live video stream, the workflow comprising a set of services for processing the segments. A transcoding component employs a subset of the services to process the segments based at least in part on the instruction graph and a distribution component then streams the segments to a device in response finishing process of the segments. With the live video processing system, segments of a video stream are processed and distributed as they are received. Thus segments of the video stream are processed and distributed prior to processing of the entire video stream.

Growing community of inventors

Oakland, CA, United States of America

Krishnan Eswaran