Plantronics, Inc.

International Business Machines Corporation

Nokia Technologies Oy

Polycom, Inc.

Poly, Inc.

Hai Jun Xu

Xi Lu

Wenxue He

Yongkang Fan

Hailin Song

Yue Chen

Hui Gao

Yanfeng Shi

Huai Nan Zhou

Yue Wang

Tianran Wang

Fangpo Xu

Wei Ming He

Lidan Qin

Yuan Yuan Ding

Wen Tao Zhang

Guanqin Zhang

Xiao Guang Song

Methods for creating a knowledge graph for a video are disclosed. Aspects include obtaining the video, processing the video to extract audio information and video information, and storing the extracted audio information and video information with a timestamp corresponding its occurrence in the video. Aspects also include creating a plurality of groups of the extracted audio information and video information based at least in part on the timestamps and extracting two or more keywords from each of the plurality of groups. Aspects further include identifying a relationship between the two or more keywords based on the extracted audio information and video information and creating a graph having a plurality of nodes and a plurality of links that connect a pair of nodes of the plurality of nodes. Each node corresponds to an extracted keyword and each link corresponds to the identified relationship between the pair of nodes.

Creating a knowledge graph for a video

A method () for operating cameras () in a cascaded network (), comprising: capturing a first view () with a first lens () having a first focal point () and a first centroid (), the first view () depicting a subject (); capturing a second view () with a second lens () having a second focal point () and a second centroid (); detecting a first location of the subject (), relative the first lens (), wherein detecting the first location of the subject (), relative the first lens (), is based on audio captured by a plurality of microphones (); estimating a second location of the subject (), relative the second lens (), based on the first location of the subject () relative the first lens (); selecting a portion () of the second view () as depicting the subject () based on the estimate of the second location of the subject () relative the second lens ().

Optimal view selection in a teleconferencing system with cascaded cameras

A method of tracking a presenter during a videoconference, comprising: detecting, in a data stream, data indicating a presence of one or more persons at a videoconferencing endpoint, the data stream comprising a plurality of frames; detecting tracker initiation data (e.g., that a person is actively talking, or that a person has raised their hand, or both) in the data stream; determining a subject of interest responsive to the tracker initiation data, and tracking the subject of interest as they move while making a presentation.

Tracker activation and deactivation in a videoconferencing system

A videoconferencing endpoint is described that uses a cascading sequence of convolutional neural networks to perform face detection and upper body detection of participants in a videoconference at the endpoint, where at least one member of the sequence of neural networks performs upper body detection, and where the final member of the sequence of neural networks performs face detection based on the results of the upper body detection. The models of the neural networks are trained on both large datasets of faces well as images that have been distorted by a wide-angle camera of the videoconferencing endpoint.

Joint upper-body and face detection using multi-task cascaded convolutional networks

Routing containerized services requests on a serverless platform can include responding to receiving on the serverless platform a request for a containerized service and detecting a potential request failure. The response to the potential request failure can include searching a database that maps version-specific requests to a plurality of containers that are configured and managed by the serverless platform. The routing can include redirecting the request to a container containing a prior version of the containerized service requested in response to determining that the prior version of the containerized service maps to a version-specific request that matches the request received.

Serverless platform request routing

A videoconferencing device at an endpoint determines whether a person is a real person standing in front of a display device or if the person is instead an image rendered by a display device. In the first instance the real person will be included in a video feed for transmission to a remote endpoint. In the second instance, images of the display device on which the person is rendered will not be included in the video feed.

Differentiating a rendered conference participant from a genuine conference participant

A videoconferencing endpoint is described that uses a combination of face detection, motion detection, and upper body detection for selecting participants of a videoconference for group framing. Motion detection is used to remove fake faces as well as to detect motion in regions around detected faces during postprocessing. Upper body detection is used in conjunction with the motion detection in postprocessing to allow saving faces that have been initially detected by face detection for group framing even if the participant has turned away from the camera, allowing the endpoint to keep tracking the participants region better than would be possible based only on an unstable result coming from face detection.

Joint use of face, motion, and upper-body detection in group framing

A method for tracking movement of a presenter during a videoconference comprises tracking the location of subject of interest; quickly detecting error conditions&#x2014;such as can occur when a subject of interest moves too quickly or when a subject of interest walks past another person&#x2014;and quickly reinitiating the tracking sequence to help ensure a positive videoconferencing experience.

Presenter-tracker management in a videoconferencing environment

A computer system modifies a compiled application to monitor performance. A request routing function of the compiled application is replaced with jump code that directs processing to a data collection function and backing up the request routing function. In response to receiving, during execution of the compiled application, an incoming request from a remote device for handling by the request routing function, the data collection function is executed based on the jump code. Using the data collection function, data associated with the incoming request is collected, the data comprising performance metrics. The backup of the request routing function is called using the data collection function to process the request and provide a response. Embodiments of the present invention further include a method and program product for modifying a compiled application to monitor performance in substantially the same manner described above.

Modifying compiled applications to monitor performance

Detecting spoofing talker in a videoconference

A method may include receiving a source video frame captured using a wide angle lens, deriving a characteristic of a target video frame from the source video frame, generating a transitional video frame using the source video frame and the target video frame, and transmitting the transitional video frame to an endpoint.

Smoothing image transitions when changing views

An approach, is presented for exchanging status updates while on and/or offline when collaborating on various activities via one or more networks. A method comprises determining one or more tasks associated with at least one collaborative activity for one or more devices participating in the at least one collaborative activity. The method also comprises determining status information for the one or more tasks with respect to the one or more devices. Further, the method also comprises causing, at least in part, an exchange of the status information among the one or more devices via at least one communication mechanism that supports offline messaging.

Growing community of inventors

Beijing, China

Hai Jun Xu