VIDEO ANALYTIC SYSTEM FOR CROWD CHARACTERIZATION

    公开(公告)号:WO2021195060A1

    公开(公告)日:2021-09-30

    申请号:PCT/US2021/023646

    申请日:2021-03-23

    Abstract: A computer-implemented method for characterizing a crowd that includes recording (1) a video stream of individuals at a location having at least one reference point for viewing; and extracting (2) the individuals from frames of the video streams. The method may further include assigning (3) tracking identification values to the individuals that have been extracted from the video streams; and measuring (4) at least one type classification from the individuals having the tracking identification values. The method may further include generating (5) a crowd designation further characterizing the individuals having the tracking identification values in the location, the crowd designation comprising at least one measurement of probability that the individuals having the tracking identification values in the location view the at least one reference point for viewing.

    USECASE SPECIFICATION AND RUNTIME EXECUTION
    2.
    发明申请

    公开(公告)号:WO2020185512A1

    公开(公告)日:2020-09-17

    申请号:PCT/US2020/021182

    申请日:2020-03-05

    Abstract: A computer-implemented method includes obtaining (810) a usecase specification and a usecase runtime specification corresponding to the usecase. The usecase includes a plurality of applications each being associated with a micro-service providing a corresponding functionality within the usecase for performing a task. The method further includes determining (820) that at least one instance of the at least one of the plurality of applications can be reused during execution of the usecase based on the usecase specification and the usecase runtime specification, and reusing (830) the at least one instance during execution of the usecase.

    DYNAMIC, CONTEXTUALIZED AI MODELS
    3.
    发明申请

    公开(公告)号:WO2022159444A1

    公开(公告)日:2022-07-28

    申请号:PCT/US2022/012910

    申请日:2022-01-19

    Abstract: A method for employing a semi-supervised learning approach to improve accuracy of a small model on an edge device is presented. The method includes collecting (1001) a plurality of frames from a plurality of video streams generated from a plurality of cameras, each camera associated with a respective small model, each small model deployed in the edge device, sampling (1003) the plurality of frames to define sampled frames, performing (1005) inference to the sampled frames by using a big model, the big model shared by all of the plurality of cameras and deployed in a cloud or cloud edge, using (1007) the big model to generate labels for each of the sampled frames to generate training data, and training (1009) each of the small models with the training data to generate updated small models on the edge device.

    SPECIFICATION AND EXECUTION OF REAL-TIME STREAMING APPLICATIONS

    公开(公告)号:WO2020185754A1

    公开(公告)日:2020-09-17

    申请号:PCT/US2020/021849

    申请日:2020-03-10

    Abstract: Systems and methods to specify and execute real-time streaming applications are provided. The method includes specifying (170) an application topology for an application including spouts, bolts, connections, a global hash table, and a topology manager. Each spout receives input data and each bolt transforms the input data, the global hash table allows in memory communication between each spout and bolt to others of the spouts and the bolts. The topology manager (805) manages the application topology. The method includes compiling (820) the application into a shared or static library for applications, and exporting a special symbol associated with the application. The runtime system (840) can be used to retrieve the application topology from the shared or static library based on the special symbol (825) and execute the application topology on a single node or distribute across multiple nodes.

    FREE FLOW FEVER SCREENING
    5.
    发明申请

    公开(公告)号:WO2021242621A1

    公开(公告)日:2021-12-02

    申请号:PCT/US2021/033550

    申请日:2021-05-21

    Abstract: A method for free flow fever screening is presented. The method includes capturing (801) a plurality of frames from thermal data streams and visual data streams related to a same scene to define thermal data frames and visual data frames, detecting and tracking (803) a plurality of individuals moving in a free-flow setting within the visual data frames, and generating (805) a tracking identification for each individual of the plurality of individuals present in a field-of-view of the one or more cameras across several frames of the plurality of frames. The method further includes fusing (807) the thermal data frames and the visual data frames, measuring (809), by a fever-screener, a temperature of each individual of the plurality of individuals within and across the plurality of frames derived from the thermal data streams and the visual data streams, and generating (811) a notification when a temperature of an individual exceeds a predetermined threshold temperature.

    FACE CLUSTERING IN VIDEO STREAMS
    6.
    发明申请

    公开(公告)号:WO2021211226A1

    公开(公告)日:2021-10-21

    申请号:PCT/US2021/021475

    申请日:2021-03-09

    Abstract: Methods and systems for video analysis and response include detecting (304) face images within video streams. Noisy images are filtered (306) from the detected face images. Batches of the remaining detected face images are clustered (602) to generate mini-clusters, constrained by temporal locality. The mini-clusters are globally clustered (606) to generate merged clusters formed of face images for respective people, using camera-chain information to constrain a set of the video streams being considered. Analytics (204) are performed on the merged clusters to identify a tracked individual's movements through an environment. A response (206) is performed to the tracked individual's movements.

    VIDEO ANALYTICS SYSTEM FOR DWELL-TIME DETERMINATIONS

    公开(公告)号:WO2021168268A1

    公开(公告)日:2021-08-26

    申请号:PCT/US2021/018803

    申请日:2021-02-19

    Abstract: Systems and methods for determining dwell time is provided. The method includes receiving (310) images of an area (201) including one or more people (901) from one or more cameras (130), and detecting (341) a presence of each of the one or more people (901) in the received images using a worker (120). The method further includes receiving (470) by the worker (120) digital facial features stored in a watch list (112) from a master controller (110), and performing facial recognition (430) and monitoring the dwell time (420) of each of the one or more people (901). The method further includes determining if each of the one or more people is in the watch list (112) or has exceeded a dwell time threshold.

    REINFORCEMENT-LEARNING BASED SYSTEM FOR CAMERA PARAMETER TUNING TO IMPROVE ANALYTICS

    公开(公告)号:WO2022256246A1

    公开(公告)日:2022-12-08

    申请号:PCT/US2022/031279

    申请日:2022-05-27

    Abstract: A method for automatically adjusting camera parameters to improve video analytics accuracy during continuously changing environmental conditions is presented. The method includes capturing (1001) a video stream from a plurality of cameras, performing (1003) video analytics tasks on the video stream, the video analytics tasks defined as analytics units (AUs), applying (1005) image processing to the video stream to obtain processed frames, filtering (1007) the processed frames through a filter to discard low-quality frames and dynamically fine-tuning (1009) parameters of the plurality of cameras. The fine-tuning includes passing the filtered frames to an AU-specific proxy quality evaluator, employing State-Action-Reward- State-Action (SARSA) reinforcement learning (RL) computations to automatically fine-tune the parameters of the plurality of cameras, and based on the reinforcement computations, applying a new policy for an agent to take actions and learn to maximize a reward.

    APPLICATION-CENTRIC DESIGN FOR 5G AND EDGE COMPUTING APPLICATIONS

    公开(公告)号:WO2022240590A1

    公开(公告)日:2022-11-17

    申请号:PCT/US2022/026661

    申请日:2022-04-28

    Abstract: A method for specifying and executing an application including multiple microservices on 5G slices within a multi-tiered 5G infrastructure is presented. The method includes managing compute requirements and network requirements of the application simultaneously by determining end-to-end application characteristics by employing (1001) an application slice specification including an application ID component, an application name component, an application metadata component, a function dependencies component, a function instances component, and an instance connections component, specifying (1003) a function slice specification including a function network slice specification and a function compute slice specification, and employing (1005) a runtime component including a resource manager, an application slice controller, and an application slice monitor, wherein the resource manager maintains a database and manages starting, stopping, updating, and deleting application instances.

    MULTI-FACTOR AUTHENTICATION FOR PHYSICAL ACCESS CONTROL

    公开(公告)号:WO2020185517A1

    公开(公告)日:2020-09-17

    申请号:PCT/US2020/021201

    申请日:2020-03-05

    Abstract: Methods and systems for authentication include determining (606), at a first worker system, that a master system that stores a current authentication-list cannot be reached by a first network. Authentication is performed (610) on an authentication request using a previously stored copy of the authentication-list at the first worker system. The authentication includes facial recognition that is performed on detected face images for a first time window, before receiving the authentication request, and for a second time window, after receiving the authentication request. Authentication removes matching detected face images after completing an authentication request to prevent other individuals from using a same identifier. Access is granted (518) to a secured area responsive to the authentication.

Patent Agency Ranking