Amazon Technologies, Inc.

Logshero Ltd.

Shine Security Ltd.

Ianir Ideses

Noam Sorek

Dotan Kaufman

Ido Yerushalmy

Eli Alshan

Guy Adam

Eduard Oks

Ziv Segal

Mark Kliger

Itay Katz

Liza Potikha

Eran Borenstein

Sharon Alpert

Asaf Yigal

Imry Kissos

Lior Fritz

Omer Meir

Tomer Levy

Matan Goldman

Ron Porat

Alon Blayer-Gat

Oren Farage

Ridge Carpenter

Avi Avraham Ben-Cohen

Michael Chertok

Noah Lirone Sarfati

Matan Karklinsky

Gabriel Pragier

Roi Rav-Hon

Georgy Melamed

Lamarr Smith

Doron Shamia Sadeh

Elizabeth Reisman

Claire Mcgowan

Systems and techniques are described for localizing a camera. An example method includes generating first image data comprising a first image plane representing the planar scene from a first point-of-view. The example includes generating second image data comprising the second image plane representing the planar scene from a second point-of-view. The example includes identifying a plurality of keypoints of the planar scene that is visible in the first image plane and the second image plane. The example includes matching the plurality of keypoints of a planar scene to a first pixel of a first image plane and a second pixel of a second image plane. The example includes generating a homography relating the planar scene, the first image plane, and the second image plane. The example includes generating, based on the homography, at least one of a calibration parameter or a localization parameter for an unlocalized camera.

Systems and methods for camera localization with planar surfaces

Systems, devices, and methods are provided for determining factually consistent generative narrations. A narrative may be generated by performing steps to determine one or more metadata messages for a first portion of a video stream, determine transcribed commentary for a second portion of the video stream, wherein the second portion includes the first portion, and determine a prompt based at least in part on the one or more metadata messages and the transcribed commentary. The prompt may be provided to a generative model that produces an output text. Techniques for performing a factual consistency evaluation may be used to determine a consistency score for the output text that indicates whether the output text is factually consistent with the one or more metadata messages and the transcribed commentary. A narrated highlight video may be generated using the consistent narrative.

Factually consistent generative narrations

Characteristics of a user's movement are evaluated based on performance of activities by a user within a field of view of a camera. Video data representing performance of a series of movements by the user is acquired by the camera. Pose data is determined based on the video data, the pose data representing positions of the user's body while performing the movements. The pose data is compared to a set of existing videos that correspond to known errors to identify errors performed by the user. The errors may be used to generate scores for various characteristics of the user's movement. Based on the errors, exercises or other activities to improve the movement of the user may be determined and included in an output presented to the user.

System for correcting user movement

Video output is synchronized to the actions of a user by determining positions of the user's body based on acquired video of the user. The positions of the user's body are compared to the positions of a body shown in the video output to determine corresponding positions in the video output. The video output may then be synchronized so that the subsequent output that is shown corresponds to the subsequent position attempted by the user. The rate of movement of the user may be used to determine output characteristics for the video to cause the body shown in the video output to appear to move at a similar rate to that of the user. If the user moves at a rate less than a threshold or performs an activity erroneously, the video output may be slowed or portions of the video output may be repeated.

System for synchronizing video output based on user activity

First video data representing performance of an activity by a first user is acquired. Poses of the first user are determined from the first video data. Second video data is generated based on the determined poses and based on appearance data that represents a second user, such as a model, paid performer, and so forth, in various poses. The resulting second video data depicts the second user performing the same poses as the first user. The second video data may then be sent to a recipient. For example, a participant in an exercise class may send a video to an instructor that depicts what appears to be the paid performer performing the poses, instead of the participant. As a result, video data showing the participant is not shared, protecting the privacy of the participant while still allowing them to participate and interact.

System for generating privacy-preserving video data

Video output is generated based on first video data that depicts the user performing an activity. Poses of the user during performance of the activity are compared with second video data that depicts an instructor performing the activity. Corresponding poses of the user's body and the instructor's body may be determined through comparison of the first and second video data. The video data is used to determine the rate of motion of the user and to generate video output in which a visual representation of the instructor moves at a rate similar to the that of the user. For example, video output generated based on an instructional fitness video may be synchronized so that movement of the presented instructor matches the rate of movement of the user performing an exercise, improving user comprehension and performance.

System for synchronizing video output based on detected behavior

Systems for assisting a user in performance of a meditation activity or another type of activity are described. The systems receive user input and sensor data indicating physiological values associated with the user. These values are used to determine a recommended type of activity and a length of time for the activity. While the user performs the activity, sensors are used to measure physiological values, and an output that is provided to the user is selected based on the measured physiological values. The output may be selected to assist the user in reaching target physiological values, such as a slower respiration rate. After completion of the activity, additional physiological values are used to determine the effectiveness of the activity and the output that was provided. The effectiveness of the activity and the output may be used to determine future recommendations and future output.

Interface for guided meditation based on user interactions

A first user generates video data for performance of an activity, such as a fitness exercise, by performing the activity in front of a camera. Based on the video data, the amount of movement of different parts of the first user's body is determined. Data representing the position of the first user over time is generated. The data may take the form of a function or a signal that is based on the function. The locations of body parts that move significantly are prioritized over other body parts when determining this data. At a subsequent time, a second user performs the activity. The number of times the second user completes the activity is counted by determining the number of times the second user reaches a position corresponding to a maximum value in the data representing the position of the first user.

System for learning and tracking performance of activities

A system and methods for classification of events, the system including a memory, a network interface, and a processor configured to monitor system events via the network interface and store the system events as an event log in the memory, collect information about events from various web resources, extract, from the event log and the collected event information, labeling data including data about urgency of events and generate, based on the extracted labeling data, labeling data structures ascribing labels to respective events, extract, from the event log and the collected event information, event features data and generate, based on the extracted event features data, event features data structures ascribing features to respective events, and formulate decision rule structures for identification of urgent events based on the labeling data structures and the event features data structures.

Methods and systems for alerting based on event classification and for automatic event classification

A system for managing a plurality of shared computer resources, comprising at least one hardware processor adapted for: in at least one of a plurality of management iterations: computing statistical values each associated with one of a plurality of streams, where the streams are input to a plurality of computing processes which are used for processing the plurality of streams, executed using at least one of a plurality of shared computer resources; identifying according to an analysis of the statistical values at least one obstructing stream; allocating at least one new shared computer resource; creating a plurality of new computing processes which execute using the at least one new shared computer resource; and instructing inputting the at least one obstructing stream to the plurality of new computing processes instead of to the plurality of computing processes while the plurality of streams are inputted to the plurality of computing processes.

System and method for managing shared computer resources

Method and system for automatic event classification

A computerized method of preemptive event handling, The method comprises monitoring, in run time at kernel level, a plurality of events of a plurality of processes executed by an operating system (OS) running on a computing device, detecting, in run time, a first event of the plurality of events, the first event being performed by a first process of the plurality of processes on the computing device, classifying, in run time, the first process as a malware in response to the detection of the first event, and preventing, in run time, the first process from running on the computing device before the first event is processed by the OS.

Preemptive event handling

A computerized method of reverting system data affected by a malware. The method comprises monitoring, in run time, a plurality of events of a plurality of processes executed by an operating system (OS) running on a computing device, logging in an event log, in run time, the plurality of events, classifying, in run time, a first process of the plurality of processes as a malware, identifying a set of events of the first process from the plurality of events using the event log, and reverting, in response to the classification, at least one system object hosted in the computing device to remove an effect of the set of events on the OS.

Reversion of system objects affected by a malware

A method of queuing network traffic events on a client terminal. The method comprises monitoring, in run time, a plurality of network traffic events triggered by a plurality of applications executed on a client terminal, extracting a plurality of network traffic event characteristics of each of the plurality of network traffic events, classifying each one of the plurality of network traffic events according to a respective the plurality of network traffic event characteristics, clustering the plurality of network traffic events in a plurality of clusters according to the classifying, and managing an opening a plurality data connections between the client terminal and a network such that the content of each cluster of the plurality of clusters is transmitted in another of the plurality data connections.

Growing community of inventors

Raanana, Israel

Ianir Ideses