公开(公告)号：US10755106B1

公开(公告)日：2020-08-25

申请号：US15975300

申请日：2018-05-09

Applicant: Amazon Technologies, Inc.

Inventor： Amr Bakry ,  Mashhour Solh

IPC: G06K9/00 ,  H04L29/08

Abstract: Systems and techniques are generally described for pattern recognition in video data. In some examples, a first set of pixels representing a first person in the first video data may be determined. First feature representation data of the first set of pixels may be generated. The first feature representation data may be determined to correspond to second feature representation data stored in memory. An identity of the first person associated with the second feature representation data may be determined. A first action associated with the first person may be determined. A notification may be sent to a computing device accessible by the first person. The notification may include a reminder for the first person to engage in the first action.

公开(公告)号：US10909826B1

公开(公告)日：2021-02-02

申请号：US15968337

申请日：2018-05-01

Applicant: Amazon Technologies, Inc.

Inventor： Mashhour Solh ,  Amr Bakry

IPC: H04N5/225 ,  H04N7/18 ,  G08B13/196 ,  H04N5/232

Abstract: Techniques are generally described for suppressing video streaming based on trajectory information. First video data captured at a first time may be received from a first camera device. A determination may be made that the first video data includes image data representing a previously-identified human. A determination may be made that first trajectory data associates movement of the previously-identified human with the first camera device at the first time. A signal may be sent to a second camera device. The signal may be effective to suppress streaming of video captured by the second camera device during a second time following the first time.

公开(公告)号：US11257226B1

公开(公告)日：2022-02-22

申请号：US16515014

申请日：2019-07-17

Applicant: Amazon Technologies, Inc.

Inventor： Mashhour Solh ,  Amr Bakry

IPC: G06T7/254 ,  G06T3/40 ,  G06K9/62 ,  G06F17/18 ,  H04N7/18

Abstract: A method of classifying motion from video frames involves generating motion frames indicative of changes in pixel values between pairs of frames. The method also involves determining one-dimensional feature values based on the video frames or motion frames, such as the statistical values or linear transformation coefficients. Each one-dimensional feature value may be stored in a buffer, from which additional temporal feature values can be extracted indicative of the change of the one-dimensional feature values across a set of frames. A classifier may receive the one-dimensional feature values and the additional temporal feature values as inputs, and determine the class of motion present in the video frames. Some classes of motion, such as irrelevant motion, may be considered irrelevant to the execution of certain motion-triggered actions, such that the method may involve suppressing the performance of a motion-triggered action based on the determined class of motion.