Motion detection using machine learning models

Systems and methods described herein provide techniques for detecting behavior of a subject by processing video data using one or more trained models configured to detect behavior of the subject. The described system processes a set of frames from the video data using different trained models. The s...

Full description

Saved in:
Bibliographic Details
Main Authors KUMAR VIVEK, GEUTHER BRIAN Q
Format Patent
LanguageChinese
English
Published 06.06.2023
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:Systems and methods described herein provide techniques for detecting behavior of a subject by processing video data using one or more trained models configured to detect behavior of the subject. The described system processes a set of frames from the video data using different trained models. The system further processes different orientations of the set of frames. Various outputs from different trained models and from processing the different orientations of the set of frames may be combined to then make a final determination as to whether a subject exhibits a particular behavior within a particular frame. 本文中所描述的系统和方法提供通过使用被配置成检测受试者行为的一个或多个经训练模型处理视频数据来检测受试者行为的技术。所描述的系统使用不同的经训练模型处理来自所述视频数据的帧集合。所述系统进一步处理所述帧集合的不同定向。来自不同的经训练模型和来自处理所述帧集合的所述不同定向的各种输出可进行组合,以接着作出关于受试者在特定帧内是否表现出特定行为的最终确定。
Bibliography:Application Number: CN202180072763