基于深度学习的人体异常行为检测综述

Overview of Human Abnormal Behavior Detection Based on Deep Learning

如今,从大量视频流中手动搜索异常事件是一项艰巨的任务,而基于深度学习的人体异常行为检测不仅省时省力,且决策能力相对可靠,从而保证了公共安全。首先回顾了异常行为检测任务的传统机器学习方法并引入深度学习方法;然后概述异常事件的定义与人体异常行为检测的过程;最后详细介绍了基于深度学习的人体异常行为检测方法。

Nowadays,it is a difficult task to manually search abnormal events from a large number of video streams.The human abnormal behavior detection based on deep learning not only saves time and effort,but also has relatively reliable decision-making ability,thus ensuring public safety.This paper firstly reviews the traditional machine learning methods for abnormal behavior detection tasks and introduces the deep learning method.Then,the definition of abnormal events and the process of human abnormal behavior detection are summarized.Finally,the human abnormal behavior detection method based on deep learning is introduced in detail.