Distributed Deep Learning Model for Intelligent Video Surveillance Systems with Edge Computing

Abstract

In this paper, we propose a Distributed Intelligent Video Surveillance (DIVS) system using Deep Learning (DL) algorithms and deploy it in an edge computing environment. We establish a multi-layer edge computing architecture and a distributed DL training model for the DIVS system. The DIVS system can migrate computing workloads from the network center to network edges to reduce huge network communication overhead and provide low-latency and accurate video analysis solutions. We implement the proposed DIVS system and address the problems of parallel training, model synchronization, and workload balancing. Task-level parallel and model-level parallel training methods are proposed to further accelerate the video analysis process. In addition, we propose a model parameter updating method to achieve model synchronization of the global DL model in a distributed EC environment. Moreover, a dynamic data migration approach is proposed to address the imbalance of workload and computational power of edge nodes. Experimental results showed that the EC architecture can provide elastic and scalable computing power, and the proposed DIVS system can efficiently handle video surveillance and analysis tasks.

Extending reference architecture of big data systems towards machine learning in edge computing environments

“…Data processing functionality was executed either in end user devices [29,30], in edge servers [40,41], or in the public cloud [45,46]. Data loading and pre-processing was performed in end user devices [29,34] or in on-site devices [45].…”

“…Data loading and pre-processing was performed in end user devices [29,34] or in on-site devices [45]. Data was extracted either from end-user devices [29], or from the edge environment [30,40]. Deployment of models required functionality in different environments.…”

“…The approach reduced gradient communication time (~ 70-80%), and speeded up training in distributed ML with little effect on accuracy. A distributed video surveillance system has been proposed, where vehicle classification/traffic flow prediction models were trained at edge nodes in parallel [40].…”

Automatically Detecting Excavator Anomalies Based on Machine Learning

“…Based on the real excavator data, we have carried out a comprehensive evaluation. The results show that the anomaly detection accuracy is as high as 99.88%, which is obviously superior to the previous methods based on expert systems and traditional statistical models.Symmetry 2019, 11, 957 2 of 18 condition data increases, it is hard to extract regular patterns from mass of data based on traditional statistic models.The recent advancement of neural network and machine learning (ML) have been successfully applied to various application scenarios [8][9][10][11][12][13][14]. However, none has used them in anomaly detection for excavators.…”

“…The recent advancement of neural network and machine learning (ML) have been successfully applied to various application scenarios [8][9][10][11][12][13][14]. However, none has used them in anomaly detection for excavators.…”