Investigating response time and accuracy in online classifier learning for multimedia publish-subscribe systems

Asra Aslam,Edward Curry

doi:10.1007/s11042-020-10277-x

Asra Aslam, Edward Curry

Open Access

https://doi.org/10.1007/s11042-020-10277-x

Copy DOI

Abstract

The enormous growth of multimedia content in the field of the Internet of Things (IoT) leads to the challenge of processing multimedia streams in real-time. Event-based systems are constructed to process event streams. They cannot natively consume multimedia event types produced by the Internet of Multimedia Things (IoMT) generated data to answer multimedia-based user subscriptions. Machine learning-based techniques have enabled rapid progress in solving real-world problems and need to be optimised for the low response time of the multimedia event processing paradigm. In this paper, we describe a classifier construction approach for the training of online classifiers, that can handle dynamic subscriptions with low response time and provide reasonable accuracy for the multimedia event processing. We find that the current object detection methods can be configured dynamically for the construction of classifiers in real-time, by tuning hyperparameters even when training from scratch. Our experiments demonstrate that deep neural network-based object detection models, with hyperparameter tuning, can improve the performance within less training time for the answering of previously unknown user subscriptions. The results from this study show that the proposed online classifier training based model can achieve accuracy of 79.00% with 15-min of training and 84.28% with 1-hour training from scratch on a single GPU for the processing of multimedia events.

Highlights

The rising interest in multimedia devices with an increase in the number of users is responsible for the evolution of multimedia content in smart environments
Confusion matrix has been shown by taking Single Shot MultiBox Detectors (SSD) as an object detection model presently in Table 9, where it contains information about expected and predicted classes detected by the proposed system
We can observe that the values of true positives and true negatives are considerably higher than the values of false positives and false negatives for most of the subscriptions

Summary

Introduction

The rising interest in multimedia devices with an increase in the number of users is responsible for the evolution of multimedia content in smart environments This imposes the Multimedia Tools and Applications (2021) 80:13021–13057 challenge of processing multimedia events in real-time irrespective of multiple application domains. Event processing systems are introduced to process data streams for the detection of events within publish/subscribe paradigm, where publish/subscribe is a message-oriented interaction paradigm in which publishers send messages. The consumers express their interest for receiving some useful information [29]. Existing multimedia applications cannot handle dynamic subscriptions belonging to multiple domains, and we need to move towards generalised multimedia event processing to achieve high accuracy in low response-time

Methods

Results

Conclusion