Content aware lean predictive model for medical video transmission in Iot networks

Abstract

The increasing popularity of IoT and Wireless Sensor Networks has enabled the great demands for video traffic. However, the wireless link capacity for interconnecting the devices fails due to the increasing the demand of the traffic. Consequently, it results in some major problems such as poor quality in video streaming, service interruption and even the performance degradation. As a result, this research provides a novel technique that combines network-centric deep learning models with the concept of visual saliency content clustering. The boosted LSTM technique is used to build the suggested deep learning model, which reduces computation and improves performance. The hybrid model also acts as a medium of platform for intelligent video streaming which adopts the adaptive distance power features to train its own deep learning network. The extensive evaluation is carried out to prove the excellence of the proposed model. The Model evaluation parameters such as prediction accuracy, sensitivity, selectivity PSNR, SSIM and QoS evaluation parameters such as delay and bandwidth used. These measurements are made using the real-time test beds, which are made up of a Raspberry Pi 3 Model B+ running the proposed framework and connected to a cloud network. The proposed predictive model has proven to be superior than other conventional models when compared to other models that are currently in use, such as multi-layer perceptron and deep LSTM networks. But this framework requires further improvisation to handle the heterogeneity multiple path. Hence the research continues its steps by proposing a novel deep reinforcement-based extreme learning machines (DRLELM) approach to examine the complexities between routes, pathways, sub-flows, and SMPTCP connections in different topologies. Using DRLELM, throughput of the network is estimated. newline