The primary objective of this work is to classify the videos, which comprises the extraction of video frames and generation of features vector and identification of action sequence in the video. A deep neural network has been proposed that combines Convolutional Neural Network and Long Short Term Memory models to achieve better spatial and temporal features. Video Classification is performed by extracting the features from the frames by considering the parameters viz., frame width and height, video sequence length etc. In the proposed model, the frames extracted from a video are fed into a two-layer LSTM. The outcome of the LSTM model is forwarded to the Convolution Layers with an additional Global Average Pooling (GAP) layer in place of the fully connected layer. The frames extracted from the video are sent as an input to the twolayered Convolutional 2D layer, which is followed by Batch normalization and Max Pooling. Fully connected layer is replaced by Global Average Pooling (GAP) layer followed by dense layer. Depending on the input data, more number of batch normalization and dense layers would be added in order to achieve more accuracy. UCF101 Dataset is considered to classify the videos. The results demonstrate that the LRCN methodology outperforms both the conventional CNN method and ConvLSTM in terms of prediction accuracy. The suggested approach also provides more accurate temporal and spatial stream identification. The results have shown that the probability of the action sequence is predicted to be around 82 percentage with an accuracy of 92.86 percentage.
Wireless Sensor Networks (WSNs) have gained an emerging importance in different application domains especially in event tracking and monitoring. The sensor nodes in WSNs are observed to have shorter lifetime due to the continuous sensing and processing operations that result in quicker energy depletion. Based on the research works carried out, it is evident that clustering in an optimal solution to reduce the energy consumption and to enhance network lifetime. It is proposed to introduce a novel clustering technique called as Energy Efficient Hierarchical Clustering (E2HC) to optimize the network lifetime while consuming minimal energy. E2HC adapts three-tier architecture comprising of Hybrid Clustering, Rotational Cluster Head Selection and Hierarchical Packet Routing. In Hybrid clustering two distinct cluster categories are proposed viz., nearby cluster and distant cluster. The selection of nearby cluster is based on three parameters: distance between sensor nodes and base station, maximum number of sensor nodes that can be included in the nearby cluster and optimal distance (distance between the base station and the distant sensor node of nearby cluster).Fuzzy C-Means Clustering algorithm is used to cluster the remaining sensor nodes into multiple distant clusters. A residual energy assisted fitness function is adopted to select the cluster head in rotational fashion. To route the information collected by the sensor nodes, hierarchical packet routing uses distance between sender and base station. Extensive analysis of simulated WSNs over the proposed E2HC shows significant improvements in reduced energy consumption and extension of the network lifetime, as compared with the existing methods.
The modern online social networks consist of many interconnected services, such as business, private, public, entertainment and social media, which are collectively represented as Social Networking Services. The size of the social network grows exponentially due to increase in social media services and their real use in the above stated applications. The main aim of this paper is to develop a cloud-based simulation model that provides video sharing, content storage and appropriate decision-making services on demand over the Internet with an objective of reducing the redundant traffic flow amidst the applications and hence to utilize the network bandwidth effectively in the online social networks. The computational engine, which is included in the proposed cloud model will inherently analyse the traffic flow due to video data transmission across social media applications. The service providers can deploy their applications without any limitation in a cloud environment and users can share information using the rich deployed applications from anywhere on demand basis or on the "push and pull" basis. The deploying services minimize the risk of redundant traffic, which not only reduce the consumption of network bandwidth in the perspective of service providers but also lower the cost of usage of mobile data in the perspective of end users. The cloud-based simulation model provides reliable services using virtualized compute and storage technologies. The performance measures such as network bandwidth usage, mobile data usage and time consumption are estimated before and after sharing of video contents among different social media applications, which are most popular in the online social networks.
The main aim of this paper is to develop a new approach for identifying independent groups among users communicating in social networks using social media applications at any instant. Grouping of users as independent clusters is of dynamic nature as communication between known and unknown users can happen randomly at any point of time. It is becoming inherent to identify the groups, where the members of the group have strong relationship who communicate frequently and consistently via social media applications. Louvain’s algorithm will identify the clusters in the community detection process but keeps the lightweight nodes in the original groups without making them into one group by considering the dependence relations. The concept of Bernstein conditions is enhanced and applied to identify the dependency among the users of social networks by formulating equivalence relations, which adhere to the properties of Reflexivity, Symmetricity and Transitivity. Then, the equivalence classes are identified which denote the individual groups of clusters where the users of one cluster are loosely coupled with the users of any other cluster but tightly coupled among the users of the same group. The strength of relationship among the users within the same and different clusters is identified with respect to the quantum of messages being propagated among the users using Louvain’s algorithm and the results of equivalence class approach are compared using the same set of communication sequences to show the relation dependency among the members in various clusters.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.