International Journal of Communication Networks and Information Security (IJCNIS)
Not a member yet
1021 research outputs found
Sort by
Enhancing the Performance of Single-Channel Blind Source Separation by Using ConvTransFormer
Get PDFIn the specialized field of audio signal processing, this study introduces a pioneering ConvTransFormer architecture aimed at enhancing the performance of single-channel blind source separation (SCBSS). This innovative architecture ingeniously combines the strengths of a multiple simple-weak attention mechanism with the triple-gating feature of a Gated Attention Unit (GAU) within the ConvTransFormer. This combination allows for a more focused and effective targeting of specific segments within the input sequence. The efficacy of this ConvTransFormer architecture is rigorously evaluated using the WSJ0-2mix dataset, a standard benchmark in the field. The results of this evaluation are significant, demonstrating substantial improvements in key performance metrics. Notably, there is an increase in the Signal-to-Interference (SI)-Signal-to-Noise Ratio improvement (SNRi) by 16.5 and in the Signal-to-Distortion Ratio improvement (SDRi)-Signal-to-Interference (SDRi) by 16.8. These improvements are crucial indicators of the quality of source separation in SCBSS. The findings of this research are groundbreaking, indicating that the proposed ConvTransFormer architecture surpasses existing methods in both SI-SNRi and SDRi performance metrics. This advancement marks a significant step forward in the field of SCBSS, offering new avenues for more effective and precise audio signal processing, especially in scenarios where isolating individual sound sources from a single- channel input is essential
Exploring Radio Cognitive Network Applications in the Digital Design of Jinzuo Furniture Cultural Tourism Display Platform Anchored in Intangible Cultural Heritage
Get PDFDigital design analysis plays an important role in the construction of Jinzuo furniture cultural tourism display platform, especially in the fields of cultural relics, archaeology, and collection, and has also attracted widespread attention from academia in the antique market. However, the current understanding of Jinzuo furniture still stays in theoretical research and lacks a systematic display platform. In the process of analyzing the regional culture of Jinzuo furniture, there are problems such as poor display effect and a small amount of display data. The reason is mainly because of the difference in regional culture and the lack of support for wireless network technology, which affects the publicity and development of Jinzuo furniture culture. Therefore, this paper proposes a digital design method based on the combination of multimedia networks and cognitive radio technology to build a Jinzuo furniture culture and tourism display platform. Firstly, cognitive radio technology is used to collect Jinzuo furniture data, and data from different platforms are summarized through multimedia technology. According to the regional cultural characteristics, the genre of Jinzuo furniture is divided, and the transmission of digital design is studied and understood by radio technology To promote the identification of regional cultural characteristics of Jinzuo furniture and the display effect of the platform. The digital design display results show that with the support of cognitive radio technology, multimedia network technology can improve the design level of Jinzuo furniture culture and tourism display platform, and use cognitive radio technology to promote the development of Jinzuo furniture culture and tourism. It can meet the requirements of Shanxi cultural construction
Using Minimum Connected Dominating Set for Mobile sink path planning in Wireless Sensor Networks
Get PDFWireless sensor networks are a motivating area of research and have a variety of applications. Given that these networks are anticipated to function without supervision for extended periods, there is a need to propose techniques to enhance the performance of these networks without consuming the essential resource sensor nodes have, which is their battery energy. In this paper, we propose a new sink node mobility model based on calculating the minimum connected dominating set of a network. As a result, instead of visiting all of the static sensor nodes in the network, the mobile sink will visit a small number or fraction of static sensor nodes to gather data and report it to the base station. The proposed model's performance was examined through simulation using the NS-2 simulator with various network sizes and mobile sink speeds. Finally, the proposed model's performance was evaluated using a variety of performance metrics, including End-To-End delay, packet delivery ratio, throughput, and overall energy consumption as a percentage
Machine Learning Algorithms for High Performance Modelling in Health Monitoring System Based on 5G Networks
Get PDFThe development of Internet of Things (IoT) applications for creating behavioural and physiological monitoring methods, such as an IoT-based student healthcare monitoring system, has been accelerated by advances in sensor technology. Today, there are an increasing number of students living alone who are dispersed across large geographic areas, therefore it is important to monitor their health and function. This research propose novel technique in high performance modelling for health monitoring system by 5G network based machine learning analysis. Here the input is collected as EEG brain waves which are monitored and collected through 5G networks. This input EEG waves has been processed and obtained as fragments and noise removal is carried out. The processed EEG wave fragments has been extracted using K-adaptive reinforcement learning. this extracted features has been classified using naïve bayes gradient feed forward neural network. The performance analysis shows comparative analysis between proposed and existing technique in terms of accuracy, precision, recall, F-1 score, RMSE and MAP. Proposed technique attained accuracy of 95%, precision of 85%, recall of 79%, F-1 measure of 68%, RMSE of 52% and MAP of 66%
An Adaptive Fractal Image Steganography Using Mandelbrot and Linear Congruent Generator
Get PDFDespite the advancements that occurred in the field of technology, information security (i.e., IS) is still deemed important and critical topic. It is still especially deemed so during the transfer process. In this research, a new approach is proposed for hiding information through the use of iterated function systems (i.e., IFS) from Fractals. This approach employs the main feature of fractals that concentrate on the idea that hackers who seek to find the hidden data shall not be able of locating it. Therefore, there is a need to carry out a decoding process in the aim of revering the conversion for securing the transmitted information. In this research, the secure information is hidden inside a fractal Mandelbrot image using the Linear Congruent Generator (i.e., LCG). Regarding the proposed system, it generates the fractal image through the use of the predefined knowledge gained from the hider site that works as a host for different types of secret messages. The knowledge that comes from the key of image dimensions, parameters of Mandelbrot, LCG key, and key agreement of cryptography method, which makes Stego-image analyses of hidden data unacceptable without the correct knowledge. Based on the results that are obtained through carrying out experiments showed the proposed method meets all the requirements for steganography. Such requirements include: the ones related to capacity, visual appearance, undetectability, robustness against extraction (i.e., security), and hit the highest capacities with a visual appearance of high quality
Analysis of the Role of Compliance Plan in AI Criminal Risk Prevention-Take AI Criminal Risk in Network Communication as Example
Get PDFTo address the criminal risks associated with artificial intelligence (AI) in network communications, such as privacy invasion, information leakage and abuse, prejudice and discrimination, and intelligent crime, it's crucial to implement robust compliance plans. These plans should ensure legal compliance by adhering to relevant laws and regulations, thus safeguarding against unlawful use of AI. Strengthening data privacy and protection is vital to prevent unauthorized access and misuse of sensitive information. Ensuring fairness and eliminating discrimination are essential to maintaining ethical AI practices and preventing biases in AI decision-making processes. Improving system transparency and interpretability is also critical; it involves making AI systems more understandable and accountable for their actions and decisions. Additionally, reinforcing security measures is necessary to defend against cyber threats and vulnerabilities, thereby reducing the probability of AI- enabled criminal activities. These comprehensive strategies are pivotal in mitigating the criminal risks of AI in network communication and promoting the responsible and ethical development of AI technology
COMPARISON AND IMPLEMENTATION OF PROTOCOLS TO COMPENSATE FOR CONGESTION IN NETWORKS WITH REINFORCEMENT LEARNING
No full textAdaptive techniques with reinforcement learning are much neededfor TCP problem avoidance in network protocols. Recentdevelopments in congestion control need a comparative analysis tosolve the issues like packet delivery, time etc, a primary study withdifferent protocols execution and comparison of techniques in thepresent work. The present work is initiating CORPL , RTCP, QTCPimplementation in network simulation to avoid congestion usingthe NS3 simulator. Work mainly focused on delivery ratio,throughput, energy consumption, network lifetime, and delay.Based on their comparison of these three protocols, QTCP givesbetter results, especially with advancements in reinforcement-basedenhancement techniques
Optimization of Deep Convolutional Neural Network with the Integrated Batch Normalization and Global pooling
Get PDFDeep convolutional neural networks (DCNN) have made significant progress in a wide range of applications in recent years, which include image identification, audio recognition, and translation of machine information. These tasks assist machine intelligence in a variety of ways. However, because of the large number of parameters, float manipulations and conversion of machine terminal remains difficult. To handle this issue, optimization of convolution in the DCNN is initiated that adjusts the characteristics of the neural network, and the loss of information is minimized with enriched performance. Minimization of convolution function addresses the optimization issues. Initially, batch normalization is completed, and instead of lowering neighborhood values, a full feature map is minimized to a single value using the global pooling approach. Traditional convolution is split into depth and pointwise to decrease the model size and calculations. The optimized convolution-based DCNN's performance is evaluated with the assistance of accuracy and occurrence of error. The optimized DCNN is compared with the existing state-of-the-art techniques, and the optimized DCNN outperforms the existing technique
Research on the implantation and dissemination strategy of short creative advertising videos in the new media era
Get PDFInterconnections and communication between people in social media enhance the information dissemination. This is leveraged by the business firms by using the social media as advertising platform for their products, brands and services. As interactive multimedia is more captivating than other media forms, the short and creative advertising videos are recently used as effective marketing tool. This work proposes a framework that taps the features of the video advertisements along with social networking properties like network clusters and network degree to determine the transmission and retransmission of video between primary and secondary level clusters. The degradation in retransmission is estimated using the connections and their influence computed as connection coefficient between the nodes. The proposed framework is validated on Facebook advertisement videos by assessing its efficacy in dissemination and implementation of the products. The detailed experimentation reveals that the proposed model is much effective in determining the dissemination and implementation of the video content with the average mean square error of 0.24. The work is robust and versatile, that it could be deployed to other social media networks by customising the estimation of transmission and retransmission rate
Revolutionizing the Industrialization of Buildings with BIM and AI: Factors Influencing Adoption and Development and Building Planning
Get PDFThe continuous development of BIM and AI has made the analysis of building industrialization a hot spot in research and made the issue of industrial reform the focus of computer and construction research. The traditional architectural planning method cannot solve the problem of building industrialization reform under AI technology, and the accuracy of architectural planning is low. Therefore, this paper proposes a model based on BIM and AI technology to analyze the development mode and planning of building reform. First, BIM and AI technology are used to analyze the building data, and the irrelevant construction industrialization data is deleted according to the development mode and planning of the data characteristics. Then, according to the judgment results of construction data, compared with the traditional architectural planning method, the content of different construction industries is deeply excavated, and the influencing factors with higher possibilities are output. After simulation test and analysis, BIM and AI technology can improve the judgment accuracy, integrity, and rationality of the influencing factors of building industrialization planning, with an accuracy rate of 96.4%, and make the correct choice of development mode to meet the judgment needs of influencing factors of architectural planning