6 research outputs found
Hyper Parameter Optimization for Transfer Learning of ShuffleNetV2 with Edge Computing for Casting Defect Detection
No full text171-177A casting defect is an expendable abnormality and the most undesirable thing in the metal casting process. In Casting Defect Detection, deep learning based on Convolution Neural Network (CNN) models has been widely used, but most of these models require a lot of processing power. This work proposes a low-power ShuffleNet V2-based Transfer Learning model for defect identification with low latency, easy upgrading, increased efficiency, and an automatic visual inspection system with edge computing. Initially, various image transformation techniques were used for data augmentation on casting datasets to test the model flexibility in diverse casting. Subsequently, a pre-trained lightweight ShuffleNetV2 model is adapted, and hyperparameters are fine-tuned to optimize the model. The work results in a lightweight, adaptive, and scalable model ideal for resource-constrained edge devices. Finally, the trained model can be used as an edge device on the NVIDIA Jetson Nano-kit to speed up detection. The measures of precision, recall, accuracy, and F1 score were utilized for model evaluation. According to the statistical measures, the model accuracy is 99.58%, precision is 100%, recall is 99%, and the F1-Score is 100 %
Blockchain-Based Voting Systems Enhancing Transparency and Security in Electoral Processes
No full textBlockchain technology, when used as designed, serves a transformative purpose in modernizing electoral systems with better transparency, security and voter confidence. Blockchain Isn’t a Voting Panacea Despite issues related to scaling, privacy, existing infrastructures, and so on, there is an immense opportunity for us to have a new way of voting that solves the security, transparency, and efficiency (again: voting should be cheap) issues existing systems have today. By using privacy-preserving solutions, decentralized protocols and smart contracts we can ensure transparency in vote counting and tamper-proof election results, thus ultimately mitigating the opportunity for voter fraud. In addition, the decentralized nature of blockchain enables not only secure voting for monumental elections, but in even resource-scarce environments, making it an ultimate solution to international democratic participation. We will talk about the most severe issues that Blockchain voting systems have to face and how they can be fixed, leading us to a future of electronic voting that is safe, secure, and everyone can access
Machine Learning for Predictive Maintenance Applications in Industrial Equipment and Manufacturing Processes
No full textUtilization of predictive maintenance, backed by machine learning, has made a difference in monitoring industrial equipment and manufacturing, cutting down on downtime, improving operational efficiency, and ensuring safety. However, current systems suffer limitations, including lack of real-time deployment, low scalability, significant computation footprints, security vulnerabilities and low interpretability. We present a novel, scalable explainable AI based predictive maintenance framework integrating lightweight deep learning models, federated learning, blockchain secure storage and adaptive self-learning mechanisms. With the application of edge AI computing, interpretable machine learning methods, and real-time industrial data processing, the proposed study realizes a cost-effective, secure, and scalable predictive maintenance solution. A practical and innovative solution for minimizing failures and enhancing manufacturing efficiency involving sustainable smart industrial approaches can be achieved by validating the proposed model in real-world industrial environments
Edge AI Deploying Artificial Intelligence Models on Edge Devices for Real-Time Analytics
No full textBecause of its on-the-go nature, edge AI has gained popularity, allowing for realtime analytics by deploying artificial intelligence models onto edge devices. Despite the promise of Edge AI evidenced by existing research, there are still significant barriers to widespread adoption with issues such as scalability, energy efficiency, security, and reduced model explainability representing common challenges. Hence, while this paper solves the Edge AI in a number of ways, with real use case of a deployment, modular adaptability, and dynamic AI model specialization. Our paradigm achieves low latency, better security and energy efficiency using light-weight AI models, federated learning, Explainable AI (XAI) and smart edge-cloud orchestration. This framework could enable generic AI beyond specific applications that depend on multi-modal data processing, which contributes to the generalization of applications across various industries such as healthcare, autonomous systems, smart cities, and cybersecurity. Moreover, this work will help deploy sustainable AI by employing green computing techniques to detect anomalies in near real-time in various critical domains helping to ease challenges of the modern world
Implementing Data Security in Wireless Sensor Networks through Location Aware Multifunctional Key Management Framework
No full textHyper Parameter Optimization for Transfer Learning of ShuffleNetV2 with Edge Computing for Casting Defect Detection
Get PDFA casting defect is an expendable abnormality and the most undesirable thing in the metal casting process. In Casting Defect Detection, deep learning based on Convolution Neural Network (CNN) models has been widely used, but most of these models require a lot of processing power. This work proposes a low-power ShuffleNet V2-based Transfer Learning model for defect identification with low latency, easy upgrading, increased efficiency, and an automatic visual inspection system with edge computing. Initially, various image transformation techniques were used for data augmentation on casting datasets to test the model flexibility in diverse casting. Subsequently, a pre-trained lightweight ShuffleNetV2 model is adapted, and hyperparameters are fine-tuned to optimize the model. The work results in a lightweight, adaptive, and scalable model ideal for resource-constrained edge devices. Finally, the trained model can be used as an edge device on the NVIDIA Jetson Nano-kit to speed up detection. The measures of precision, recall, accuracy, and F1 score were utilized for model evaluation. According to the statistical measures, the model accuracy is 99.58%, precision is 100%, recall is 99%, and the F1-Score is 100 %
