Recently, deep image-hiding techniques have attracted considerable attention in covert communication and highcapacity information hiding.However, these approaches some limitations.For example, a cover image lacks self-adaptability, leakage, or weak concealment.To address issues, this study proposes universal adaptable method.First, domain mechanism is designed by combining the Atrous convolution, which makes better use of relationship between secret domain.Second, to improve perceived human...

The incipient vehicular applications and the explosive growth of data generally led to increase in demand for communication, computation, storage resources, as well stringent performance requirement on latency wireless network capacity. To address these challenges, Vehicular Edge Computing is an envisioned promising solution that extends computation capability edge offloading service provided proximity vehicles. computing (VEC) D2D communication have been designed benefit from gain optimize...

Machine learning (ML) models are essential to securing communication networks. However, these vulnerable adversarial examples (AEs), in which malicious inputs modified by adversaries produce the desired output. Adversarial training is an effective defense method against such attacks but relies on access a substantial number of AEs, prerequisite that entails significant computational resources and inherent limitation poor performance clean data. To address problems, this study proposes novel...

Forecasting passenger flow at metro transit stations is a useful method for optimizing the organization of transportation and enhancing operational safety efficiency. Aiming problem that traditional ARIMA model has poor performance in predicting flow, hybrid prediction based on ARIMA-Kalman filtering proposed. In this regard, training experimental samples are integrated with Kalman filter to create recursion equation, which then utilized estimate flow. The simulation experiment results...

In recent years, various adversarial defense methods have been proposed to improve the robustness of deep neural networks.Adversarial training is one most potent defend against attacks.However, difference in feature space between natural and examples hinders accuracy model training.This paper proposes a learnable distribution method, aiming construct same for data utilizing Gaussian mixture model.The centroid built classify samples constrain sample features.The are pushed method generates...

Different kinds of attacks on the network have greatly increased due to exponential increase in number users over last decade. This has hampered transactions online especially when they are financially based. Hence, there is every need develop and design new cybersecurity techniques curb these cyber-crimes. study presents accurate classification Transmission Control Protocol (TCP) User Datagram (UDP) flows using Machine Learning (ML) based parameter optimization. The method was validated...

Deep learning models have enabled significant performance improvements to remote sensing image processing. Usually, a large number of training samples is required for detection models. In this study, dynamic simulation strategy designed generate in real time during training. The few adversarial examples are not only directly involved the but also used fit distribution model noise, helping real-time generated be similar examples. noise randomly according model, and random variation inputs...

<abstract><p>Adversarial examples have been shown to easily mislead neural networks, and many strategies proposed defend them. To address the problem that most transformation-based defense will degrade accuracy of clean images, we an Enhanced Image Transformation Generative Adversarial Network (EITGAN). Positive perturbations were employed in EITGAN counteract adversarial effects while enhancing classified performance samples. We also used image super-resolution method mitigate...