The complexity of many IT services and facilities has been continuously increasing, and the complexity of related monitoring systems and the difficulty of managing it are also growing rapidly. The integration and anal...
详细信息
With recent advances in autonomous driving, voice control systems have become increasingly adopted as human-vehicle interaction methods. This technology enables drivers to use voice commands to control the vehicle and...
详细信息
ISBN:
(纸本)9781665416597
With recent advances in autonomous driving, voice control systems have become increasingly adopted as human-vehicle interaction methods. This technology enables drivers to use voice commands to control the vehicle and will be soon available in Advanced Driver Assistance systems (ADAS). Prior work has shown that Siri, Alexa and Cortana, are highly vulnerable to inaudible command attacks. This could be extended to ADAS in real-world applications and such an inaudible command threat is difficult to detect due to microphone nonlinearities. In this paper, we aim to develop a more practical solution by using camera views to defend against inaudible command attacks where ADAS are capable of detecting their environment via multi-sensors. To this end, we propose a novel multimodal deep learning classification system to defend against inaudible command attacks. Our experimental results confirm the feasibility of the proposed defense methods and the best classification accuracy reaches 89.2%. Code is available at https://***/ITSEG-MQ/sensor-Fusion-Against-VoiceCommand-Attacks.
Ocular regions such as iris and sclera yield high accuracy in user's biometrics as well as liveness detection systems. The primary purpose for ocular recognition system is the accurate segmentation of the regions ...
详细信息
ISBN:
(纸本)9781728160344
Ocular regions such as iris and sclera yield high accuracy in user's biometrics as well as liveness detection systems. The primary purpose for ocular recognition system is the accurate segmentation of the regions of interest that plays the key role in retaining the accuracy and restraining the errors in the whole system. However, accurate ocular regions segmentation in the images in a physical environment is very challenging owing to the images with low resolution, occlusion, blur, ghost effect, unusual glint, and off-angles. Deep learning algorithms with a convolutional neural network (CNN) has achieved promising results for ocular regions segmentation. However, previous CNN-based methods are unable to find the true boundary of ocular regions in non-ideal situations, which results in reduced reliability and accuracy. To overcome these challenges, we present OcularNet, a deep learning-based lite-residual encoder-decoder network to determine the accurate ocular regions such as iris and sclera. In this way, the true ocular regions can be segmented with the transfer of high-frequency information using residual skip connections. Additionally, the proposed Ocular-Net does not enhance performance on the cost of increasing depth, complexity or number of parameters, in fact, it has much fewer parameters than the previous state-of-the-art methods. We performed comprehensive experiments and obtained optimum performance on iris and sclera datasets.
Authors of the article draw their attention to the study of methodological decomposition and parallelization applied for the tasks of synthesis and operation of complex physical and chemical systems in order to develo...
详细信息
ISBN:
(纸本)9781728189901
Authors of the article draw their attention to the study of methodological decomposition and parallelization applied for the tasks of synthesis and operation of complex physical and chemical systems in order to develop a horizontally scalable distributed information system. This system relies on algorithms rooted in a unified approach to the description and modelling of variable physical and chemical processes in the systems under investigation. Then the results of process modelling are applied to synthesize this system model. This model in its turn is used for the tasks of synthesis and operation of studied complex systems. In terms of user interface, the information system of this kind is expandable over various practical domains associated with complex physical and chemical systems (power engineering, chemical technology, cellular biology, meteorology, etc.).
The proceedings contain 6 papers. The topics discussed include: accelerating microstructural analytics with Dask for volumetric x-ray images;enabling system wide shared memory for performance improvement in PyCOMPSs a...
ISBN:
(纸本)9780738110868
The proceedings contain 6 papers. The topics discussed include: accelerating microstructural analytics with Dask for volumetric x-ray images;enabling system wide shared memory for performance improvement in PyCOMPSs applications;experiences in developing a distributed agent-based modeling toolkit with python;data engineering for HPC with python;python workflows on HPC systems;and distributed asynchronous array computing with the jetlag environment.
A distributed denial of service attack is a major security challenge in modern communications networks. In this article, we propose models that capture all the key performance indicators of synchronized denial of serv...
详细信息
Traffic sign recognition (TSR) system is essential for autonomous vehicle and is vulnerable to security threats from adversarial attacks. The existing adversarial attacks for TSR are invasive and suffer from poor conc...
详细信息
Traffic sign recognition (TSR) system is essential for autonomous vehicle and is vulnerable to security threats from adversarial attacks. The existing adversarial attacks for TSR are invasive and suffer from poor concealment and high computational complexity, and thus have low feasibility in real-world scenarios. This paper proposes a non-invasive modulated LED illumination-based adversarial attack scheme. By generating luminance flashes imperceptible to human eyes through fast intensity modulation of lighting such as LED streetlights and exploiting the rolling shutter mechanism of CMOS sensors of in-vehicle imaging system, the proposed attack scheme can successfully perform adversarial attacks on TSR system by implanting luminance information perturbations into the images acquired by autonomous vehicle and thus poisoning the image data fed into TSR system. Depending on the modulation frequency and pattern of LED illumination, the proposed attack scheme enables denial of service (DoS) attack that leads to traffic sign detection failure and escape attack that leads to traffic sign misclassification, with the advantages of superior concealment, low computational complexity and high practical feasibility. Experiments are conducted with two benchmark datasets (GTSDB and GTSRB) and two state-of-the-art models of TSR detection and TSR classification, YOLOv5m and Sill-Net respectively, in both the digital and physical world. Experimental results show that the proposed DoS attack on the TSR detection model (YOLOv5m) can reach the success rate of 90.00% and the proposed escape attack on the TSR classification model (Sill-Net) can achieve the success rate of 35.00%.
Anomaly detection is increasingly important to handle the amount of sensor data in Edge and Fog environments, Smart Cities, as well as in Industry 4.0. To ensure good results, the utilized ML models need to be updated...
详细信息
Anomaly detection is increasingly important to handle the amount of sensor data in Edge and Fog environments, Smart Cities, as well as in Industry 4.0. To ensure good results, the utilized ML models need to be updated periodically to adapt to seasonal changes and concept drifts in the sensor data. Although the increasing resource availability at the edge can allow for in-situ execution of model training directly on the devices, it is still often offloaded to fog devices or the cloud. In this paper, we propose Local-Optimistic Scheduling (LOS), a method for executing periodic ML model training jobs in close proximity to the data sources, without overloading lightweight edge devices. Training jobs are offloaded to nearby neighbor nodes as necessary and the resource consumption is optimized to meet the training period while still ensuring enough resources for further training executions. This scheduling is accomplished in a decentralized, collaborative and opportunistic manner, without full knowledge of the infrastructure and workload. We evaluated our method in an edge computing testbed on real-world datasets. The experimental results show that LOS places the training executions close to the input sensor streams, decreases the deviation between training time and training period by up to 40% and increases the amount of successfully scheduled training jobs compared to an in-situ execution.
Load balancing is the process of improving the performance of the system by sharing of workload among the processors. The workload of a machine means the total processing time it requires to execute all the tasks assi...
详细信息
Massive deployment of diverse ultra-low power wireless devices in different application areas has given rise to a plethora of heterogeneous architectures and communication protocols. It is challenging to provide conve...
详细信息
ISBN:
(纸本)9781728105703
Massive deployment of diverse ultra-low power wireless devices in different application areas has given rise to a plethora of heterogeneous architectures and communication protocols. It is challenging to provide convergent access to these miscellaneous collections of communicating devices. In this paper, we propose VGATE, an edge-based virtualized IoT gateway for bringing these devices together in a single framework using SDRs as technology-agnostic radioheads. SDR platforms, however, suffer from large unpredictable delays. We design a GNU Radio-based ieee 802.15.4 experimental setup using LimeSDR, where the data path is time-stamped at various points of interest to get a comprehensive understanding of the characteristics of the delays. Our analysis shows that GNU Radio processing and LimeSDR buffering delays are the major delays. We decrease the LimeSDR buffering delay by decreasing the USB transfer size but show that this comes at the cost of increased processing overhead. We modify the USB transfer packet size to investigate which USB transfer size provides the best balance between buffering delay and processing overhead across two different host computers. Our experiments show that for the best measured configuration the mean and jitter of latency decreases by 37% and 40% respectively for the host computer with higher processing resources. We also show that the throughput is not affected by these modifications.
暂无评论