Accurate and up-to-date road maps are of great importance in a wide range applications. Unfortunately, automatic extraction from high-resolution remote sensing images remains challenging due to the occlusion trees buildings, discriminability roads, complex backgrounds. To address these problems, especially connectivity completeness, this article, we introduce novel deep learning-based multistage framework accurately extract surface centerline simultaneously. Our consists three steps:...

Road surface extraction from remote sensing images using deep learning methods has achieved good performance, while most of the existing are based on fully supervised learning, which requires a large amount training data with laborious per-pixel annotation. In this article, we propose scribble-based weakly road method named ScRoadExtractor, learns easily accessible scribbles such as centerlines instead densely annotated ground truths. To propagate semantic information sparse to unlabeled...

Road surface extraction from high-resolution remote sensing images has many engineering applications; however, extracting regularized and smooth road maps that reach the human delineation level is a very challenging task, substantial time-consuming manual work usually unavoidable. In this article, to solve problem, we propose novel framework by introducing graph neural network (GNN) for processing preconstructed easily accessible centerlines. The proposed formulates problem as two-sided...

Through-silicon-via (TSV) enables vertical connectivity between stacked chips or interposer and is a key technology for 3-D integrated circuits (ICs). While arrays of TSVs are needed in IC, there only exists frequency-dependent resistance, inductance, conductance capacitance circuit model pair with coupling them. In this paper, we develop simple yet accurate multiport TSV network (e.g., coupled array) by decomposing the into number pairs then applying models each We call new pair-based...

In the process of mosaicking regional synthetic aperture radar (SAR) intensity images, multiple images with significant brightness anomalies can cause a considerable number pixels to exceed grayscale quantization range. Applying traditional color harmonization methods increases this issue, causing loss information. We propose multi-objective gray consistency correction method designed explicitly for SAR address this. constructed two-objective optimization model ensure image and mitigate...

High speed serial links usually have extremely tight requirement on the quality of signal channels, in terms insertion loss and return loss. Along with an end-to-end channel design, transition from plated-through-hole (PTH) via to fan-out traces printed circuit board (PCB) creates unavoidable impedance discontinuity, which greatly impacts performance. It is important understand model this discontinuity for optimization purpose. This paper discusses several approaches improving channel's...

Fast growing bandwidth demands for networking and data center applications continuously face difficult signal integrity challenges. Among them, serializer/deserializer (SerDes) package pin-out its corresponding printed circuit board (PCB) via crosstalk have become a significant source of high jitter limit the performance speed serial links. In this paper, variety far-end (FEXT) near-end (NEXT) SerDes patterns are designed investigated to explore interfaces running 25Gb/s beyond. The impact...

A novel 3-PSS/7R spatial hybrid redundantly driven mechanism (HRDM) is proposed. The by constant velocity (CV) motor and servo at the same time. CV provides main power for system, plays role of motion regulation. This system not only has advantages traditional mechanical such as stable operation, large bearing capacity, but also can output controllable adjustable flexible trajectory. degrees freedom are determined middle limb, which called both motor. Based on closed vector method, position...

Drawing and updating road networks are both time-consuming labor-intensive. Deep learning technology high-resolution remote sensing images have provided opportunities for automatic extraction. However, recent convolutional neural network (CNN) based segmentation methods shown serious problems on connectivity; tracing with single starting point perform well in connectivity but often result part areas unreached. We propose a multiple points tracer which benefits from methods. compare our...

Tonle Sap Lake is the largest river-connected lake, buffer area and ecological zone of Mekong River, which plays a huge role in dispelling flood peak compensating water, conservation biological diversity. The river-lake relationship between River unique has always been major focus international community. land terrain under-water topography were used to analyze morphological characteristics Cambodia Delta Lake. Long series hydrological data controlling stations water level variation volume...

High performance ASIC packages are typically mounted on the PCB using BGA solder ball technology; package to board transition creates impedance discontinuity in multi-gigabit signaling channel. It is important understand and model this accurately improve end channel design system level. Usually when simulated, instead of modeling with together one model, also known as piece separate models built for individual then cascaded circuit simulator. If not setup correctly field solvers, i.e. port...

Differential signaling is widely used in high speed data communications. Inter-symbol interference (ISI) and crosstalk between differential pair, however, heavily affect the integrity of as measured by timing jitter amplitude noise eye diagram. To reduce impact ISI, a pre-emphasis filter commonly used, but it increases noise. In this paper, we first propose formula-based models considering combined effect crosstalk, filter. With given input patterns, our achieve within 5% difference compared...

With the thriving of wearable devices and widespread use smartphones, e-healthcare system emerges to cope with high demand health services. However, this integrated smart is vulnerable various attacks, including intrusion attacks. Traditional detection schemes generally lack classifier diversity identify attacks in complex scenarios that contain a small amount training data. Moreover, cloud-based attack may result higher latency. In paper, we propose an Edge-assisted Anomaly Detection (EAD)...

Joint design time and post-silicon optimization for analog circuits has been an open problem in literature because of the complex nature circuit modeling optimization. In this paper we formulate co-optimization digitally tuned to optimize parametric yield, subject power area constraints. A general framework combing branch-and-bound algorithm gradient ascent method is proposed. We demonstrate our with two examples high-speed serial link, transmitter phase-locked-loop (PLL) design. Simulation...

Power distribution networks (PDNs) are designed and analyzed iteratively. Random walk is among the most efficient methods for PDN analysis. We develop in this paper an incremental on-demand random to reduce iterative analysis time. During each iteration, we map design changes as positive or negative walks observed nodes. To update result, only need apply these extra walks, instead of doing all from scratch. show that different execution orders do not affect accuracy but runtime because...

A hybrid CS-DL method for aircraft classification in complex electromagnetic environment is introduced. To classify from interfered radar echoes, the authors propose a novel based on compressed sensing (CS) and deep-learning (DL). After recovering spectrum polluted by jamming signals using CS, they exploit sparse auto-encoder (SAE) to extract modulation features then aircraft. The tested 536 flights of three types airplanes, results show that correct rate reaches 75% even when 41% pulses are...

The lateral velocity is of importance in cases like target identification and traffic management. Conventional Doppler methods are not capable measuring velocities since they quantify only the radial component. Based on spectrogram characteristic laterally moving targets, an algorithm based fractional Fourier transform has been studied signal processing literature. searches peak position transformation, calculates from position. performance analysis this carried out paper, which shows that...