With the explosive growth in amount of medical data, traditional data analysis methods can hardly meet demand, especially complex tasks such as disease diagnosis, patient monitoring and personalized treatment. A deep learning-based system for intelligent recognition has emerged, which is capable automatically extracting features from massive efficiently learning through multi-layer neural networks, thus significantly improving diagnostic accuracy efficiency. The not only covers a variety...

Traditional methods for seismic damage evaluation require manual extractions of intensity measures (IMs) to properly represent the record-to-record variation ground motions. Contemporary such as convolutional neural networks (CNNs) time series classification and face a challenge in training due huge task ground-motion image encoding. Presently, no consensus has been reached on understanding most suitable encoding technique size (width × height channel) CNN-based evaluation. In this study, we...

Abstract The availability of reliable probabilistic capacity models reinforced concrete (RC) columns is a cornerstone for high‐confidence seismic fragility and risk analyses highway bridges. Existing studies often perform physics‐based pushover or moment–curvature the modeling RC columns, which may encounter nonconvergent problems under high levels nonlinearities in structural material constitutive elements, become computationally inefficient especially when analysis model contains plenty...

Wood-frame structures are used in almost 90% of residential buildings the United States. It is thus imperative to rapidly and accurately assess damage wood-frame wake an earthquake event. This study aims develop a machine-learning-based seismic classifier for portfolio 6,113 near New Madrid Seismic Zone (NMSZ) which synthesized ground motions adopted characterize potential earthquakes. classifier, based on multilayer perceptron (MLP), compared with existing fragility curves developed same...

Contemporary deep learning approaches for post-earthquake damage assessments based on 2D convolutional neural networks (CNNs) require encoding of ground motion records to transform their inherent 1D time series images, thus requiring high computing and resources. This study develops a CNN model avoid the costly image encoding. The is compared with wavelet feedforward network (FNN) evaluate prediction performance computational efficiency. A case benchmark reinforced concrete (r/c) building...

This paper takes a half-through steel truss arch bridge as an example. A seismic analysis is conducted with nonlinear finite element method. Contrast models are established to discuss the effect of simplified method for main girder on accuracy result. The influence wave direction and wave-passage behaviors analysed well superstructure ring interaction which mostly related supported bearings wind resistant springs. In end, application cable-sliding aseismic devices discussed put forward...

In this paper, we propose a novel graph-based salient object detection algorithm which exploits higher order potential to capture the cross-scale grouping cues instead of using multi-scale graph model or naive fusion (i.e. individually compute saliency result for each scale and then combine them). And, investigate importance affinities in labeling. We take both local (spatial distribution) nonlocal (feature priors into account learn pairwise similarity values semi-supervised manner, thereby...

Bridges have recently been exposed to an increasing number of natural hazards such as earthquakes and tsunamis. These extreme events resulted in transverse offsets, overturning moments, even dropping-off superstructures due their weak connection substructures. outcomes are potentially prevented or mitigated by developing deploying sliding, modular, adaptive, replaceable, two-dimensional (SMART) shear keys fuse elements between The novelty SMART is enable adaptive control both the force...

Abstract The crack pattern of steel reinforced ultrahigh performance concrete (UHPC) beam is usually characterized by many densely distributed fine cracks (i.e., multiple microcracks) along with localized macrocrack, and the width development rate height smaller than that normal since fibers reinforcement bars are supposed to be effective in controlling propagation UHPC beam. However, an prediction formula still underdeveloped for present study aims formulate a equation based on equations...

The design and application of medical imaging diagnostic assistance systems have become an indispensable facet modern medicine. Technologies based on machine learning can handle vast intricate data, offering precise support significantly enhancing the efficiency accuracy clinical diagnoses. This study focuses development implementation efficient system utilizing algorithms such as Principal Component Analysis (PCA), Support Vector Machines (SVM), Multi-Layer Perceptrons (MLP). Through...

This paper defines series of important performance evaluation parameters to evaluate Amplitude-Phase (AP) algorithm comparing with that Block Adaptive Quantization (BAQ) algorithm. The procedure is carried out in two domains, raw data domain and image domain. Numerical experiments based on ERS-2 show AP provides us more Compression Ratio (CR) choices than BAQ for certain CR, at least one choice whose better or equal BAQ. These algorithms neither affect spatial resolution nor generate...

Severe fatigue and noise problems of modular bridge expansion joints (MBEJs) are often induced by vehicle loads. However, the dynamic characteristics single-support MBEJs have yet to be further investigated. To better understand vibration mechanism under loads, a 3D finite element model MBEJ with five center beams is built. Successive loads given out vertical responses each analyzed successive Dynamic amplification factors (DAFs) also calculated along increasing velocities from 20 km/h 120...