Intelligent Machine Learning Based Brain Tumor Segmentation through Multi-Layer Hybrid U-Net with CNN Feature Integration

Brain tumors are a pressing public health concern, characterized by their high mortality and morbidity rates.Nevertheless, the manual segmentation of brain tumors remains a laborious and error-prone task, necessitatingthe development of more precise and efficient methodologies. To address this formidable challenge, we proposean advanced approach for segmenting brain tumorMagnetic Resonance Imaging (MRI) images that harnesses theformidable capabilities of deep learning and convolutional neural networks (CNNs). While CNN-based methodshave displayed promise in the realm of brain tumor segmentation, the intricate nature of these tumors, markedby irregular shapes, varying sizes, uneven distribution, and limited available data, poses substantial obstacles toachieving accurate semantic segmentation. In our study, we introduce a pioneering Hybrid U-Net framework thatseamlessly integrates the U-Net and CNN architectures to surmount these challenges. Our proposed approachencompasses preprocessing steps that enhance image visualization, a customized layered U-Net model tailoredfor precise segmentation, and the inclusion of dropout layers to mitigate overfitting during the training process.Additionally, we leverage the CNN mechanism to exploit contextual information within brain tumorMRI images,resulting in a substantial enhancement in segmentation accuracy.Our experimental results attest to the exceptionalperformance of our framework, with accuracy rates surpassing 97% across diverse datasets, showcasing therobustness and effectiveness of our approach. Furthermore, we conduct a comprehensive assessment of ourmethod’s capabilities by evaluating various performance measures, including the sensitivity, Jaccard-index, andspecificity. Our proposed model achieved 99% accuracy. The implications of our findings are profound. Theproposed Hybrid U-Net model emerges as a highly promising diagnostic tool, poised to revolutionize brain tumorimage segmentation for radiologists and clinicians.

Optimization Design of the Multi-Layer Cross-Sectional Layout of An Umbilical Based on the GA-GLM

Marine umbilical is one of the key equipment for subsea oil and gas exploitation,which is usually integrated by a great number of different functional components with multi-layers.The layout of these components directly affects manufacturing,operation and storage performances of the umbilical.For the multi-layer cross-sectional layout design of the umbilical,a quantifiable multi-objective optimization model is established according to the operation and storage requirements.Considering the manufacturing factors,the multi-layering strategy based on contact point identification is introduced for a great number of functional components.Then,the GA-GLM global optimization algorithm is proposed combining the genetic algorithm and the generalized multiplier method,and the selection operator of the genetic algorithm is improved based on the steepest descent method.Genetic algorithm is used to find the optimal solution in the global space,which can converge from any initial layout to the feasible layout solution.The feasible layout solution is taken as the initial value of the generalized multiplier method for fast and accurate solution.Finally,taking umbilicals with a great number of components as examples,the results show that the cross-sectional performance of the umbilical obtained by optimization algorithm is better and the solution efficiency is higher.Meanwhile,the multi-layering strategy is effective and feasible.The design method proposed in this paper can quickly obtain the optimal multi-layer cross-sectional layout,which replaces the manual design,and provides useful reference and guidance for the umbilical industry.

Multi-layer perceptron-based data-driven multiscale modelling of granular materials with a novel Frobenius norm-based internal variable

One objective of developing machine learning(ML)-based material models is to integrate them with well-established numerical methods to solve boundary value problems(BVPs).In the family of ML models,recurrent neural networks(RNNs)have been extensively applied to capture history-dependent constitutive responses of granular materials,but these multiple-step-based neural networks are neither sufficiently efficient nor aligned with the standard finite element method(FEM).Single-step-based neural networks like the multi-layer perceptron(MLP)are an alternative to bypass the above issues but have to introduce some internal variables to encode complex loading histories.In this work,one novel Frobenius norm-based internal variable,together with the Fourier layer and residual architectureenhanced MLP model,is crafted to replicate the history-dependent constitutive features of representative volume element(RVE)for granular materials.The obtained ML models are then seamlessly embedded into the FEM to solve the BVP of a biaxial compression case and a rigid strip footing case.The obtained solutions are comparable to results from the FEM-DEM multiscale modelling but achieve significantly improved efficiency.The results demonstrate the applicability of the proposed internal variable in enabling MLP to capture highly nonlinear constitutive responses of granular materials.

Target Controllability of Multi-Layer Networks With High-Dimensional Nodes

This paper studies the target controllability of multilayer complex networked systems,in which the nodes are highdimensional linear time invariant(LTI)dynamical systems,and the network topology is directed and weighted.The influence of inter-layer couplings on the target controllability of multi-layer networks is discussed.It is found that even if there exists a layer which is not target controllable,the entire multi-layer network can still be target controllable due to the inter-layer couplings.For the multi-layer networks with general structure,a necessary and sufficient condition for target controllability is given by establishing the relationship between uncontrollable subspace and output matrix.By the derived condition,it can be found that the system may be target controllable even if it is not state controllable.On this basis,two corollaries are derived,which clarify the relationship between target controllability,state controllability and output controllability.For the multi-layer networks where the inter-layer couplings are directed chains and directed stars,sufficient conditions for target controllability of networked systems are given,respectively.These conditions are easier to verify than the classic criterion.

Dynamic Multi-Layer Perceptron for Fetal Health Classification Using Cardiotocography Data

Fetal health care is vital in ensuring the health of pregnant women and the fetus.Regular check-ups need to be taken by the mother to determine the status of the fetus’growth and identify any potential problems.To know the status of the fetus,doctors monitor blood reports,Ultrasounds,cardiotocography(CTG)data,etc.Still,in this research,we have considered CTG data,which provides information on heart rate and uterine contractions during pregnancy.Several researchers have proposed various methods for classifying the status of fetus growth.Manual processing of CTG data is time-consuming and unreliable.So,automated tools should be used to classify fetal health.This study proposes a novel neural network-based architecture,the Dynamic Multi-Layer Perceptron model,evaluated from a single layer to several layers to classify fetal health.Various strategies were applied,including pre-processing data using techniques like Balancing,Scaling,Normalization hyperparameter tuning,batch normalization,early stopping,etc.,to enhance the model’s performance.A comparative analysis of the proposed method is done against the traditional machine learning models to showcase its accuracy(97%).An ablation study without any pre-processing techniques is also illustrated.This study easily provides valuable interpretations for healthcare professionals in the decision-making process.

A flexible ultra-broadband multi-layered absorber working at 2 GHz-40 GHz printed by resistive ink

A flexible extra broadband metamaterial absorber(MMA)stacked with five layers working at 2 GHz–40 GHz is investigated.Each layer is composed of polyvinyl chloride(PVC),polyimide(PI),and a frequency selective surface(FSS),which is printed on PI using conductive ink.To investigate this absorber,both one-dimensional analogous circuit analysis and three-dimensional full-wave simulation based on a physical model are provided.Various crucial electromagnetic properties,such as absorption,effective impedance,complex permittivity and permeability,electric current distribution and magnetic field distribution at resonant peak points,are studied in detail.Analysis shows that the working frequency of this absorber covers entire S,C,X,Ku,K and Ka bands with a minimum thickness of 0.098λ_(max)(λ_(max) is the maximum wavelength in the absorption band),and the fractional bandwidth(FBW)reaches 181.1%.Moreover,the reflection coefficient is less than-10 dB at 1.998 GHz–40.056 GHz at normal incidence,and the absorptivity of the plane wave is greater than 80%when the incident angle is smaller than 50°.Furthermore,the proposed absorber is experimentally validated,and the experimental results show good agreement with the simulation results,which demonstrates the potential applicability of this absorber at 2 GHz–40 GHz.

Target layer state estimation in multi-layer complex dynamical networks considering nonlinear node dynamics

In many engineering networks, only a part of target state variables are required to be estimated.On the other hand,multi-layer complex network exists widely in practical situations.In this paper, the state estimation of target state variables in multi-layer complex dynamical networks with nonlinear node dynamics is studied.A suitable functional state observer is constructed with the limited measurement.The parameters of the designed functional observer are obtained from the algebraic method and the stability of the functional observer is proven by the Lyapunov theorem.Some necessary conditions that need to be satisfied for the design of the functional state observer are obtained.Different from previous studies, in the multi-layer complex dynamical network with nonlinear node dynamics, the proposed method can estimate the state of target variables on some layers directly instead of estimating all the individual states.Thus, it can greatly reduce the placement of observers and computational cost.Numerical simulations with the three-layer complex dynamical network composed of three-dimensional nonlinear dynamical nodes are developed to verify the effectiveness of the method.

Multi-layer network embedding on scc-based network with motif

Interconnection of all things challenges the traditional communication methods,and Semantic Communication and Computing(SCC)will become new solutions.It is a challenging task to accurately detect,extract,and represent semantic information in the research of SCC-based networks.In previous research,researchers usually use convolution to extract the feature information of a graph and perform the corresponding task of node classification.However,the content of semantic information is quite complex.Although graph convolutional neural networks provide an effective solution for node classification tasks,due to their limitations in representing multiple relational patterns and not recognizing and analyzing higher-order local structures,the extracted feature information is subject to varying degrees of loss.Therefore,this paper extends from a single-layer topology network to a multi-layer heterogeneous topology network.The Bidirectional Encoder Representations from Transformers(BERT)training word vector is introduced to extract the semantic features in the network,and the existing graph neural network is improved by combining the higher-order local feature module of the network model representation network.A multi-layer network embedding algorithm on SCC-based networks with motifs is proposed to complete the task of end-to-end node classification.We verify the effectiveness of the algorithm on a real multi-layer heterogeneous network.

Dynamic interwell connectivity analysis of multi-layer waterflooding reservoirs based on an improved graph neural network

The analysis of interwell connectivity plays an important role in the formulation of oilfield development plans and the description of residual oil distribution. In fact, sandstone reservoirs in China's onshore oilfields generally have the characteristics of thin and many layers, so multi-layer joint production is usually adopted. It remains a challenge to ensure the accuracy of splitting and dynamic connectivity in each layer of the injection-production wells with limited field data. The three-dimensional well pattern of multi-layer reservoir and the relationship between injection-production wells can be equivalent to a directional heterogeneous graph. In this paper, an improved graph neural network is proposed to construct an interacting process mimics the real interwell flow regularity. In detail, this method is used to split injection and production rates by combining permeability, porosity and effective thickness, and to invert the dynamic connectivity in each layer of the injection-production wells by attention mechanism.Based on the material balance and physical information, the overall connectivity from the injection wells,through the water injection layers to the production layers and the output of final production wells is established. Meanwhile, the change of well pattern caused by perforation, plugging and switching of wells at different times is achieved by updated graph structure in spatial and temporal ways. The effectiveness of the method is verified by a combination of reservoir numerical simulation examples and field example. The method corresponds to the actual situation of the reservoir, has wide adaptability and low cost, has good practical value, and provides a reference for adjusting the injection-production relationship of the reservoir and the development of the remaining oil.

A Well Productivity Model for Multi-Layered Marine and Continental Transitional Reservoirs with Complex Fracture Networks

Using the typical characteristics of multi-layered marine and continental transitional gas reservoirs as a basis,a model is developed to predict the related well production rate.This model relies on the fractal theory of tortuous capillary bundles and can take into account multiple gas flow mechanisms at the micrometer and nanometer scales,as well as the flow characteristics in different types of thin layers(tight sandstone gas,shale gas,and coalbed gas).Moreover,a source-sink function concept and a pressure drop superposition principle are utilized to introduce a coupled flow model in the reservoir.A semi-analytical solution for the production rate is obtained using a matrix iteration method.A specific well is selected for fitting dynamic production data,and the calculation results show that the tight sandstone has the highest gas production per unit thickness compared with the other types of reservoirs.Moreover,desorption and diffusion of coalbed gas and shale gas can significantly contribute to gas production,and the daily production of these two gases decreases rapidly with decreasing reservoir pressure.Interestingly,the gas production from fractures exhibits an approximately U-shaped distribution,indicating the need to optimize the spacing between clusters during hydraulic fracturing to reduce the area of overlapping fracture control.The coal matrix water saturation significantly affects the coalbed gas production,with higher water saturation leading to lower production.

Multi-layer phenomena in petawatt laser-driven acceleration of heavy ions

Laser-accelerated high-flux-intensity heavy-ion beams are important for new types of accelerators.A particle-in-cell program(Smilei) is employed to simulate the entire process of Station of Extreme Light(SEL) 100 PW laser-accelerated heavy particles using different nanoscale short targets with a thickness of 100 nm Cr, Fe, Ag, Ta, Au, Pb, Th and U, as well as 200 nm thick Al and Ca. An obvious stratification is observed in the simulation. The layering phenomenon is a hybrid acceleration mechanism reflecting target normal sheath acceleration and radiation pressure acceleration, and this phenomenon is understood from the simulated energy spectrum,ionization and spatial electric field distribution. According to the stratification, it is suggested that high-quality heavy-ion beams could be expected for fusion reactions to synthesize superheavy nuclei. Two plasma clusters in the stratification are observed simultaneously, which suggest new techniques for plasma experiments as well as thinner metal targets in the precision machining process.

The role of polyurethane foam compressible layer in the mechanical behaviour of multi-layer yielding supports for deep soft rock tunnels

The polyurethane foam(PU)compressible layer is a viable solution to the problem of damage to the secondary lining in squeezing tunnels.Nevertheless,the mechanical behaviour of the multi-layer yielding supports has not been thoroughly investigated.To fill this gap,large-scale model tests were conducted in this study.The synergistic load-bearing mechanics were analyzed using the convergenceconfinement method.Two types of multi-layer yielding supports with different thicknesses(2.5 cm,3.75 cm and 5 cm)of PU compressible layers were investigated respectively.Digital image correlation(DIC)analysis and acoustic emission(AE)techniques were used for detecting the deformation fields and damage evolution of the multi-layer yielding supports in real-time.Results indicated that the loaddisplacement relationship of the multi-layer yielding supports could be divided into the crack initiation,crack propagation,strain-hardening,and failure stages.Compared with those of the stiff support,the toughness,deformability and ultimate load of the yielding supports were increased by an average of 225%,61%and 32%,respectively.Additionally,the PU compressible layer is positioned between two primary linings to allow the yielding support to have greater mechanical properties.The analysis of the synergistic bearing effect suggested that the thickness of PU compressible layer and its location significantly affect the mechanical properties of the yielding supports.The use of yielding supports with a compressible layer positioned between the primary and secondary linings is recommended to mitigate the effects of high geo-stress in squeezing tunnels.

基于改进U-net的少样本煤岩界面图像分割方法

煤岩图像语义分割技术是煤岩界面识别的重要研究方向,现有的语义分割模型通常依赖于大样本数据集进行训练,然而目前已标注的煤岩图像数据样本难以获取,并且缺乏公开数据集。针对以上问题,提出了一种基于改进U-net模型的样本煤岩界面图像分割模型。将裁剪后具有更强特征提取能力且结构上更为简单的VGG16替换U-net的原始骨干特征提取网络,提升对图像信息的特征提取能力并获得更快的训练速度,在U-net网络的跳跃连接和解码器上采样部分引入注意力机制模块,对提取的特征层进行处理,提升模型对煤岩界面图像关键特征的提取能力,提高分割精度。使用迁移学习方法对改进的模型进行预训练,提高模型泛化能力同时避免过拟合,使模型更适用于小样本数据集训练。通过使用自制的煤岩界面数据集对所改进的网络模型性能进行验证,并将该模型与经典Unet、DeepLabv3+、PspNet、HrNet网络模型进行了对比。试验结果表明:在同样使用由125幅煤岩界面图片构建的小样本数据集进行训练的情况下,所提改进模型相较于经典U-net模型在分割精确度和检测效率方面都有显著提升,模型精确度提高了1.84%,平均交并比提高了5.34%,类别平均像素准确率提高了0.48%,检测速度增幅为5.3%。同时,与其他网络模型相比,所提改进模型在小样本煤岩界面图像的语义分割中优势显著,表明所提改进思路的有效性。

基于残差U-Net和自注意力Transformer编码器的磁场预测方法

利用有限元方法对几何结构复杂的电机和变压器进行磁场分析,存在仿真时间长且无法复用的问题。因此,该文提出一种基于残差U-Net和自注意力Transformer编码器的磁场预测方法。首先建立永磁同步电机(PMSM)和非晶合金变压器(AMT)有限元模型,得到深度学习训练所需的数据集;然后将Transformer模块与U-Net模型结合,并引入短残差机制建立ResUnet-Transformer模型,通过预测图像的像素实现磁场预测;最后通过Targeted Dropout算法和动态学习率调整策略对模型进行优化,解决拟合问题并提高预测精度。计算实例证明,ResUnet-Transformer模型在PMSM和AMT数据集上测试集的平均绝对百分比误差(MAPE)均小于1%,且仅需500组样本。该文提出的磁场预测方法能减少实际工况和多工况下精细模拟和拓扑优化的时间和资源消耗,亦是虚拟传感器乃至数字孪生的关键实现方法之一。

一种改进U-Net网络的心电图分类算法研究

基于CPSC-2018十二导联数据,提出了一种U-Net网络和注意力机制结合的心电图分类算法。首先,针对数据集数据长度长短不一的问题,对数据进行等长处理和归一化处理。然后,利用U-Net网络中跳层连接和编码解码方式,对预处理后较长的数据进行处理。在U-Net网络解码的最后一层加入注意力机制对抗噪声,提升模型的有效信息关注度和准确性。最后,利用CPSC-2018数据集进行验证。实验结果表明:所提模型能够取得较好的分类效果,识别房颤(AF)和右束支传导阻滞(RBBB)心律失常的精准率、召回率、F1值都可以达到90%以上,平均F1值可以达到82.5%。

基于改进U-Net的根系表型参数测量系统

为解决背景噪声干扰下,从微根管采集的原位根系图像中难以直接提取准确的表型参数问题,提出一种基于改进U-Net的微根管根系表型参数测量系统。在U-Net网络中引入优化后的空洞空间金字塔池化模块(Atrous Spatial Pyramid Pooling,ASPP)和高效通道注意力模块(Efficient Channel Attention,ECA),增大感受野,提升模型捕捉根系细节特征的能力,获取精确的根系分割图像。结果表明,改进的U-Net模型平均交并比和平均像素精度分别为87.07%和91.85%,相较原始U-Net分别提高了2.49%和2.3%。与WinRHIZO根系分析软件测量值相比,根长度和面积决定系数分别为0.951 8和0.984 9,Spearman相关系数分别为0.972 5和0.975 7,可以实现根系长度和面积的准确测量。

面向青花瓷碎片图像的U-Net++拼接网络

针对现有图像拼接方法存在拼接处伪影以及非重叠区域内容失真,导致较低的准确性和鲁棒性的问题,提出一种基于U-Net++消除伪影的青花瓷碎片图像拼接方法.首先估计待拼接图像单应性矩阵;然后将单应性矩阵应用于结构拼接阶段,得到图像粗拼接结果;最后以图像粗拼接结果作为先验信息,在内容校正阶段改进现有的U-Net,利用U-Net++细化粗拼接结果,得到最终图像精确拼接.以青花瓷碎片图像数据集与相关经典方法进行实验的结果表明,在3个评价指标中,所提方法的峰值信噪比提高约13%,均方根误差降低约33%,均方误差降低57%左右;该方法具有较小的误差比,不仅能够提高图像拼接质量,而且表现出较好的鲁棒性.

基于U-NET的双分支海上SAR溢油检测模型

为提高海上溢油SAR(Synthetic Aperture Radar)检测的准确率,本文提出一种基于U-NET和注意力门的海上溢油SAR检测模型(AW-net),该模型将U-NET中传统的单输入编码器替换为双分支编码器,分别输入纹理特征和SAR灰度特征,并进一步采用注意力门融合纹理信息和灰度信息。实验利用1景海丝一号(HISEA-1)SAR数据构建样本训练集进行AW-net模型训练,分别应用1景HISEA-1 SAR数据和1景Radarsat-2SAR数据开展模型测试,溢油检测准确率均优于U-NET、AttentionU-NET和FCN等语义分割模型,说明该模型具有较强的强鲁棒性和应用潜力。

基于AttentionR2U-net的岩石(体)关键节理智能识别与参数提取

针对岩石(体)表面复杂节理网中关键节理的智能识别与参数提取问题,提出一种基于AttentionR2U-net网络与节理几何特征模型耦合识别的方法.在R2U-net网络的基础上引入注意门(attentiongate)改进网络,通过定性与定量的方法对边坡节理图像和混凝土、龟裂土、常见脆性岩石裂隙图像的识别结果分别作准确性及泛化能力检验;利用AttentionR2U-net网络耦合节理几何特征的方法识别关键节理,提取原始节理和关键节理的几何参数并对其迹长、面积及倾角作差异性分析.研究结果表明:针对岩石(体)节理识别,本文算法的Dice相似系数从U-net网络的0.965提升至0.990,且明显优于传统算法,故本文算法在岩石(体)节理识别上具有更强的可靠性与优越性;针对混凝土、龟裂土和大理岩、花岗岩、砂岩等脆性岩石裂隙的识别,本文算法的Dice相似系数均在0.953以上,故本文算法具有较强的泛化能力.与原始节理网络相比,关键节理网络优势迹长由0.732m显著增大至1.835m,节理倾角分布形式和优势倾角组均不变,优势迹长和倾角的节理占比均显著增大.

改进的U-Net算法在管道内焊缝缺陷图像分割中的应用

【目的】图像处理技术在管道焊缝识别系统中的应用已经成为了机器视觉在焊缝检测中主要应用方向。对焊缝表面缺陷进行识别是应用的关键技术。为了提高焊缝表面缺陷识别效果,需要对焊缝图像进行有效分割。针对管道内焊缝边界区域可能出现的模糊不清,导致分割结果不准确的现象,需要采取相应的技术有段进行改善。【方法】针对管道内焊缝缺陷图像分割问题,提出一种改进的U-Net图像分割方法。以管道内焊缝图像为研究对象,采用改进型U-Net网络对管道内焊缝缺陷图像进行识别和分割,经过网络训练和模型测试后,将分割结果与原U-Net网络、FCN网络进行对比。【结果】结果表明,在改进型U-Net网络对管道内焊缝缺陷图像的分割中,相似性系数(Dice)、平均交并比(mIoU)两项评价指标分别达到0.8420和0.8514,相较于FCN网络分别提升13.44%和8.68%,相较于原U-Net网络分别提升6.51%和3.31%。【结论】因此,该文提出的改进后的U-Net网络对管道内焊缝缺陷的识别和分割具有更好的效果,也为研究管道焊缝缺陷识别系统提供可靠基础,减少人工检测的成本和时间。