期刊文献+
共找到2,099篇文章
< 1 2 105 >
每页显示 20 50 100
多尺度非对称注意力遥感去雾Transformer
1
作者 王旭阳 梁宇航 《广西师范大学学报(自然科学版)》 北大核心 2026年第2期77-89,共13页
雾霾干扰会导致遥感图像结构模糊、细节丢失,严重影响下游视觉任务的准确性。为此,本文提出一种异构增强的遥感图像去雾网络,从空间结构建模与频率信息整合2个层面提升特征恢复能力。具体而言,设计多尺度非对称注意力Transformer模块,... 雾霾干扰会导致遥感图像结构模糊、细节丢失,严重影响下游视觉任务的准确性。为此,本文提出一种异构增强的遥感图像去雾网络,从空间结构建模与频率信息整合2个层面提升特征恢复能力。具体而言,设计多尺度非对称注意力Transformer模块,引入方向感知机制以增强模糊边缘与纹理细节的建模;同时构建基于小波变换高低频自适应增强模块,使用Haar小波分解分离频域信息,分别通过高频与低频子模块强化边缘轮廓与结构表达。2个模块分别嵌入特征提取与融合阶段,协同缓解传统方法方向性建模不足与高频特征易丢失等问题。在保持低计算开销的前提下,本文方法在HAZE1K与RICE数据集上的平均PSNR/SSIM性能分别达到24.9936/0.9099与33.1802/0.8942,在细节恢复方面表现出显著优势。 展开更多
关键词 遥感图像去雾 transformer 非对称注意力 高低频特征增强 小波变换 方向感知建模 深度学习
在线阅读 下载PDF
基于CNN-BiLSTM-Transformer的舰船中压直流全电推进系统故障诊断设计 被引量:2
2
作者 张建良 韩涛 季瑞松 《实验技术与管理》 北大核心 2025年第1期11-18,共8页
针对舰船中压直流全电推进系统结构复杂度高、单元耦合性强、运行环境多变等特点造成的故障诊断准确性低和实时性差等问题,开展了基于CNN-BiLSTM-Transformer的故障诊断设计。首先,基于卷积神经网络CNN构建单点特征级联网络,开展单一时... 针对舰船中压直流全电推进系统结构复杂度高、单元耦合性强、运行环境多变等特点造成的故障诊断准确性低和实时性差等问题,开展了基于CNN-BiLSTM-Transformer的故障诊断设计。首先,基于卷积神经网络CNN构建单点特征级联网络,开展单一时刻下故障信号空间特征的深入提取,以提升故障特征提取的有效性;其次,以双向长短期记忆网络BiLSTM为核心设计多点特征依赖网络,利用门控机制和双向时序学习机制,实现故障信号在多个时刻之间特征依赖关系的有效学习,以提升故障诊断的准确性;然后,以Transformer为核心建立序列特征并行处理网络,通过自注意力机制实现对故障特征上下文关系的精确刻画,进而利用多头注意力机制实现特征序列的并行处理,以提升故障诊断的实时性;最后,设计舰船中压直流全电推进系统故障诊断实验方案,并开展不同故障模式下的诊断性能评估。该文方法在多种故障模式下诊断准确率和实时性均优于现有的主流故障诊断方法,有助于为舰船中压直流全电推进系统的安全运行提供更有力的技术保障。 展开更多
关键词 舰船 中压直流 全电推进系统 故障诊断 transformer
在线阅读 下载PDF
基于循环双向Transformer的伪CT图像生成方法
3
作者 肖宁 赵俊 +3 位作者 贾保平 强彦 赵涓涓 吕亚丽 《计算机辅助设计与图形学学报》 北大核心 2025年第6期994-1005,共12页
磁共振成像引导的放射治疗可以根据肿瘤和对器官的威胁情况实时调整治疗计划,依靠使用磁共振成像生成伪计算机断层扫描进行放射治疗.目前,伪层析成像的生成技术基于对抗性网络的生成方法,但它在训练过程中使用像素级损失更新网络参数,... 磁共振成像引导的放射治疗可以根据肿瘤和对器官的威胁情况实时调整治疗计划,依靠使用磁共振成像生成伪计算机断层扫描进行放射治疗.目前,伪层析成像的生成技术基于对抗性网络的生成方法,但它在训练过程中使用像素级损失更新网络参数,很容易导致模式崩溃,生成不稳定的伪计算机断层扫描.为了精准地实现基于磁共振图像的伪计算机断层扫描生成,利用视觉Transformer的上下文敏感性以及卷积算子的归纳偏置,提出一种循环双向Transformer医学图像合成方法.在编码预测阶段,循环双向Transformer利用U-Net编码得到的码本表示图像,并使用非自回归编码与向量量化方式缩短生成码本的长度,生成局部真实并且全局一致的图像;使用归一化互信息作为损失函数,并加入了循环一致性损失解决数据不匹配的问题.在颅脑磁共振成像数据集TCGA-GBM与CPTAC-GBM上进行一系列实验,验证了所提方法在图像生成任务上的有效性;该方法的MAE, PSNR和SSIM分别达到86.3, 25.96 dB和0.897;与对比方法相比,该方法也表现出优越的性能. 展开更多
关键词 双向transformer 伪计算机断层扫描 循环一致性 颅脑磁共振成像 U型网络
在线阅读 下载PDF
基于多编码器及双路径Transformer网络的单通道盲分离 被引量:2
4
作者 陈柄旭 高勇 《通信技术》 2025年第9期950-954,共5页
针对通信混合信号的单通道盲分离问题,提出了一种基于多编码器及双路径Transformer网络的新型分离方法。首先,为了在多个编码空间学习源混合信号的特征,设计了两个编码器,分别提取信号的波形特征与调制信息;其次,构建双路径Transformer... 针对通信混合信号的单通道盲分离问题,提出了一种基于多编码器及双路径Transformer网络的新型分离方法。首先,为了在多个编码空间学习源混合信号的特征,设计了两个编码器,分别提取信号的波形特征与调制信息;其次,构建双路径Transformer模块,通过局部路径和全局路径以直接上下文感知方式对信号序列进行建模;最后,利用双向循环神经网络充分捕捉序列数据中的长时依赖关系和特征,并用解码器重构源信号。仿真结果表明,所提出的网络可以有效地对两路混合信号进行分离,与卷积分离网络相比有约1 dB的性能提升。 展开更多
关键词 单通道盲分离 transformer 多编码器 双向门控循环单元
在线阅读 下载PDF
MS-SwinCE:融合多尺度Swin Transformer与对比特征增强的遥感图像变化检测
5
作者 朱永进 李熹 +1 位作者 李永胜 苏湘粤 《计算机系统应用》 2025年第12期67-74,共8页
遥感图像变化检测在城市扩张与灾害监测等领域中具有重要意义,然而,现有方法在特征提取能力、抗伪变化干扰以及多尺度融合方面仍存在不足之处.本文提出一种融合多尺度Swin Transformer、对比特征增强模块(contrastive feature enhanceme... 遥感图像变化检测在城市扩张与灾害监测等领域中具有重要意义,然而,现有方法在特征提取能力、抗伪变化干扰以及多尺度融合方面仍存在不足之处.本文提出一种融合多尺度Swin Transformer、对比特征增强模块(contrastive feature enhancement module, CFEM)与双向特征金字塔网络(bi-directional feature pyramid network,BiFPN)的变化检测模型MS-SwinCE.该模型利用局部窗口和移位机制增强长程依赖建模能力, CFEM精准提取变化差异并抑制噪声, BiFPN实现多尺度语义信息的高效融合.实验结果表明, MS-SwinCE在LEVIR-CD数据集上,相较于ChangeFormer, IoU提升了1.18%, F1分数提升了0.70%, Precision提升了0.32%, Recall提升了1.06%;在WHU-CD数据集上,相较于BIT, IoU提升了1.84%, F1分数提升了1.06%, Precision提升了0.47%, Recall提升了1.66%.此外,在保持较高精度的同时,模型参数量为31.66M,明显低于精度相近的ChangeFormer(41.03M),在精度与效率间实现了良好权衡.消融实验进一步验证了各模块的有效性与协同增益. 展开更多
关键词 遥感图像变化检测 Swin transformer 对比特征增强模块 多尺度特征融合 双向特征金字塔网络
在线阅读 下载PDF
基于NRBO-Transformer-BiLSTM的柔性薄壁轴承细粒度故障诊断
6
作者 郭明军 陈昕昀 +2 位作者 石淇 李鑫 赵学智 《现代制造工程》 北大核心 2025年第11期136-145,34,共11页
针对现有柔性薄壁轴承的细粒度故障识别及其现有诊断方法特征提取不足等问题,提出一种基于牛顿-拉夫逊优化(Newton-Raphson-Based Optimizer,NRBO)算法优化的Transformer编码层与双向长短期记忆(Bi-directional Long Short-Term Memory,... 针对现有柔性薄壁轴承的细粒度故障识别及其现有诊断方法特征提取不足等问题,提出一种基于牛顿-拉夫逊优化(Newton-Raphson-Based Optimizer,NRBO)算法优化的Transformer编码层与双向长短期记忆(Bi-directional Long Short-Term Memory,BiLSTM)神经网络解码层相结合的故障诊断方法。该方法首先使用时变滤波经验模态分解(Time-Varying Filter Empirical Mode Decomposition,TVFEMD)柔性薄壁轴承振动加速度信号,同时采用NRBO算法对其时变滤波带宽和B样条阶数等参数进行优化,依据互相关系数准则筛选主要本征模态分量(Intrinsic Mode Function,IMF),计算其时频域特征,结合振动加速度信号的时域、频域特征构建多域特征数据集,并按比例划分为训练集和测试集;其次将训练集输入模型,通过NRBO算法对模型初始学习率、BiLSTM神经网络的隐藏层节点数以及Transformer模型的正则化系数进行优化;通过测试集对优化模型进行测试,并与其他细粒度故障诊断模型对比。结果表明,所提方法准确率达99.60%,高于其他模型。该方法可为柔性薄壁轴承细粒度的智能诊断提供一种新的研究思路,对其他相关领域的智能化健康管理亦可提供有益借鉴。 展开更多
关键词 柔性薄壁轴承 故障诊断 牛顿-拉夫逊优化算法 双向长短期记忆网络 transformer模型 时变滤波经验模态分解
在线阅读 下载PDF
Cause Analysis and Preventive Measures of Over-standard Direct Resistance of Main Transformer
7
作者 SU Changbao YANG Gan +1 位作者 LI Qizhong ZHANG Zezhong 《外文科技期刊数据库(文摘版)工程技术》 2020年第1期043-044,共2页
Transformers are the core of substations. Once they fail, they will cause major losses. It may cause equipment failure at a slight scale, and a fire may be caused at a serious level, endangering normal operation safet... Transformers are the core of substations. Once they fail, they will cause major losses. It may cause equipment failure at a slight scale, and a fire may be caused at a serious level, endangering normal operation safety. Therefore, it is very important to analyze the causes of transformer failures and take corresponding preventive measures. In daily power transmission and supply, transformer DC resistance test is the main item of main transformer test and the main basis for comprehensive judgment of transformer failure. This article analyzes the cause of a main transformer direct resistance exceeding the standard and proposes some preventive measures. Later management provides basic guarantee. 展开更多
关键词 SUBSTATION transformer direct resistance FAULT preventive measures
原文传递
Wavelet transform and gradient direction based feature extraction method for off-line handwritten Tibetan letter recognition 被引量:3
8
作者 黄鹤鸣 达飞鹏 韩晓旭 《Journal of Southeast University(English Edition)》 EI CAS 2014年第1期27-31,共5页
To improve the recognition accuracy of off-line handwritten Tibetan characters the local gradient direction histograms based on the wavelet transform are proposed as the recognition features.First for a Tibetan charac... To improve the recognition accuracy of off-line handwritten Tibetan characters the local gradient direction histograms based on the wavelet transform are proposed as the recognition features.First for a Tibetan character sample image the first level approximation component of the Haar wavelet transform is calculated.Secondly the approximation component is partitioned into several equal-sized zones. Finally the gradient direction histograms of each zone are calculated and the local direction histograms of the approximation component are considered as the features of the character sample image.The proposed method is tested on the recently developed off-line Tibetan handwritten character sample database.The experimental results demonstrate the effectiveness and efficiency of the proposed feature extraction method.Furthermore compared with the detail components the approximation component contributes more to the recognition accuracy. 展开更多
关键词 pattern recognition wavelet transform gradient direction TIBETAN handwritten character
在线阅读 下载PDF
RESEARCH OF PROBLEMS ON REALIZING DIRECT ALGORITHM OF WAVELET TRANSFORM 被引量:2
9
作者 Tang BaopingZhong YoumingQin ShurenDepartment of Mechanical Engineering,Chongqing University,Chongqing 400044, China 《Chinese Journal of Mechanical Engineering》 SCIE EI CAS CSCD 2003年第2期136-140,共5页
Direct algorithm of wavelet transform (WT) is the numerical algorithmobtained from the integral formula of WT by directly digitization. Some problems on realizing thealgorithm are studied. Some conclusions on the dire... Direct algorithm of wavelet transform (WT) is the numerical algorithmobtained from the integral formula of WT by directly digitization. Some problems on realizing thealgorithm are studied. Some conclusions on the direct algorithm of discrete wavelet transform (DWT),such as discrete convolution operation formula of wavelet coefficients and wavelet components,sampling principle and technology to wavelets, deciding method for scale range of wavelets, measuresto solve edge effect problem, etc, are obtained. The realization of direct algorithm of continuouswavelet transform (CWT) is also studied. The computing cost of direct algorithm and Mallat algorithmof DWT are still studied, and the computing formulae are obtained. These works are beneficial todeeply understand WT and Mallat algorithm. Examples in the end show that direct algorithm can alsobe applied widely. 展开更多
关键词 wavelet transform direct algorithm mallat algorithm
在线阅读 下载PDF
Directional Filter for SAR Images Based on Nonsubsampled Contourlet Transform and Immune Clonal Selection 被引量:3
10
作者 Xiao-Hui Yang Li-Cheng Jiao Deng-Feng Li 《International Journal of Automation and computing》 EI 2009年第3期245-253,共9页
A directional filter algorithm for intensity synthetic aperture radar (SAR) image based on nonsubsampled contourlet transform (NSCT) and immune clonal selection (ICS) is presented. The proposed filter mainly foc... A directional filter algorithm for intensity synthetic aperture radar (SAR) image based on nonsubsampled contourlet transform (NSCT) and immune clonal selection (ICS) is presented. The proposed filter mainly focuses on exploiting different features of edges and noises by NSCT. Furthermore, ICS strategy is introduced to optimize threshold parameter and amplify parameter adaptively. Numerical experiments on real SAR images show that there are improvements in both visual effects and objective indexes. 展开更多
关键词 directional filter nonsubsampled contourlet transform (NSCT) immune clonal selection optimization (ICSO) syntheticaperture radar (SAR).
在线阅读 下载PDF
Three-dimensional transformation of magnetization direction and magnetic field component at low latitudes based on vertical relationship 被引量:6
11
作者 He Tao Xiong Sheng-Qing Wang Wan-Yin 《Applied Geophysics》 SCIE CSCD 2022年第1期91-106,145,共17页
The transformation of the magnetization direction and the magnetic fi eld component is one of the important methods in magnetic data processing and transformation,which can be conducted in both wavenumber and spatial ... The transformation of the magnetization direction and the magnetic fi eld component is one of the important methods in magnetic data processing and transformation,which can be conducted in both wavenumber and spatial domains.The transformation method in the wavenumber domain has simpler processing expression and higher processing effi ciency than in the spatial domain;however,they are unstable at low latitude.In this paper,the conclusion that the sum is 0 of two vertical magnetic fi eld components(magnetization inclinations are also perpendicular)in 2D is used for the 3D transformation of the magnetization direction and the magnetic field component.In addition,the transformation method at low latitudes based on vertical relationship(VMT)is proposed,which is an iterative algorithm that converts the transformation of the magnetization direction and the magnetic field component at the low latitude into the high latitude.This method restrains the instability of transformation of constant and variable magnetization direction and magnetic fi eld components in low latitudes.The accuracy,stability,and practicality are verifi ed from synthetic models and real data. 展开更多
关键词 the transformation of magnetization direction and magnetic field component wavenumber domain low latitudes vertical relationship iterative algorithm.
在线阅读 下载PDF
Study of wavelet transform type high-current transformer 被引量:2
12
作者 卢文科 朱长纯 +1 位作者 刘君华 张建军 《Journal of Coal Science & Engineering(China)》 2002年第2期75-79,共5页
The wavelet transformation is applied to the high current transformer.The high current transformer elaborated in the paper is mainly applied to the measurement of AC/DC high current.The principle of the transformer is... The wavelet transformation is applied to the high current transformer.The high current transformer elaborated in the paper is mainly applied to the measurement of AC/DC high current.The principle of the transformer is the Hall direct measurement principle.The transformer has the following three characteristics:firstly, the effect of the remnant field of the iron core on the measurement is decreased;secondly,because the temperature compensation is adopted,the transformer has good temperature charactreristic;thirdly,be cause the wavelet transfomation technology is adopted,the transformer has the capacity of good antijanming. 展开更多
关键词 wavelet transformation high current SENSOR Hall direct measurement transformer
在线阅读 下载PDF
Microstructures and phase transformation in directionally solidified TiAl-Nb alloys 被引量:2
13
作者 Fu-qiang Zhang Xian-fei Ding +6 位作者 Hai Nan Ru-yue You Qiang Wang Jian-ping He Yan-qing Su Yong-feng Liang Jun-pin Lin 《China Foundry》 SCIE CAS 2020年第6期402-408,共7页
Intermetallic Ti-45Al-8Nb-(W,B,Y)(at.%)and Ti-46Al-5Nb alloys are directionally solidified at a constant growth rate of 30μm·s-1 using a Bridgman type apparatus.The quenched microstructures and lengths of differ... Intermetallic Ti-45Al-8Nb-(W,B,Y)(at.%)and Ti-46Al-5Nb alloys are directionally solidified at a constant growth rate of 30μm·s-1 using a Bridgman type apparatus.The quenched microstructures and lengths of different phase regions were observed and measured after various growing times of 0-30 min.Results show that the phase transformations in different phase regions are mainly depending on the high temperature microstructure and the supercooling degree during quenching process.After isothermal holding,the primary phase grows into the liquid phase,the dendrites change from equiaxed to columnar grains,and the length of the L+βphase region,L+β+αphase region and mushy zone varies,indicating that the entire directional solidification process can be described by a static equilibrium-nonequilibrium-dynamic equilibrium evolution process.In addition,the gap between the original growth interface and front interface shows that the actual crystal growth rate is not equal to the drawing velocity during directional solidification. 展开更多
关键词 TiAl alloys directional solidification quenched microstructure phase transformation
在线阅读 下载PDF
基于多层多向Transformer的红外弱小目标检测 被引量:3
14
作者 王潇 刘贞报 《航空学报》 EI CAS CSCD 北大核心 2024年第14期1-13,共13页
针对基于卷积神经网络的红外图像弱小目标检测方法面临卷积神经网络的感受野有限,扩展感受野的下采样操作容易导致特征丢失,以及卷积网络对局部对比特征提取能力有限的问题,提出了一种基于多层多向Transformer的红外图像弱小目标检测算... 针对基于卷积神经网络的红外图像弱小目标检测方法面临卷积神经网络的感受野有限,扩展感受野的下采样操作容易导致特征丢失,以及卷积网络对局部对比特征提取能力有限的问题,提出了一种基于多层多向Transformer的红外图像弱小目标检测算法。首先使用感受野较大、对比特征提取能力较强的Transformer网络作为基本单元,设计U型深度神经网络和多特征层融合网络将局部特征以及全局特征进行融合。同时在解码网络设计双向注意力算子,利用自注意力计算机制分别计算空间方向和特征方向的注意力特征,进一步提高深度网络提取弱小目标和周边区域对比特征的能力。另外在主干网络最后添加数量约束网络,通过统计对比检测结果中目标的数量,减小误检目标的数目,提高目标检测的准确度。最后在多个实验数据集上和已有方法进行了对比实验,在各个评价参数上取得了最多35%的提升,证明所提红外弱小目标检测算法具有较高的准确性。 展开更多
关键词 红外弱小目标检测 transformer 多特征层融合 双向注意力算子 数量约束
原文传递
基于Transformer和关系图卷积网络的信息传播预测模型 被引量:2
15
作者 吕锡婷 赵敬华 +1 位作者 荣海迎 赵嘉乐 《计算机应用》 CSCD 北大核心 2024年第6期1760-1766,共7页
针对在信息传播动态演化中,结构特征和时序特征以及两者间的交互表达难以有效捕获的问题,提出一种基于Transformer和关系图卷积网络的信息传播预测模型(TRGCN)。首先,构建由社交关系图和传播级联图组合而成的异构图,使用关系图卷积网络(... 针对在信息传播动态演化中,结构特征和时序特征以及两者间的交互表达难以有效捕获的问题,提出一种基于Transformer和关系图卷积网络的信息传播预测模型(TRGCN)。首先,构建由社交关系图和传播级联图组合而成的异构图,使用关系图卷积网络(RGCN)提取图中各节点的结构特征;其次,使用双向长短期记忆(Bi-LSTM)网络对各节点的时间嵌入重新编码,引入时间衰减项以不同的权重赋予不同时间位置的节点,获得节点的时序特征;最后,将结构特征和时序特征输入Transformer进行融合,得到时空特征以预测信息传播。在Twitter、Douban和Memetracker这3个真实数据集上的实验结果表明,相较于对比实验中的最优模型,TRGCN的Hits@100指标分别提升3.18%,5.96%和3.34%,Map@100指标分别提升11.60%,19.72%和8.47%,验证了所提模型的有效性和合理性。 展开更多
关键词 信息传播预测 transformer 关系图卷积网络 双向长短期记忆网络 时空特征
在线阅读 下载PDF
Evaluation of the Direct UTM Coordinates Transformation Method Based on the Standard 7-Parameters Transformation 被引量:1
16
作者 Mohammed Anwer Jassim Darin Mohammed Tofiq Mohammed 《Journal of Civil Engineering and Architecture》 2022年第8期413-419,共7页
GNSS(global navigation satellite system)observations produce the geodetic position including latitude,longitude,and altitude(or ellipsoidal height)concerning the global reference datum WGS84(Word Geodetic System 1984)... GNSS(global navigation satellite system)observations produce the geodetic position including latitude,longitude,and altitude(or ellipsoidal height)concerning the global reference datum WGS84(Word Geodetic System 1984),which usually should be converted to another local datum to get the desired position meaning in a physical sense,coordinates of points in the local datum are usually calculated by the seven-parameter transformation method.This paper aims to validate the methods of position transformation between WGS84 and the Iraqi local datum Karbala 1979 using the UTM(universal transverse Mercator)projected coordinates directly.The proposed algorithm was tested for 10 ground control points in Erbil city and many selected points in other different cities over all Iraqi territory.The control points are measured by the CHCNAV i73 GNSS receiver.For the evaluation procedure,the RMSE(root mean square error)of the transformed coordinates is calculated with an average value of±10.715 m as an estimated uncertainty of the direct UTM coordinates transformation method over Erbil city territory,and more than±12 m over different places over Iraqi territory. 展开更多
关键词 direct UTM coordinates transformation global datum WGS84 local Iraqi datum Karbala 1979
在线阅读 下载PDF
基于上下文知识增强型Transformer网络的抑郁检测 被引量:4
17
作者 张亚洲 和玉 +1 位作者 戎璐 王祥凯 《计算机工程》 CAS CSCD 北大核心 2024年第8期75-85,共11页
抑郁症作为一种常见的心理健康问题,严重影响人们的日常生活甚至是生命安全。鉴于目前的抑郁症检测存在主观性和人工干预等缺点,基于深度学习的自动检测方式成为热门研究方向。对于最易获取的文本模态而言,主要的挑战在于如何建模抑郁... 抑郁症作为一种常见的心理健康问题,严重影响人们的日常生活甚至是生命安全。鉴于目前的抑郁症检测存在主观性和人工干预等缺点,基于深度学习的自动检测方式成为热门研究方向。对于最易获取的文本模态而言,主要的挑战在于如何建模抑郁文本中的长距离依赖与序列依赖。为解决该问题,提出一种基于上下文知识的增强型Transformer网络模型RoBERTa-BiLSTM,旨在从抑郁文本序列中充分提取和利用上下文特征。结合序列模型与Transformer模型优点,建模单词间上下文交互,为抑郁类别揭示与信息表征提供参考。首先,利用RoBERTa方法将词汇嵌入到语义向量空间;其次,利用双向长短期记忆网络(BiLSTM)模型有效捕获长距离上下文语义;最后,在DAIC-WOZ和EATD-Corpus 2个大规模数据集上进行实证研究。实验结果显示,RoBERTa-BiLSTM模型的准确率分别达到0.74和0.93以上,召回率分别达到0.66和0.56以上,能够准确地检测抑郁症。 展开更多
关键词 抑郁检测 序列模型 深度学习 transformer模型 双向长短期记忆模型
在线阅读 下载PDF
Identification of Sources of Potential Fields with Directional Wavelet Transform:Application to Mineral Exploration
18
作者 Dianhua Cao Anjian Wang 《地学前缘》 EI CAS CSCD 北大核心 2009年第S1期146-146,共1页
For the sustainable supply of mineral resources,blind deposits are becoming the emphasis of exploration after long-period exploitation of exposed deposits.The collection and analysis of gravity or magnetic data repres... For the sustainable supply of mineral resources,blind deposits are becoming the emphasis of exploration after long-period exploitation of exposed deposits.The collection and analysis of gravity or magnetic data represents one of the cheapest forms of large-scale geophysical exploration.With the identification of potential fields,we can get the map of worms or skeletonizations showing the three-dimension structure of shallow crust. 展开更多
关键词 potential field directional wavelet transform multidirectionaland multiscale edges detection mineral exploration
在线阅读 下载PDF
A DECOMPOSITION DEPENDED ON DIRECTIONS FOR NONSINGULAR LINEAR TRANSFORMATION
19
作者 张慎学 《Applied Mathematics and Mechanics(English Edition)》 SCIE EI 1996年第8期789-794,共6页
In this paper. we give a decomposition depending on p(1≤p≤n-2) orthonormaldirections assigned for nonsingular linear transformation F on a n-dimension (n≥3)Euclidean space En, and then prove foal there exist q(q=n-... In this paper. we give a decomposition depending on p(1≤p≤n-2) orthonormaldirections assigned for nonsingular linear transformation F on a n-dimension (n≥3)Euclidean space En, and then prove foal there exist q(q=n-p) quasi-Principaldirections.for F depending on the preceding p orthonormal directions. As applicance ofthe preceding result, we derive that there exist at least two orthonormal principaldirections of strain in arbitrary plane of body which is in homogeneous deformation,and strain energy density is.function of 5 real numbers under arbitrary quasi-principalbase.for the preceding nonsingular linear transformation. 展开更多
关键词 NONSINGULAR linear transformation quasi-principal directions.Decomposition
在线阅读 下载PDF
Identification of Potential Mineral Fields with Directional Wavelet Transform:Application to Mineral Exploration
20
作者 Dianhua Cao Anjian Wang 《地学前缘》 EI CAS CSCD 北大核心 2009年第S1期171-172,共2页
For the sustainable supply of mineral resources,blind deposits are becoming the emphasis of exploration after long-period exploitation of exposed deposits.The collection and analysis of gravity or magnetic data repres... For the sustainable supply of mineral resources,blind deposits are becoming the emphasis of exploration after long-period exploitation of exposed deposits.The collection and analysis of gravity or magnetic data represents one of the cheapest ways of large-scale geophysical exploration. 展开更多
关键词 potential field directional wavelet transform multidirectional and multiscale edges detection mineral exploration
在线阅读 下载PDF
上一页 1 2 105 下一页 到第
使用帮助 返回顶部