Results 81 to 90 of about 12,984 (168)
A DC-UNet-based image processing method for detecting fractures along roadway sections of coal mines [PDF]
ObjectiveDetecting fractures along roadway sections of coal mines allows for the characterization of the geologic conditions of roadways, thus providing guidance for roadway tunneling and support engineering.
Chenhui TIAN +5 more
core +1 more source
Unsupervised intrusion detection model based on temporal convolutional network [PDF]
Most existing intrusion detection models rely on long short-term memory (LSTM) networks to consider time-dependencies among data. However, LSTM’s sequential data processing significantly increases computational complexity and memory consumption during ...
DING Jiawei, FENG Guanghui, LIAO Jinju
core +1 more source
[Electrocardiogram signal classification based on fusion method of residual network and self-attention mechanism]. [PDF]
Sheng Wu Yi Xue Gong Cheng Xue Za Zhi, 2023 Yuan C, Liu Z, Wang C, Yang F.
europepmc +1 more source
智能科学与技术学报针对数字货币量化交易中存在的问题，即大量且复杂因子以及因子状态空间维度较高，导致交易模型制定策略的准确性和风险控制能力难以兼顾，提出了一种改进的OAC模型——OAC_LSTM_ATT。该模型采用了LSTM和多头注意力机制来优化OAC的网络结构，从而提高OAC对时间序列数据的建模能力和泛化能力。通过这种融合，智能体在量化交易环境中可以更加灵活和准确地做出交易决策，进一步提高交易策略的质量和效果。实验结果显示，在比特币市场中，累计收益率达到了16.36%，最大回撤率为9.08%，夏普比为0.014 ...
许波, 贺一峻, 李祥霞
doaj +1 more source
[Whole-brain parcellation for macaque brain magnetic resonance images based on attention mechanism and multi-modality feature fusion]. [PDF]
Nan Fang Yi Ke Da Xue Xue Bao, 2023 Wu X, Zhang Y, Zhang H, Zhong T.
europepmc +1 more source
[Medical nucleus image segmentation network based on convolution and attention mechanism]. [PDF]
Sheng Wu Yi Xue Gong Cheng Xue Za Zhi, 2022 Zhi P, Deng J, Zhong Z.
europepmc +1 more source
Vehicle trajectory prediction based on spatio-temporal Transformer feature fusion [PDF]
In complex traffic environments, autonomous vehicles must thoroughly analyze the motion direction, speed, and other information of surrounding traffic objects to accurately predict future trajectories. A network model based on spatio-temporal Transformer
WAN Zilu, WANG Wei, ZHAO Wenhong
core +1 more source
Gaoya dianqi基于无人机航拍的电力巡检成为目前绝缘子缺陷检测方法的主流，但当遇到图像特征不够明显或干扰特征较多等问题时，绝缘子缺陷识别困难，检测精度不高。由此，提出了一种基于改进YOLOv5s的绝缘子缺陷检测方法。首先，重新设计卷积模块，然后将CA注意力机制与其相融合，并且在主干网络加入注意力机制与颈部网络的特征图进行多尺度特征融合，抑制复杂环境下的干扰特征，专注缺陷特征提取；其次，对空间金字塔池化结构（SPPF）进行改进，扩大感受野，减少被模型过滤掉的有用信息；接着，将Transformer与C3模块中的残差结构 ...
彭晏飞, 袁晓龙, 赵涛, 陈炎康
doaj
Dianxin kexue现有基于深度学习的调制识别在训练阶段需要大量IQ信号样本，而复杂电磁环境中很难获取大量样本，导致基于深度学习的调制识别算法泛化性能下降。针对网络泛化能力差的问题，提出了一种基于信号增强的调制识别（signal enhancement based modulation recognition，SEBMR）算法。首先，设计了捕获IQ信号全局特征的特征提取及重构模块；其次，提出了基于辅助分类生成对抗网络（auxiliary classifier generative adversarial network ...
程风云, 周金
doaj +1 more source
Gaoya dianqi, 2022 为提高换流站智能化水平，充分利用巡检机器人硬件能力，文中基于变压器声音信息提出一种直流偏磁声纹识别方法，该方法可以无需特定降噪算法应对瞬态和稳态噪声。首先，对变压器声音信号进行分析；其次，为提高有效声音信息权重，结合变压器声音信号特点，使用W-50FMCC特征来表征声音信息；再次，基于多头注意力机制和残差结构设计了非自回归端到端偏磁声纹识别模型，使用信道补偿算法提升特征正交性并获取相似度得分，完成识别；最后通过实验进行了验证。实验表明，该方法可以直接对变压器声学信号的偏磁情况进行准确的识别，无需降噪算法。
刘建华 +5 more
doaj