Results 1 to 10 of about 16,430 (195)
基于显著性信息的Fit CutMix数据增强算法在医学影像上的应用
智能科学与技术学报, 2023 深度卷积神经网络是图像分类领域的主流算法之一，但是其训练需要大量标注数据，在阿尔茨海默病医学影像等小数据集上易出现过拟合现象。数据增强算法可以用于扩充数据量，其中 CutMix 数据增强算法近来被广泛应用，但是现有方法生成的增强图像往往忽略原始图像显著区域，且增强图像的标签设计考虑的因素较为单一。针对这些问题，提出Fit CutMix数据增强算法。该算法一是利用基于显著性极值迁移的区域替换策略生成增强样本，集中源样本与目标样本中显著性高的区域；二是综合源样本与目标样本的面积和显著性信息赋予增强样本标签 ...
罗欣欢, 王奕璇, 李炜, 陈曦
doaj +1 more source
智能科学与技术学报, 2022 交通态势感知是智能交通系统的重要研究方向。已有研究大多关注如何使用物理传感器感知当下交通态势并预测未来交通状况。然而，物理传感器性能易因天气影响、电磁干扰、能源限制等问题出现不稳定或失效情况，导致其采集的数据稀疏或缺失，使其对交通态势感知滞后且不准确。社会媒体数据为及时感知完善的交通态势信息提供了新的增强方式。面向当下异常交通情况频发的城市交通管控现状，社会传感与物理传感数据互为补充，可进一步满足城市交通高效管理需求。基于此，对基于社会媒体数据的交通事件检测和交通状况预测工作展开分析研究 ...
陈苑文, 王晓, 李灵犀, 王飞跃
doaj +1 more source
Radio frequency fingerprint data augmentation for indoor localization based on diffusion model [PDF]
, 2023 The radio frequency fingerprint indoor localization method ensures the accuracy by collecting a sufficient amount of fingerprints in the offline state to build a dense fingerprint database.A data augmentation method called FPDiffusion was proposed based ...
Hanxiao CHANG +4 more
core +1 more source
Dizhen xuebao, 2020 为了快速、高效地从地震数据中识别地震事件和拾取震相，本文利用基于样本增强的卷积神经网络自动震相拾取方法，将西藏林芝地区L0230台站3个月数据作为训练集，该区内另外6个台站连续1个月的波形数据作为测试集，采用高斯噪声、随机噪声拼接、随机挑选噪声、随机截取地震事件等4种样本增强的方法扩增训练集，以提高自动震相拾取技术的准确率。结果显示：样本增强前模型在测试集上的地震事件识别准确率为80%，样本增强后提升至97%，表明样本增强有效地提高了模型的泛化性能和抗干扰能力；在0.5 s误差范围内 ...
An Li +4 more
doaj +1 more source
Data augmentation scheme for federated learning with non-IID data [PDF]
, 2023 To solve the problem that the model accuracy remains low when the data are not independent and identically distributed (non-IID) across different clients in federated learning, a privacy-preserving data augmentation scheme was proposed.Firstly, a data ...
Di WANG, Lingtao TANG, Shengyun LIU
core +1 more source
Compression Algorithm of Face Recognition Model Based on Unlabeled Knowledge Distillation [PDF]
, 2022 When transplanting face recognition technology to mobile devices,it often needs to be processed by accelerated algorithms such as model compression.Knowledge distillation is a model compression method that has a wide range of practical applications and ...
CHENG Xiang-ming, DENG Chun-hua
core +1 more source
Composite Tor traffic features extraction method of webpage in actual network flow based on SDN [PDF]
, 2022 Website fingerprinting (WF) methods for Tor webpage traffic are often based on the separated Tor traffic or even the separated Tor webpage traffic.However, distinguishing Tor traffic from the original traffic of the actual network and Tor webpage traffic
Hongping YAN +5 more
core +1 more source
Research on intrusion detection for maritime meteorological sensor network based on balancing generative adversarial network [PDF]
, 2023 Aiming at the problem that the resources of maritime mobile terminals were limited and the network traffic was imbalanced in the MMSN (maritime meteorological sensor network) environment, which made it difficult to detect network intrusion accurately, a ...
Guifu ZHANG +3 more
core +1 more source
Kongjian kexue xuebao, 2023 利用GUVI探测器观测的OI 135.6 nm气辉辐射数据，研究了2003年10月和11月两次持续时间不同的强磁暴期间气辉辐射强度经纬度分布及纬度–高度分布的变化。结果表明，全球大部分区域内OI 135.6 nm气辉辐射强度在磁暴期间会随着Dst指数的降低而增强，与磁暴的剧烈程度呈现准正相关关系。从纬度上看，气辉辐射强度的增强会由磁静期的磁赤道两侧附近向高纬地区延伸至南北半球50°附近，增幅基本在80%以上，最大可超过200%，且经纬度分布在磁暴期间的变化情况呈现出南北不对称性和纬向不对称性；从高度上看，
李 春来, 吕 建永, 李 婧媛
doaj +1 more source
Adaptive Weight Based Broad Learning Algorithm for Cascaded Enhanced Nodes [PDF]
, 2022 In the era of intelligence,continuous autonomous learning and optimization need to be carried out on the big data platform,and the first step of continuous autonomous learning is data enhancement.This paper proposes a broad learning method based on ...
CAI Xin-yu, FENG Xiang, YU Hui-qun
core +1 more source