排序方式: 共有23条查询结果,搜索用时 15 毫秒
1.
恶意服务常利用域名生成算法(DGA)逃避域名检测,针对DGA域名隐蔽性强、现有检测方法检测速度较慢、实用性不强等问题,采用深度学习技术,提出了一种基于Deep-IndRNN的DGA域名检测方法。方法运用词袋模型(BoW)将域名向量化,然后通过Deep-IndRNN提取域名字符间特征,并使用Sigmoid函数对域名分类检测。其主要特点在于:通过将Deep-IndRNN的多序列输入拼接为单向量输入,以单步处理代替循环处理,同时结合Deep-IndRNN能保存更长时间记忆的特点,可有效释放深度学习时占用的GPU、CPU等系统资源,且在保证高准确率和精确度的前提下提高训练、检测速度。实验结果表明,基于Deep-IndRNN的DGA域名检测方法在检测任务中具有较高的准确率和精确度,相比于DNN、CNN、LSTM、BiLSTM、CNN-LSTM-Concat等同类检测方法,能显著提高训练、检测速度,是有效可行的。 相似文献
2.
现有的焦距检测方法通常由于检测仪器光源波长与光学系统不完全匹配从而产生纵向色差影响检测结果。针对这一问题,研究光学系统纵向色差的变化规律,并确定在400 nm~1 000 nm波段用于表示其函数关系的Conrady公式和复消色差特性公式。根据光学系统近焦位置的离焦量与位置呈线性关系的特性, 提出使用菲索干涉仪测量5种不同波长的焦距位置,获得单透镜和双胶合镜头的纵向色差曲线。实验结果表明: 在400 nm~1 000 nm波段单色系统和消色差系统的纵向色差的函数关系分别符合Conrady公式和复消色差特性公式,研究结果为焦距的理论计算和精确检测提供了新的思路和参考。 相似文献
3.
4.
5.
6.
对基于工业CT图像重构的网格模型进行网格简化时,大多数现有网格模型简化算法会丢失特征,出现网格质量不好的问题。因此提出一种网格模型保特征简化方法,该方法用三角形折叠法对原始模型进行简化,当简化后模型的平均二面角角度误差达到允许误差后,再使用边折叠法对模型进行简化。在三角形折叠法中提出了利用被折叠三角形的法向量、各个顶点的高斯曲率及其在周边三角形上的投影确定该三角形的折叠点,利用局部体积误差与二面角角度误差的无因次化和确定折叠代价的方法;在边折叠法中提出了将二面角角度误差引入到二次误差测度(QEM)法的折叠代价中的改进QEM法。实验结果表明:与其他算法相比,该方法能够生成保特征、高质量、低几何误差的网格模型。 相似文献
7.
<正>With the development of the compressive sensing theory,the image reconstruction from the projections viewed in limited angles is one of the hot problems in the research of computed tomography technology.This paper develops an iterative algorithm for image reconstruction,which can fit most cases.This method gives an image reconstruction flow with the difference image vector,which is based on the concept that the difference image vector between the reconstructed and the reference image is sparse enough.Then the l_2-norm minimization method is used to reconstruct the difference vector to recover the image for flat subjects in limited angles.The algorithm has been tested with a thin planar phantom and a real object in limited-view projection data.Moreover,all the studies showed the satisfactory results in accuracy at a rather high reconstruction speed. 相似文献
8.
9.
10.
Feasibility of similarity coefficient map for improving morphological evaluation of T2* weighted MRI for renal cancer 下载免费PDF全文
The purpose of this paper is to investigate the feasibility of using a similarity coefficient map(SCM) in improving the morphological evaluation of T2* weighted(T2*W) magnatic resonance imaging(MRI) for renal cancer.Simulation studies and in vivo 12-echo T2*W experiments for renal cancers were performed for this purpose.The results of the first simulation study suggest that an SCM can reveal small structures which are hard to distinguish from the background tissue in T2*W images and the corresponding T2* map.The capability of improving the morphological evaluation is likely due to the improvement in the signal-to-noise ratio(SNR) and the carrier-to-noise ratio(CNR) by using the SCM technique.Compared with T2* W images,an SCM can improve the SNR by a factor ranging from 1.87 to 2.47.Compared with T2* maps,an SCM can improve the SNR by a factor ranging from 3.85 to 33.31.Compared with T2*W images,an SCM can improve the CNR by a factor ranging from 2.09 to 2.43.Compared with T2* maps,an SCM can improve the CNR by a factor ranging from 1.94 to 8.14.For a given noise level,the improvements of the SNR and the CNR depend mainly on the original SNRs and CNRs in T2*W images,respectively.In vivo experiments confirmed the results of the first simulation study.The results of the second simulation study suggest that more echoes are used to generate the SCM,and higher SNRs and CNRs can be achieved in SCMs.In conclusion,an SCM can provide improved morphological evaluation of T2*W MR images for renal cancer by unveiling fine structures which are ambiguous or invisible in the corresponding T2*W MR images and T2* maps.Furthermore,in practical applications,for a fixed total sampling time,one should increase the number of echoes as much as possible to achieve SCMs with better SNRs and CNRs. 相似文献