A novel approach for improving antenna bandwidth is described using a 6-element Yagi-Uda array as an example. The new approach applies Central Force Optimization, a deterministic metaheuristic, and Variable Z0 technol...A novel approach for improving antenna bandwidth is described using a 6-element Yagi-Uda array as an example. The new approach applies Central Force Optimization, a deterministic metaheuristic, and Variable Z0 technology, a novel, proprietary design and optimization methodology, to produce an array with 33.09% fractional impedance bandwidth. This array’s performance is compared to its CFO-optimized Fixed Z0counterpart, and to the performance of a 6-ele- ment Dominating Cone Line Search-optimized array. Both CFO-optimized antennas exhibit better performance than the DCLS array, especially with respect to impedance bandwidth. Although the Yagi-Uda antenna was chosen to illustrate this new approach to antenna design and optimization, the methodology is entirely general and can be applied to any antenna against any set of performance objectives.展开更多
This paper investigates the effect of adding three extensions to Central Force Optimization when it is used as the Global Search and Optimization method for the design and optimization of 6-elementYagi-Uda arrays. Tho...This paper investigates the effect of adding three extensions to Central Force Optimization when it is used as the Global Search and Optimization method for the design and optimization of 6-elementYagi-Uda arrays. Those exten</span><span><span style="font-family:Verdana;">sions are </span><i><span style="font-family:Verdana;">Negative</span></i> <i><span style="font-family:Verdana;">Gravity</span></i><span style="font-family:Verdana;">, </span><i><span style="font-family:Verdana;">Elitism</span></i><span style="font-family:Verdana;">, and </span><i><span style="font-family:Verdana;">Dynamic</span></i> <i><span style="font-family:Verdana;">Threshold</span></i> <i><span style="font-family:Verdana;">Optimization</span></i><span style="font-family:Verdana;">. T</span></span><span style="font-family:Verdana;">he basic CFO heuristic does not include any of these, but adding them substan</span><span style="font-family:Verdana;">tially improves the algorithm’s performance. This paper extends the work r</span><span style="font-family:Verdana;">eported in a previous paper that considered only negative gravity and which </span><span style="font-family:Verdana;">showed a significant performance improvement over a range of optimized a</span><span style="font-family:Verdana;">rrays. Still better results are obtained by adding to the mix </span><i><span style="font-family:Verdana;">Elitism</span></i><span style="font-family:Verdana;"> and </span><i><span style="font-family:Verdana;">DTO</span></i><span style="font-family:Verdana;">. An overall improvement in best fitness of 19.16% is achieved by doing so. While the work reported here was limited to the design/optimization of 6-</span></span></span><span><span><span style="font-family:""> </span></span></span><span><span><span style="font-family:""><span style="font-family:Verdana;">element Yagis, the reasonable inference based on these data is that any antenna design/optimization problem, indeed any Global Search and Optimiza</span><span style="font-family:Verdana;">tion problem, antenna or not, utilizing Central Force Optimization as the Gl</span><span style="font-family:Verdana;">obal Search and Optimization engine will benefit by including all three extensions, probably substantially.展开更多
不同成像模式设备采集的医学图像存在不同程度的分布差异,无监督域自适应方法为了将源域训练的模型泛化到无标注的目标域,通常是将差异分布最小化,使用源域和目标域的共有特征进行结果预测,但会忽略目标域的私有特征.为了解决该问题,文...不同成像模式设备采集的医学图像存在不同程度的分布差异,无监督域自适应方法为了将源域训练的模型泛化到无标注的目标域,通常是将差异分布最小化,使用源域和目标域的共有特征进行结果预测,但会忽略目标域的私有特征.为了解决该问题,文中提出基于目标域增强表示的医学图像无监督跨域分割方法(Enhanced Target Domain Representation Based Unsupervised Cross-Domain Medical Image Segmentation,TreUCMIS).首先,通过共有特征学习获取源域和目标域的共有特征,通过图像重构训练目标域特征编码器,提取目标域完整特征.然后,通过目标域的无监督自学习方式,加强深层特征和浅层特征的共有性.最后,对齐使用共有特征和完整特征得到的预测结果,利用目标域的完整特征分割目标,提高模型在目标域的泛化性.在两个具有CT和MRI双向域自适应任务的医学图像分割数据集(腹部、心脏)上的实验表明TreUCMIS的有效性与优越性.展开更多
文摘A novel approach for improving antenna bandwidth is described using a 6-element Yagi-Uda array as an example. The new approach applies Central Force Optimization, a deterministic metaheuristic, and Variable Z0 technology, a novel, proprietary design and optimization methodology, to produce an array with 33.09% fractional impedance bandwidth. This array’s performance is compared to its CFO-optimized Fixed Z0counterpart, and to the performance of a 6-ele- ment Dominating Cone Line Search-optimized array. Both CFO-optimized antennas exhibit better performance than the DCLS array, especially with respect to impedance bandwidth. Although the Yagi-Uda antenna was chosen to illustrate this new approach to antenna design and optimization, the methodology is entirely general and can be applied to any antenna against any set of performance objectives.
文摘This paper investigates the effect of adding three extensions to Central Force Optimization when it is used as the Global Search and Optimization method for the design and optimization of 6-elementYagi-Uda arrays. Those exten</span><span><span style="font-family:Verdana;">sions are </span><i><span style="font-family:Verdana;">Negative</span></i> <i><span style="font-family:Verdana;">Gravity</span></i><span style="font-family:Verdana;">, </span><i><span style="font-family:Verdana;">Elitism</span></i><span style="font-family:Verdana;">, and </span><i><span style="font-family:Verdana;">Dynamic</span></i> <i><span style="font-family:Verdana;">Threshold</span></i> <i><span style="font-family:Verdana;">Optimization</span></i><span style="font-family:Verdana;">. T</span></span><span style="font-family:Verdana;">he basic CFO heuristic does not include any of these, but adding them substan</span><span style="font-family:Verdana;">tially improves the algorithm’s performance. This paper extends the work r</span><span style="font-family:Verdana;">eported in a previous paper that considered only negative gravity and which </span><span style="font-family:Verdana;">showed a significant performance improvement over a range of optimized a</span><span style="font-family:Verdana;">rrays. Still better results are obtained by adding to the mix </span><i><span style="font-family:Verdana;">Elitism</span></i><span style="font-family:Verdana;"> and </span><i><span style="font-family:Verdana;">DTO</span></i><span style="font-family:Verdana;">. An overall improvement in best fitness of 19.16% is achieved by doing so. While the work reported here was limited to the design/optimization of 6-</span></span></span><span><span><span style="font-family:""> </span></span></span><span><span><span style="font-family:""><span style="font-family:Verdana;">element Yagis, the reasonable inference based on these data is that any antenna design/optimization problem, indeed any Global Search and Optimiza</span><span style="font-family:Verdana;">tion problem, antenna or not, utilizing Central Force Optimization as the Gl</span><span style="font-family:Verdana;">obal Search and Optimization engine will benefit by including all three extensions, probably substantially.
文摘不同成像模式设备采集的医学图像存在不同程度的分布差异,无监督域自适应方法为了将源域训练的模型泛化到无标注的目标域,通常是将差异分布最小化,使用源域和目标域的共有特征进行结果预测,但会忽略目标域的私有特征.为了解决该问题,文中提出基于目标域增强表示的医学图像无监督跨域分割方法(Enhanced Target Domain Representation Based Unsupervised Cross-Domain Medical Image Segmentation,TreUCMIS).首先,通过共有特征学习获取源域和目标域的共有特征,通过图像重构训练目标域特征编码器,提取目标域完整特征.然后,通过目标域的无监督自学习方式,加强深层特征和浅层特征的共有性.最后,对齐使用共有特征和完整特征得到的预测结果,利用目标域的完整特征分割目标,提高模型在目标域的泛化性.在两个具有CT和MRI双向域自适应任务的医学图像分割数据集(腹部、心脏)上的实验表明TreUCMIS的有效性与优越性.
文摘具有混合记忆的自步对比学习(Self-paced Contrastive Learning,SpCL)通过集群聚类生成不同级别的伪标签来训练网络,取得了较好的识别效果,然而该方法从源域和目标域中捕获的行人数据之间存在典型的分布差异,使得训练出的网络不能准确区别目标域和源域数据域特征。针对此问题,提出了双分支动态辅助对比学习(Dynamic Auxiliary Contrastive Learning,DACL)框架。该方法首先通过动态减小源域和目标域之间的局部最大平均差异(Local Maximum Mean Discrepancy,LMMD),以有效地学习目标域的域不变特征;其次,引入广义均值(Generalized Mean,GeM)池化策略,在特征提取后再进行特征聚合,使提出的网络能够自适应地聚合图像的重要特征;最后,在3个经典行人重识别数据集上进行了仿真实验,提出的DACL与性能次之的无监督域自适应行人重识别方法相比,mAP和rank-1在Market1501数据集上分别增加了6.0个百分点和2.2个百分点,在MSMT17数据集上分别增加了2.8个百分点和3.6个百分点,在Duke数据集上分别增加了1.7个百分点和2.1个百分点。
