期刊文献+
共找到3篇文章
< 1 >
每页显示 20 50 100
基于坐标系人脸注册和集成分类器的3D人脸识别方法 被引量:1
1
作者 卜宇 刘俊霞 陈卫东 《西南师范大学学报(自然科学版)》 CAS 北大核心 2017年第10期106-114,共9页
由于三维人脸识别精度很大程度上取决于对齐模块的质量,提出一种内在坐标系注册人脸、两层集成分类器进行人脸识别的方法.首先进行3D人脸注册,人脸注册是基于由鼻尖、鼻子水平面以及人脸垂直对称平面表征的内在坐标系统,只需注册1次人... 由于三维人脸识别精度很大程度上取决于对齐模块的质量,提出一种内在坐标系注册人脸、两层集成分类器进行人脸识别的方法.首先进行3D人脸注册,人脸注册是基于由鼻尖、鼻子水平面以及人脸垂直对称平面表征的内在坐标系统,只需注册1次人脸扫描;然后进行预处理,去除噪声和填充空洞区域;最后通过主成分分析(PCA)提取图像特征,两层集成分类器采用马氏余弦(MahCos)匹配度和加权波达计数法(WBC)对人脸区域分类并进行合并和重新排列.FRGC v2.0和CurtinFaces数据库的实验结果验证了此方法的有效性.与Mesh shift方法、三维PCA方法、ICP匹配方法和LDA方法相比,此方法提高了识别性能,且在计算复杂度上也有较大优势. 展开更多
关键词 三维人脸识别 人脸注册 主成分分析 波达计算法 预处理
下载PDF
COMPUTER VISUALIZATION OF ELECTROMAGNETIC WAVE INTERACTION WITH STRUCTURES
2
作者 余雄庆 杨景佐 《Transactions of Nanjing University of Aeronautics and Astronautics》 EI 1995年第1期80-86,共7页
In this paper, a computer visualization approach is proposed for electromagnetic wave interaction with structures by mains of finite difference-time doain method (F-D) and computer graphics. By visualization of FDTD, ... In this paper, a computer visualization approach is proposed for electromagnetic wave interaction with structures by mains of finite difference-time doain method (F-D) and computer graphics. By visualization of FDTD, Phenomena such as wave propagation, penetration through structures, renection and absorption by structures are observed. Visualization of electromagnetic wave interactions with two wing-shaped structures is demonstrated. These examples indicate that the approach describe in the paper offers an effective way for investigating electromagnetic wave phenomena and is helpful to the engineers in controlling radar signature of the targets. 展开更多
关键词 electromagnetic wave scattering finite difference method radar targets computer visualization ANIMATION
下载PDF
Computationally Efficient 2DDOA Estimation for Uniform Planar Arrays:RDROOTMUSIC Algorithm 被引量:3
3
作者 YE Changbo ZHU Beizuo +1 位作者 LI Baobao ZHANG Xiaofei 《Transactions of Nanjing University of Aeronautics and Astronautics》 EI CSCD 2021年第4期685-694,共10页
The problem of two-dimensional direction of arrival(2D-DOA)estimation for uniform planar arrays(UPAs)is investigated by employing the reduced-dimensional(RD)polynomial root finding technique and 2D multiple signal cla... The problem of two-dimensional direction of arrival(2D-DOA)estimation for uniform planar arrays(UPAs)is investigated by employing the reduced-dimensional(RD)polynomial root finding technique and 2D multiple signal classification(2D-MUSIC)algorithm.Specifically,based on the relationship between the noise subspace and steering vectors,we first construct 2D root polynomial for 2D-DOA estimates and then prove that the 2D polynomial function has infinitely many solutions.In particular,we propose a computationally efficient algorithm,termed RD-ROOT-MUSIC algorithm,to obtain the true solutions corresponding to targets by RD technique,where the 2D root-finding problem is substituted by two one-dimensional(1D)root-finding operations.Finally,accurate 2DDOA estimates can be obtained by a sample pairing approach.In addition,numerical simulation results are given to corroborate the advantages of the proposed algorithm. 展开更多
关键词 uniform planar array(UPA) direction of arrival(DOA)estimation RD-ROOT-MUSIC algorithm
下载PDF
上一页 1 下一页 到第
使用帮助 返回顶部