In this paper the low pass filter is discussed in the noisy case. And a regularized low pass filter is presented. The convergence property of the regularized low pass filtering algorithm is proved in theory and tested...In this paper the low pass filter is discussed in the noisy case. And a regularized low pass filter is presented. The convergence property of the regularized low pass filtering algorithm is proved in theory and tested by numerical results.展开更多
Presents the generation of wavelet functions using the filter H(ω)=(1+e -iω 2) N|F(e -iω )|F(z) is a rational function), and then a class of symmetric wavelets including orthogonal B spline wavelets, and finally ...Presents the generation of wavelet functions using the filter H(ω)=(1+e -iω 2) N|F(e -iω )|F(z) is a rational function), and then a class of symmetric wavelets including orthogonal B spline wavelets, and finally the best estimation of regular order obtained by analysing their regularity.展开更多
在相关滤波器跟踪算法中引入正则化后可以有效提高跟踪效率,但需要花费大量精力调整预定义参数,此外还有目标响应发生在非目标区域会导致跟踪漂移等问题,因此提出一种自动全局上下文感知相关滤波器(Automatic Global Context Awareness ...在相关滤波器跟踪算法中引入正则化后可以有效提高跟踪效率,但需要花费大量精力调整预定义参数,此外还有目标响应发生在非目标区域会导致跟踪漂移等问题,因此提出一种自动全局上下文感知相关滤波器(Automatic Global Context Awareness Correlation Filter,AGCACF)跟踪算法.首先,在跟踪过程中利用目标局部响应变化实现自动空间正则化,将自动空间正则化模块加入目标函数,使滤波器专注于目标对象的学习;其次,跟踪器利用目标全局上下文信息,结合自动空间正则化,使滤波器能及时学习到更多与目标有关的信息,减少背景对跟踪性能的影响;接着,在滤波器中加入时间正则化项,来充分学习目标在相邻帧之间的变化,从而获得更准确的模型样本.实验结果表明,与其他跟踪算法相比,AGCACF跟踪算法在距离精度和成功率方面具备更好的跟踪效果.展开更多
文摘In this paper the low pass filter is discussed in the noisy case. And a regularized low pass filter is presented. The convergence property of the regularized low pass filtering algorithm is proved in theory and tested by numerical results.
基金supported by National Natural Science Foundation of China(No.61806006)Jiangsu University Superior Discipline Construction ProjectTalent Introduction Project(No.B12018)。
文摘Presents the generation of wavelet functions using the filter H(ω)=(1+e -iω 2) N|F(e -iω )|F(z) is a rational function), and then a class of symmetric wavelets including orthogonal B spline wavelets, and finally the best estimation of regular order obtained by analysing their regularity.