...
首页> 外文期刊>International journal of biomedical imaging >Evolution-Operator-Based Single-Step Method for Image Processing
【24h】

Evolution-Operator-Based Single-Step Method for Image Processing

机译:基于进化算子的单步图像处理方法

获取原文
           

摘要

This work proposes an evolution-operator-based single-time-stepmethod for image and signal processing. The key component of theproposed method is a local spectral evolution kernel (LSEK) thatanalytically integrates a class of evolution partial differentialequations (PDEs). From the point of view PDEs, the LSEK providesthe analytical solution in a single time step, and is of spectralaccuracy, free of instability constraint. From the point ofimage/signal processing, the LSEK gives rise to a family oflowpass filters. These filters contain controllable time delay andamplitude scaling. The new evolution operator-based method isconstructed by pointwise adaptation of anisotropy to thecoefficients of the LSEK. The Perona-Malik-type of anisotropicdiffusion schemes is incorporated in the LSEK for image denoising.A forward-backward diffusion process is adopted to the LSEK forimage deblurring or sharpening. A coupled PDE system is modifiedfor image edge detection. The resulting image edge is utilized forimage enhancement. Extensive computer experiments are carried outto demonstrate the performance of the proposed method. The majoradvantages of the proposed method are its single-step solution andreadiness for multidimensional data analysis.
机译:这项工作为图像和信号处理提出了一种基于演化算子的​​单步方法。所提出的方法的关键部分是局部谱演化核(LSEK),它可以对一类演化偏微分方程(PDE)进行解析积分。从PDE的角度来看,LSEK可在单个时间步中提供分析解决方案,并且具有光谱准确性,不受不稳定性约束。从图像/信号处理的角度来看,LSEK引起了一系列低通滤波器。这些滤波器包含可控制的时间延迟和幅度缩放。新的基于演化算子的​​方法是通过将各向异性与LSEK系数进行逐点调整而构造的。 LSEK中采用了Perona-Malik型各向异性扩散方案进行图像降噪,而LSEK采用了向前-向后扩散过程进行图像去模糊或锐化。修改了耦合的PDE系统以进行图像边缘检测。所得到的图像边缘被用于图像增强。进行了广泛的计算机实验,以证明该方法的性能。该方法的主要优点是它的单步解决方案和易于进行多维数据分析的能力。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号