Tuesday, February 15, 2011

Neural Network Concept神经网络

基本的NN概念

1.Is a control method basically not depends on the Mathematical Model of operational objectives & environment
2.Is suitable for the controlled objectives with uncertainty, complex, inaccuracy or high non-linearity
3.Has strong Adapting, Study functions, Fault tolerance, potential distributed concurrent computing and can treat multi transferred information
特点
structure ,function, link-weight coefficient,learning method
其实可以理解成一个输入输出模型,中间有带权重值的传递函数。

 两个类似的神经网络可以合并,达到用一个网络识别多种信息的功能。
最神奇的,NN具有强大的容错功能,假如输入的花有一个元素有问题,并不影响输出结果的正确性,神经网络仍可以判断出输入是花。


Iterative procedure for image reconstruction and OSEM

除了简单的back projection 和 改进的filtered back projection,用的更多的图像重建方法是迭代法,其中包括EM。

下面是Medical imaging 里面用的较多的OSEM



Image Reconstruction:Back projection

Image reconstructions from projections
经过ct transmission projection
based on Redon equation












Back Projection
BP是实现radon的方法之一
     






















How to realize BP?



















Drawbacks of BP


Filtered Back Projection(Convolution method)