轻松实现GM模型的操作 GM(1-N)

This library contains helper functions to read image frames into matlab from Video file using DirectShow. This library has borrowed codes from Ashwin Thangali s "DirectShow (DirectX) based AVI file reader", and it can read multiple video formats besides avi. Features: a) can read frames from rmvb, avi, etc. b) handles most codecs installed on the system d) its implementation is only based on the IMediaDet interface -- Depending on your Matlab version, the .mexw32 files may need to be renamed as .dll -- The library usage is as follows, [avi_hdl, avi_inf] = dxVideoOpen(avi_filename) pixmap = dxVideoReadMex(avi_hdl, frame_num) img = reshape(pixmap,[avi_inf.Height,avi_inf.Width,3]) dxVideoCloseMex(avi_hdl) -- recompile the library I have compiled the code with the DirectShow lib in Microsoft DirectX 9.0 SDK Update (December 2004) -- known limitation a)only works for windows :( b)may be slow for big image frames and some codecs.

此程序实现构造小波基 (this program structure of this small Porgy program structure wavelet)

In this document we proposed a face recognition algorithm that enhance the recognition rate and reduce the complexity.

OMP算法，用于图像的重构过程中，比较经典，尤其是数字信号处理中(OMP algorithm for image reconstruction process, the more classic, especially in digital signal processing)

generalized predictive control

利用MATLAB对神经网络进行编程,用newff()创建两层前向网络。网络输入范围[-1 1]，第一层有10个tansig神经元(using MATLAB right neural network programming with newff () to the creation of a two-tier network. Network input range [-1 1], the first layer 10 tansig neurons)

用递推最小二乘法辨识系统参数(6位的M序列作为输入信号,递推300次)(using recursive least squares identification system parameters (6 M sequence as the input signal, recursive 300))

最优分割的基本思路是给定一个数据序列和分层数，通过搜索所有可能的划分方案，找到段内离差平方和的总和最小的一种方案作为最终划分方案。这里给出最优分割法的Matlab源程序(The basic idea is optimal partition for a given data sequence and hierarchical number divided by searching all possible solutions to find the segment and the sum of the squared deviations smallest division of a scheme as a final solution. This gives the optimal segmentation Matlab source)

这个算法是信源数估计中经典的AIC算法，我已验证过。(This algorithm is the source of the AIC estimates classical algorithm, I have verified.)

说明：  nmf盲源分离，使用欠定的方法，可以用来分离两路语音信号，人声与其他音的分离(nmf_bss.The underdetermined method can be used to separate two voice signals, human voice and other voices.)