sift

振动信号预处理。去噪、去趋势项等。滑动平均法消除趋势项。(Vibration signal preprocessing. De-noising, the trend term.)

自己修改的基于closed-form的程序，对外国程序在时间上有两倍的改善，能对在复杂背景下的图片，通过画一些简单的前景线，背景线，抠图出前景(it is about matting foreground when the interested ared is located under complixed background,and then choosing some other beautiful background,you can create new image,which brings you some special feelings.thanks.)

双数复小波阈值除噪程序，可以实现双数复小波分解及重构。并能实现阈值除噪。(Double the number of complex wavelet threshold denoising procedures can achieve double the number of complex wavelet decomposition and reconstruction. And to achieve the threshold noise removal.)

说明：  1.学习使用fft2,ifft2,fftshift,ifftshift函数。用fft2函数对”lena.bmp”做二维傅立叶变换，并显示其频谱，用fftshift函数将频谱坐标原点移动到图像中心。用ifft2函数做傅立叶反变换。（提示：显示频谱时，可以先对频谱取对数，这样可以压缩动态范围很大的频谱，获得更好的显示效果）(1. Learn to use fft2, ifft2, fftshift, ifftshift functions. With the FFT 2 function, the lena.bmp'The two-dimensional Fourier transform is performed and the spectrum is displayed. The origin of the spectrum coordinate is moved to the center of the image by using the FFT shift function. The inverse Fourier transform is done with ifft2 function. (Note: when displaying the spectrum, the logarithm of the spectrum can be taken first, so that the spectrum with a large dynamic range can be compressed to obtain better display effect.))

李纯明博士的无需初始化模型的源代码，作者亲自写的。还有测试图象，相信对大家肯定有好处。(Economics Dr. initialization model without the source code, written by the author himself. There test image, I think we will benefit.)

很好用的三维重建软件，VisualSFM软件允许上传一系列图像，它从这些图像中找到每一个图像的特定特征，利用这些特征信息重建出3D模型的稀疏点云，而后还可进行稠密点云重建。(The VisualSFM software, a good 3D reconstruction software, allows uploading a series of images. It finds the specific features of each image from these images, and USES these feature information to reconstruct the sparse point cloud of the 3D model, and then it can reconstruct the dense point cloud.)

CZT变换进行信号频谱分析，解决了FFT法频域分辨不高的问题，源码开发，特别适用(Chirp Z Transform)

根据MPEG7中图像的SCD描述算子，提取图像在HSV空间中的直方图。并且可以与源图像进行比较。(MPEG 7 Scalable Color Descriptor experiment)

说明：  image 数组 高速读写 不用指针(image array IO C#)

关于指纹图像处理的有关论文资料，希望各位同仁多多 指教。(Fingerprint image on paper deal with the relevant information, I hope that Tongren exhibitions.)