meanshift_filter

说明：  再来一个绘图程序，花了我一个上午时间才搞定。(then a drawing program, I spent a morning time to get.)

说明：  这是个读取视频中每一帧并按顺序保存的小工具，因网上现成的太少，故自己编写(This is a small tool for reading vedio and save every frame in it .)

说明：  以稀疏基有离散余弦变换基（DCT）和快速傅立叶变换基（FFT）做为稀疏基，高斯随机矩阵、部分哈达玛矩阵为测量矩阵，L1范数、正交匹配追踪算法（OMP）为重建算法进行压缩感知算法实现。 以f = cos(2*pi/256*t) + sin(2*pi/128*t)做为原信号，取原信号f的20%做为输入进行压缩感知重建。(The sparse basis includes discrete cosine transform (DCT) and fast Fourier transform (FFT) as sparse basis, Gaussian random matrix and partial Hadamard matrix as measurement matrix, L1 norm and orthogonal matching pursuit algorithm (OMP) as reconstruction algorithm. In this paper, the reconstructed signal is reconstructed by using the original sensing signal (sinf * t) as the original input signal (i.e. sinf * t * 2) + 2 * s * t * as the input signal.)

用于全息成像的算法，采用球面波照射物体，进行图像还原，代码可运行。(The algorithm for holographic imaging uses spherical wave to illuminate objects and restore images, so the code can run.)

VisualC++数字图像处理技术详解第2版光盘-第三章(VisualC++ digital image processing technology Detailed Version 2 CD- Chapter 3)

对静态图像应用canny算法提取边缘特征(The edge feature is extracted by canny algorithm of static image)

使用盒维数法求一幅大小为N*N图象的分形维数(The use of box dimension method the size of an N* N image fractal dimension)

Legendary server controller for game enthusiasts to study, can

图像清晰度评价函数说明 1、熵: 表示图像所包含的平均信息量的多少,嫡值越大则所含信息量越多。 文件名：entropy.m 结果：EN 2、交叉嫡：反映两幅图像的差异,交叉嫡越小,则融合图像和原图像的差别越小。 文件名：cross_entropy.m 结果：平均交叉嫡MCE，均方根交叉嫡RCE 3、峰值信噪比: PSNR越高,说明融合效果与质量越好。 文件名：psnr.m 结果：PSNR 4、Qabf: 评价边缘或梯度质量，越大边缘越明显 文件名：Qabf.m 结果：Qabf（QAB/F) 5、平均梯度（Average Gradient）:也称为清晰度，反映了图像中的微小细节反差与纹理变化特征，同时也反映了图像的清晰度，越大越好。 文件名：avg_gradient 调用：outval = avg_gradient(img) 6、结构相似性指数： SSIM指原图像和融合图像的相似程度,值越大越相似 文件名：ssim.m 结果：SSIM 7、互信息：MI(mutual information) 8、NMI: Normalized mutual information(Image sharpness evaluation function)

边缘检测的各种经典算法，很值得一看。机会不容错过。快来看呀。(edge detection algorithm for the classic, is an eye-catcher. The opportunity should not be missed. Come see ah.)