line_sobel

opencv图像库函数集，有关最新的opencv新增函数的应用(opencv functions)

可以实现Ransac拟合直线和椭圆，很方便，可以直接用。(Can be achieved Ransac fitting a straight line and ellipse, very convenient, and can be directly used.)

将dct生成的8*8块图像数据进行zigzag排序，以及恢复算法，在脆弱水印中使用。(The program zigzag sort 8*8 image data generated by dct, and restore algorithm, which is used in fragile watermarking.)

利用K均值聚类的方法实现人体的检测，混合高斯模型建立背景并实时更新(The use of K-means clustering method to achieve the human body detection, Gaussian mixture model to establish the background and real-time updates)

说明：  implementation of mser region detector

这是个criminisi算法的源码，可以对图像进行相应的修补。效果很好。(This is a criminisi algorithm source code, images can be appropriate patch. With good results.)

tsne 快速降维算法 效果好于PCA 算法(tsne fast dimensionality reduction algorithm is better than PCA algorithm)

阴影去除，用于视频监控中的运动目标检测后提取运动目标(Shadow Removal for moving object detection in video surveillance Moving Target Extraction)

"基于四元数卷积的彩色图像边缘检测"一文的matlab源代码，很有帮助(matlab source code of "Colour image edge detector based on quaternion convolution" Electronics Letters, 34(10), May 14 1998, 969-971)

针对散焦模糊图像，用matlab实现了先基于倒谱相关性对PSF的参数估计，再用维纳滤波实现复原的算法。(Defocus blurred image using matlab first estimated based on the parameters of the cepstrum on PSF, and then the Wiener filter recovery algorithm.)