sbs

光流法的程序，利用传统光流法思想，亲测有效，效果不错(Optical flow method program, pro-test, good results)

基于alpha matting的抠图技术(Alpha matting based matting techniques)

matlab 直方图统计可以进行统计进行直方图的绘制(Matlab statistical histogram can be conducted for the drawing histogram)

细节增强的直方图均衡法，在增强图像对比度的同时还能保留细节细心，避免了细节吞噬现象(Detail enhancement, histogram equalization, but also retained the same time enhance the image contrast details carefully, to avoid the details of phagocytosis)

C# 使用Halcon进行模板匹配定位(C# uses Halcon for template matching and location)

提出一种基于非采样contourlet变换的红外与可见光图像融合方法。高频系数结合活性度量和多分辨率系数间相关性来实现加权融合；低频系数通过局部梯度进行活性度量，再采用加权与选择相结合的规则。(Proposed a contourlet transform infrared and visible image fusion method based on non-sampling. The high-frequency coefficient binding activity to measure the correlation between multi-resolution coefficient weighted fusion activity measure low frequency coefficients of the local gradient, and then weighted and selection combining rules.)

2007年ICIP国际会议论文的源码，用于实现对图像的反卷积恢复。(2007 International Conference on ICIP paper source for the realization of the restoration of image deconvolution.)

斯坦福大学的姚邦鹏开发的一个图像分类算法，使用random forest实现了图像的精细区域描述，赢得了PASCAL VOC2011图像分类竞赛中的winner prize.经测试，程序完全可运行，且提供了MAC和Windows下的两种程序。(Image Classification: An Integration of Randomization and Discrimination in A Dense Feature Representation The goal of our method is to identify the discriminative fine-grained image region that distinguishes different classes. To achieve this goal we sample image regions from dense sampling space and use a random forest algorithm with discriminative classifier. Each node of the tree of random forest is trained and tested with fine-grained image patches combining the information from upstream nodes together. We implemented each node of the tree with a discriminative SVM classifier, which makes the node as a strong classifier. )

首先生成如下形式的图片：图中的白条为7象素宽，210象素高，白条间的宽度为17象素。 应用多种滤波器进行处理：算术均值，几何均值，谐波均值，逆谐波均值，中值滤波，中点滤波，最大值、最小值滤波等。注意，需改变模板大小。详细比较上述的结果，并分析原因。 (First generation following form images: figure IOUs 7 pixels wide, 210 pixels high, the width is 17 pixels between IOUs. Using a variety of filters for processing: the arithmetic mean, geometric mean, harmonic mean, inverse harmonic mean, median filtering, filtering midpoint, maximum, minimum filtering. Note that the need to change the template size. A detailed comparison of these results and analyze the reasons.)

说明：  基于局部Fisher准则的非线性核Fisher辨别分析，应用于有监督的特征提取与高维数据的有效降维。(Kernel Local Fisher Discriminant Analysis for Supervised Dimensionality Reduction.)