Proceduresbasedonwaveformmatching

一种比较新颖的方法，通过对以往提出几种核估计方法的融合，建立新核，达到更好的图像恢复效果(A relatively new method, proposed in the past by the nuclear fusion of several estimation methods, the establishment of new nuclear, to achieve better image restoration effect)

灭点标定，利用拍摄图像的两组灭点，求解摄像机标定的焦距、旋转矩阵、平移向量(Vanishing point calibration, using two sets of images taken vanishing point, solving the camera calibration of the focal length, rotation matrix, translation vector)

说明：  采用双相位编码对图像进行加密解密，采用双随机相位加密技术得到的 S 分量的相息图, 在加密的过 程中引进了随机相位因子, 在不知密钥的情况下解 密出 S 分量几乎不可能, 从而保证了的安全性(Two phase encoding is used to encrypt and decrypt the image)

Tae Hoon Kim教授的Nonparametric Higher-Order Learning for Interactive Segmentation源码(Tae Hoon Kim, Professor Nonparametric Higher-Order Learning for Interactive Segmentation)

利用非局部特性并结合低秩解法的图像降噪算法，速度很快(Nonlocal properties combined with low rank Solution image noise reduction algorithm, fast)

此demo演示的是一个能够检测并跟踪道路上的标记线并且在适当时机发出报警的系统，此系统可使汽车在行驶过程中能够避免偏离车道。系统由一台摄像机，一台计算机，以及一个扬声器组成，摄像机采集视频信息交由计算机处理，根据处理结果控制扬声器是否报警。(This demo demonstrates a system that detects and tracks the marking lines on the road and sends out the alarm at the right time, which enables the car to avoid deviating from the driveway during the driving process. The system is composed of a camera, a computer, and a loudspeaker. The video information is processed by the camera to control the alarm of the speaker according to the processing results.)

说明：  也是水平集方法的一个matlab源代码 比较经典的写的很详细 还可以做3d处理的(level is a collection of source code comparison Matlab classic written in great detail can be done to deal with the 3d)

基于扩展卡尔曼滤波器的粒子滤波器源程序The iterated extended kalman particle filter(based on extended Kalman Filter The particle filter iterated extend source ed Tracking particle filter)

介绍基于彩色不变理论的SIFT算法（CSIFT算法）的文章，文章是英文的(Color change based on the theory introduced SIFT algorithm (CSIFT algorithm) article, the article is in English)

一种简单高效地基于压缩感知的跟踪算法。首先利用符合压缩感知RIP条件的随机感知矩对多尺度图像特征进行降维，然后在降维后的特征上采用简单的朴素贝叶斯分类器进行分类。该跟踪算法非常简单，但是实验结果很鲁棒，速度大概能到达40帧/秒(A simple and efficient tracking algorithm based on compressed sensing. Firstly, with the random sensing matrix compressed sensing RIP conditions for multi-scale image feature dimension reduction, and then use the naive Bias classifier simple classification in the feature reduction after the. The tracking algorithm is very simple, but the results are robust, speed can reach 40 frames per second)