dct

基于DCT变换的彩色图像压缩；本程序的压缩率高，可达16+，图像恢复出来的效果较好(Color image compression based on DCT transform)

lda线性特征提取，用于人脸识别，首先进行小波特征提取后用lda提取特征。(lda linear feature extraction for face recognition, first of all, after feature extraction using wavelet feature extraction using lda.)

低信噪比下周期平稳信号的稳健检测算法，希望对大家有用(Robust detection algorithms low SNR cycle stationary signals)

对特征值进行非极大值抑制的角点检测，并且选用了合适的阈值加强检测效果(corner detection by non-maxima suppression of Eigenvalues and selected the appropriate threshold to enhance detection)

三维激光点云配准是点云三维建模的关键问题之一。经典的 ICP 算法对点云初始位置要求较高且配准 效率较低，提出了一种改进的 ICP 点云配准算法。该算法首先利用主成分分析法实现点云的初始配准，获得较好 的点云初始位置，然后在经典 ICP 算法的基础上，采用 k － d tree 结构实现加速搜索，并利用方向向量夹角阈值去除 错误点对，提高算法的效率。实验表明，本算法流程在保证配准精度的前提下，显著提高了配准效率。 (Three-dimensional laser point cloud registration is one of the key three-dimensional point cloud model. High classical ICP algorithm to the initial position of the point cloud registration requirements and low efficiency, proposed an improved ICP point cloud registration algorithm. Firstly, the use of principal component analysis of the initial point cloud registration, get a better initial position of the point cloud, then the basis of classical ICP algorithm using k- d tree structure to achieve speed up the search, and using the direction vector angle the removal of the threshold point error and improve the efficiency of the algorithm. Experiments show that the algorithm processes to ensure the accuracy of registration under the premise, significantly improve the efficiency of registration.)

说明：  svm 图像分割 用kmeans得到两类分割的图，在显示的图像中用鼠标取点得到2*num个坐标位置的二维向量，前num个为正样本，后num各为负样本(image segmentation)

图像倒频谱生成并分析，可以用于运动模糊图像的PSF参数估计(Cepstrum image generation and analysis, can be used for motion-blurred image PSF parameter estimation)

这个是广义高斯分布随机数产生程序，它可以用做广义高斯分布形状参数的估计，以及用在图像处理，独立成分分析，信号处理，数字水印等方面。(This is the random numbers about the generalized Gaussian distribution.It can be used in image processing, ICA, signal processing,digital watermark and so on.)

Object Recognition with Hierarchical Kernel Descriptors,In IEEE International Conference on Computer Vision and Pattern Recognition (CVPR), 2011.(Hierarchical Kernel Descriptors.)

从串口实时采集数据，读取家速度传感器的值，并且显示波形(Real-time collection of data from the serial port, read the value of home speed sensor, and display the waveform)