something

RN feed-fordward, para realizar redes neuronales fordwar, se utiliza la funcion de activacion sigmoideal

如果你有一个三维校准对象，你可以应付一个单一的形象。为了获得令人满意的校正结果，对象应包括整个图像以及可能的。此外，多张图片的支持。那么， 图像应捕获不同的观点改变了相机 方向和距离。如果在一个平面标定目标单一 形象是不够的和一套图像是需要解决所有 摄像机的参数。的坐标平面控制点应 被选中，这样的Z坐标成为为零。三维坐标股 毫米和图像坐标单位是像素。校准坐标 系统是用右手。原产地的图像坐标系中 左上角， X轴是正确的和Y轴向下。(If you have a three-dimensional calibration object, you can cope with a single image. Calibration In order to obtain satisfactory results, the target should include the entire image, as well as possible. In addition, the support of many pictures. Then, the image should capture different points of view camera has changed the direction and distance. If a planar calibration target at a single image is not enough and a set of images are necessary to solve all the camera parameters. The coordinates of horizontal control points should be selected, such as Z coordinates become zero. Unit-mm three-dimensional coordinates and image coordinates are pixel units. Calibration are right-handed coordinate system. Origin of the image coordinate system in the upper left corner, X-axis and Y are correct under axial.)

point rotation using Quaternion. implementation in matlab. Program is a tool to rotate poin in Cartesian coordinates in three dimensions about any angle. Run moje2.m

用于盲信号分离的LMS算法，可以从混合的观测向量中分离出需要的信号，判读的标准是计算串音误差(For the LMS algorithm for blind signal separation can be observed from the hybrid vector needs to separate out the signals, the standard interpretation is to calculate the crosstalk error)

利用matlab语言实现递推最小二乘算法，里面包含程序运行时的xls数据(Matlab language implementation using recursive least squares algorithm, which contains the program runs when the xls data)

基于IOS自组织分析法的iris数据集分类，matlab实现，含注释。(The iris data set classification based IOS self-organization analysis, Matlab implementation, including annotations.)

LMS algorith and DE in matlab

对于规则LDPC码的EXIT图的程序，对优化LDPC起到很好的帮助(For regular LDPC code EXIT chart program optimization LDPC play a very good help)

用有限差分法编程求解拉普拉斯方程在网格结点处的值并统计迭代次数(Solving the value of the Laplace equation in a grid junction with the finite difference method and statistical programming iterations)

texture segmentation with gabor filters