PC20121224

根据RBM深度网络（Hinton，2006）进行MINST手写文字的分类器的训练，利用训练得到的权值制作了这个小程序，通过这个程序可以看出训练结果对数据集内的测试样本和训练样本都能进行很好的识别，但是对其他的手写字体识别就没有那么好了。(According to RBM depth network (Hinton, 2006) conducted MINST handwritten text classifier training, using the training to get the right value to produce this small program, this program can be seen through the results of the training data set and the training sample test samples are can be well identified, but on the other handwriting recognition is not so good.)

说明：  图像分割代码，c++实现(Image segmentation code, c++ Achieve)

说明：  用于三维空间坐标系变换的7参数模型，在设计坐标系变换的各行业应用中用的最多，从数学角度来说也是最严密的转换方法。 由于结果中最多可求得七个转换参数，即三个平移参数、三个旋转参数(Ex、Ey、Ez)和一个尺度缩放因子(m)，因此，通常也被称为七参数法。 (Three-dimensional coordinate system transformation for the 7 parameter model, coordinate system transformation in the design of applications using the most, the industry, from the mathematical point of view, the conversion is the most rigorous methods. Since the results can be obtained up to seven transformation parameters, ie three translation parameters, three rotation parameters (Ex, Ey, Ez) and a zoom scale factor (m), therefore, often referred to as seven-parameter method.)

多类 kmeans 图像分割，基于灰度的图像分割。(Kmeans image segmentation, segmentation based on gray-scale images.)

bgslibrary_x86_v1.9.2,是MFC的实现文件，包含37种背景减除算法，对于做目标提取的研究非常有用。(The BGSLibrary was developed by Andrews Sobral and provides an easy-to-use C++ framework based on OpenCV to perform background subtraction (BGS) in videos. )

经典的数字图像处理的形态学处理，包含源代码(Digital image processing morphological processing)

一幅大小为256×256的256级灰度图像（每像素8bit） 如果图像的各个像素之间是统计独立的，求该图像的信息熵 -离散无记忆信源 如果图像的各个像素之间不是统计独立的，求该图像的信息熵 -图像像素的概率分布模型(A 256 gray scale image (8bit per pixel) with a size of 256 x 256 If each pixel of the image is statistically independent, the information entropy of the image is calculated Discrete memoryless source If each pixel of the image is not statistically independent, the information entropy of the image is calculated Probability distribution model of image pixels)

用于BMP图像文件的滚动显示，带上下和水平滚动条。(This file is used to display BMP image files, with up-down and horizontal scroll bars.)

在分析聚束合成孔径雷j盘(SAR)方位向时频关系的基础上，提出了一种新的全孔径分辨率SAR成像算法。 根据SAR在条带与聚束模式下回波信号之间的差别，利用现有成熟的条带SAR成像算法实现先对聚束SAR各方位向子 孔径成像，而后再将各方位子孔径带宽合成为全孔径带宽，以形成具有全孔径分辨率的SAR图像，极大降低了系统PRF 的要求。(In the analysis of spotlight synthetic aperture radar dish j (SAR) azimuth based on the time-frequency relationship, a new full-aperture resolution SAR imaging algorithm. According to SAR in strip and spotlight mode, the difference between the echo signal, the use of existing mature band SAR imaging algorithm first bit of the spotlight SAR the parties to the sub-aperture imaging, and then place then all the bandwidth synthetic aperture the bandwidth for the full aperture to form a full aperture resolution SAR images, greatly reducing the system PRF requirements.)

本程序和实现对一幅包含人脸的bmp图像进行光纤补偿，对人脸区域进行检测实现人眼和人嘴的检测与定位，最后勾绘出人脸的特征以实现人脸检测的目的。(This procedure and implementation of a compensate the fiber contains face of BMP image, the face region detection to realize human eyes and mouth detection and positioning, finally drew out the facial features to achieve the purpose of face detection. )