微软人脸检测相关研究技术报告，内有详细人脸检测的研究和实现方法...

该程序能够实现动态画圆，每隔一段时间画一个填充颜色的圆。-The program can dynamically draw a circle, from time to time to draw a fill color circle.

生产成多项式的基函数，输出为一个矩阵，可产生任意阶。-Production into the polynomial basis functions, the output of a matrix, can produce arbitrary order.

DICOM协议的服务器端程序，支持图像回调，图像检索等功能。-agreement DICOM server software, support image correction, image retrieval capabilities.

曲线的绘制的控件代码:具有水平和垂直滚动条,带有测试事例!一次可绘制三个分量-curve mapping the control code : with horizontal and vertical scroll, with a test case! Rendering can be a component ... 3

此程序用C++开发， 属于神经网络中经典的hopfeild算法 程序， 包括源代码。-this program with C development, neural networks belonging to the classic hopfeild algorithm procedures, including source code.

java实现的一个画图程序。功能齐全，类似于win的画图板的一个简化版本-realize a java drawing program. Functions, similar to win the drawing board of a simplified version of

surf算法的具体实现： Surf算法原理 (1)、构建Hessian矩阵 (2)、构建尺度空间 图像的尺度空间是这幅图像在不同解析度下的表示，由式(4)知，一幅图像j(X)在不同解析度下的表示可以利用高斯核G(￡)的卷积来实现，图像的尺度大小一般用高斯标准差来表示[6]。在计算视觉领域，尺度空间被象征性的表述为一个图像金字塔，其中，输入图像函数反复与高斯函数的核卷积并反复对其进行二次抽样，这种方法主要用于Sift算法的实现，但每层图像依赖于前一层图像，并且图像需要重设尺寸，因此，这种计算方法运算量较大，而SURF算法申请增加图像核的尺寸，这也是SIFT算法与SURF算法在使用金字塔原理方面的不同。算法允许尺度空间多层图像同时被处理，不需对图像进行二次抽样，从而提高算法性能。图1(a)是传统方式建立一个如图所示的金字塔结构，图像的寸是变化的，

Protel 2004 教程 pdf格式-Guide pdf format

触摸屏触摸屏

Cognitive radio frequency spectrum detection-The spectrum sensing of a wideband frequency range is studied by dividing it into multiple subbands. It is assumed that in each subband either a primary user (PU) is active or absent in a additive white Gaussian noise environment with an unknown variance. It is also assumed that at least a minimum given number of subbands are vacant of PUs. In this multiple interrelated hypothesis testing problem, the noise variance is estimated and a generalised likelihood ratio detector is proposed to identify possible spectrum holes at a secondary user (SU). Provided that it is known that a specific PU can occupy a subset of subbands simultaneously, a grouping algorithm which allows faster spectrum sensing is proposed. The collaboration of multiple SUs can also be considered in order to enhance the detection performance. The collaborative algorithms are compared in terms of the required exchange information among SUs in some collaboration