Description: DoG(Difference of Gaussian)滤波算子，主要用于边缘特征提取，用于模式识别中的分割预处理。其主要参数为两个高斯函数的方差，针对方差设计可以对不同的图像特征情况下有不同的表现。-DoG (Difference of Gaussian) filtering operator, the edge feature extraction is mainly used for pattern recognition of the partition pretreatment. The main parameters for the two Gaussian function of variance, the variance for design characteristics of the different images have different circumstances performance. Platform: |
Size: 2048 |
Author:wang |
Hits:
Description: 人工智能在围棋程序中的应用-- 本文介绍了人工智能在围棋程序中的应用与发展，对比了围棋与国际象棋博弈算法的差别和复杂度，从而分析围棋算法的难点，讨论各种博弈算法（气位理论、模式匹配与博弈树）在围棋程序中的融合运用。并给出了围棋死活程序的算法。-Go artificial intelligence in the application process- This article describes the procedure of artificial intelligence in the Go application and development, compared with the international chess game Go difference algorithm and complexity, so Go algorithm difficult to discuss a variety of game algorithm (gas-bit theory, pattern matching and game tree) in the Go to use the integration process. And gives life and death procedures Go algorithms. Platform: |
Size: 7168 |
Author:李荣春 |
Hits:
Description: 计算有限差分与传播光束法结合的程序，很有帮助-Calculation of finite difference beam method combined with the dissemination of the procedure, very helpful Platform: |
Size: 92160 |
Author:黄志祥 |
Hits:
Description: 两个模式识别算法实现，一个是线性区别函数另一个是混合高斯模型方法。本人的大作业，经验证可用。-Two pattern recognition algorithm, the difference between a linear function of the other is the Gaussian mixture model. I am a big operation, experience certificates available. Platform: |
Size: 252928 |
Author:苏冠华 |
Hits:
Description: a program which implements a finite difference algorithm for a predator-prey system with spatial variation in 2D Platform: |
Size: 1024 |
Author:jahan |
Hits:
Description: 小波变换在信号处理、图像处理、模式识别、计算机视觉等
方面广泛应用。本文提出了一种基于Haar小波的时问序列相似
模式匹配模型．它首先对时间序列数据进行Haar小波变换．以
降低数据的维度。首先将所要查询的时间序列进行降维，也就是
对时间序列数据进行小波变换，对序列进行标准化131，得到降维
之后的小波序列．这样可以得到系数子集的良好近似采用尺度
序列表示原始序列．并将其看成多维空间中的一个点，通过计算
两序列差的平方和的平方根作为这两个时间序列的距离函效
(即欧氏距离)。用小波变换保持了局部性质。而欧氏距离就是要
判断相似性的距离函数．所以当在作完变换之后。通过低频分量
来压缩数据，成功的计算出实际距离的下界。如果计算的结果小
于一个由用户所定义的门槛值．则认为这两个时问序列是相似
的。
在-Of wavelet transform in signal processing, image processing, pattern recognition, computer vision, etc.
Aspects of wider application. In this paper, Haar wavelet-based time series similarity
Pattern matching model. It is first time-series data, Haar wavelet transform. To
To reduce the data dimension. Will first want to check the time series dimension reduction, that is,
Pairs of time series data using wavelet transform, standardize the sequence 131, to be reduced-
After the wavelet sequence. This can be a good approximation coefficients of a subset of the use of scale
Sequence indicated that the original sequence. And to treat it as a multi-dimensional space a point, by calculating the
Two sequences and the square root of the square of the difference as the distance between these two time series letter effect
(Ie, Euclidean distance). Maintained with the local nature of the wavelet transform. The Euclidean distance is to
To determine simila Platform: |
Size: 323584 |
Author:lx |
Hits:
Description: Block Matching Algorithms for Motion Estimation
This project contains the project report and source code by Aroh Barjatya for Digital Image Processing Class at Utah State University.
Following is a short description of the m files in this zip
motionsEstAnalysis.m Script to execute all Algorithms
motionEstES.m Exhaustive Search Algorithm
motionEstTSS.m Three Step Search Algorithm
motionEstNTSS.m New Three Step Search Algorithm
motionEstSESTSS.m Simple And Efficient Search Algorithm
motionEst4SS.m Four Step Search Algorithm
motionEstDS.m Diamond Search Algorithm
motionEstARPSm Adaptive Root Pattern Search Algorithm
costFuncMAD.m Mean Absolute Difference Function
minCost.m minimum cost among macro blocks
motionComp.m motion compensated image creator
imgPSNR.m finds image PSNR w.r.t. reference image
The test images can be found at http://cc.usu.edu/~arohb/caltrain.zip-Block Matching Algorithms for Motion Estimation
This project contains the project report and source code by Aroh Barjatya for Digital Image Processing Class at Utah State University.
Following is a short description of the m files in this zip
motionsEstAnalysis.m Script to execute all Algorithms
motionEstES.m Exhaustive Search Algorithm
motionEstTSS.m Three Step Search Algorithm
motionEstNTSS.m New Three Step Search Algorithm
motionEstSESTSS.m Simple And Efficient Search Algorithm
motionEst4SS.m Four Step Search Algorithm
motionEstDS.m Diamond Search Algorithm
motionEstARPSm Adaptive Root Pattern Search Algorithm
costFuncMAD.m Mean Absolute Difference Function
minCost.m minimum cost among macro blocks
motionComp.m motion compensated image creator
imgPSNR.m finds image PSNR w.r.t. reference image
The test images can be found at http://cc.usu.edu/~arohb/caltrain.zip Platform: |
Size: 118784 |
Author:Yashil |
Hits:
Description: 该程序用VC绘制了金刚石图案，根据等分点不同，和半径的不同有不同的形状-The program drew a diamond pattern with the VC, according to the difference equal portions, and the radius of different shapes have different Platform: |
Size: 1956864 |
Author:crtdzd |
Hits:
Description: 对于运动图像分析菜单的一点说明
该菜单为主框架里的一个菜单项，对应第九章运动分析内容，该菜单提供了两种分析方法，提供的两幅静态图像为随书光盘所带“各章测试图”中“第九章”文件夹内的“图9-5a.bmp”和“图9-5b.bmp”。
归一化相位相关方法和相位差方法是两种运动估计方法，根据算法原理的不同，其输出形式也不相同。
例如，对于所提供的128*128大小的一对测试图，相位差的方法输出形式为：“相对于第一幅图像，x方向移动1像素，y方向移动3像素” ，其含义是相对于第一幅图像，第二幅图像在x正方向有1个像素的位移，在y方向有3个像素的位移。
而相位相关方法以图像中心为基准。例如，相位相关输出为“相对于第一幅图像，x方向移动为63个像素，y方向移动为66个像素”，代表中心在x方向移动1个像素，在y方向移动了3个像素。其运动方向不能根据结果的正负号判断
-For the analysis of moving images that explain the menu
The menu-based framework of a menu item, the corresponding movement of the contents of Chapter IX, the menu offers two analysis methods, providing the two still images brought to CD with the book, "each chapter test pattern" in the "Ninth chapter "in the folder" Figure 9-5a.bmp "and" Figure 9-5b.bmp ".
Normalized phase correlation method and the phase difference of two motion estimation methods is, according to the principle of different algorithms, the output is not the same form.
For example, provide a 128* 128 size of the test chart, form the output phase approach: "relative to the first image, x 1 pixel direction, y direction by 3 pixels", its meaning is relative to the the first image, the second image in the x direction is 1 pixel displacement, in the y direction of the displacement of 3 pixels.
The phase correlation method to the image center as the base. For example, the phase correlation output is "relative to the Platform: |
Size: 546816 |
Author:2510 |
Hits:
Description: library DXP this g
A really good web developmen
Fingerprint image extraction pdf
An Introduction to Design Pattern
Digital Image Processing classic
OGRE 3D 1.7 Beginners Guide
OReilly Knoppix Hacks 2nd
Iterative Methods for Linear and
10 sets of exercises linear algeb
Finite Difference Methods for Dif-library DXP this g
A really good web developmen
Fingerprint image extraction pdf
An Introduction to Design Pattern
Digital Image Processing classic
OGRE 3D 1.7 Beginners Guide
OReilly Knoppix Hacks 2nd
Iterative Methods for Linear and
10 sets of exercises linear algeb
Finite Difference Methods for Dif Platform: |
Size: 13600768 |
Author:liuH |
Hits:
Description: 画阵列天线3维方向图,调用方式antenna3(n,deta_d,deta_p,loc,unit)
返回值可用于进一步分析截面方向性，各参数含义如下:
n 表示天线元的个数
deta_d 代表阵元天线间距与波长的比值，此处默认阵列为等间距阵
deta_p 代表阵元天线电流相位差，此处默认阵列为等幅，等差相位阵列
loc 代表单元天线与阵列取向关系(0表示单元天线与阵列垂直，1表示单元天线与阵列平行)
unit 代表单元天线的类型(0表示理想点源，1表示电偶极子天线，2表示半波振子)
其中loc默认值为0(垂直)，unit默认值为2(半波振子).
例：有一10元阵列，阵元间距为0.5个波长,相位差为0,单元天线取向与阵列垂直,单元天线为半波振子
则做图时函数调用方式为：antenna3(10,0.5,0,0,2)或利用默认值：antenna3(10,0.5,0)-Painting 3-D antenna array pattern, the call mode antenna3 (n, deta_d, deta_p, loc, unit)
Return value can be used for further analysis section of the direction, the meaning of each parameter as follows:
n the number of the antenna element
deta_d antenna array element spacing on behalf of the ratio of the wavelength, where the default array spaced array
on behalf of the antenna array elements deta_p current phase, where the default array amplitude, phase array arithmetic
Representative element antenna array loc orientation relationship (0 vertical array antenna unit, a unit of the antenna array, said parallel)
representative of the type of unit antenna unit (the ideal point source 0, 1 electric dipole antenna, and 2 half-wave dipole)
Where loc default value is 0 (vertical), unit default is 2 (half-wave dipole).
Example: a 10 per array, array element spacing is 0.5 wavelengths, the phase difference is 0, element antenna array, the vertical orientation and the unit for the Platform: |
Size: 1024 |
Author:wang |
Hits:
Description: * Coupling between classes and class libraries is a major maintenance
* headache. To ease this problem, often the client talks to an
* abstraction description, which in turn calls an implementation.
* Sometimes these must evolve - when one changes there can be a need
* to change the other. The bridge design pattern lets the abstraction
* and its implementation evolve separately.
*
* So, what is the difference between a bridge and an interface? Interfaces
* can be used when creating bridges - but it should be noted that bridges
* have additional possibilities. Both the abstraction and the
* implementation may evolve over time and be the parent of derived classes.
* The operations needed in the implementation could be defined in an
* interface if there are no standard methods which are available at the
* top-level of the implementation.
*
*/-* Coupling between classes and class libraries is a major maintenance
* headache. To ease this problem, often the client talks to an
* abstraction description, which in turn calls an implementation.
* Sometimes these must evolve- when one changes there can be a need
* to change the other. The bridge design pattern lets the abstraction
* and its implementation evolve separately.
*
* So, what is the difference between a bridge and an interface? Interfaces
* can be used when creating bridges- but it should be noted that bridges
* have additional possibilities. Both the abstraction and the
* implementation may evolve over time and be the parent of derived classes.
* The operations needed in the implementation could be defined in an
* interface if there are no standard methods which are available at the
* top-level of the implementation.
*
*/ Platform: |
Size: 1024 |
Author:lcx |
Hits:
Description: MATLAB函数参考手册，查看matlab函数作用以及功能。- SVMLSPex02.m
Two Dimension SVM Problem, Two Class and Separable Situation
Difference with SVMLSPex01.m:
Take the Largrange Function (16)as object function insteads ||W||,
so it need more time than SVMLSex01.m
Method from Christopher J. C. Burges:
"A Tutorial on Support Vector Machines for Pattern Recognition", page 9
Objective: min "f(A)=-sum(ai)+sum[sum(ai*yi*xi*aj*yj*xj)]/2" ,function (16)
Subject to: sum{ai*yi}=0 ,function (15)
and ai>=0 for any i, the particular set of constraints C2 (page 9, line14).
The optimizing variables is "Lagrange Multipliers": A=[a1,a2,...,am],m is the number of total samples. Platform: |
Size: 561152 |
Author:王东东 |
Hits:
Description: SMI方向图和性能(波束形成算法，自己修改）
优点： 收敛速度快
缺点：1 当阵元输出含有较强的期望信时，或者期望信号与干扰信号相关时，性能急剧下降.2 由于权向量含有方向矢量，因此对阵列的幅相差非常敏感；3 ,期望信号的功率不能过大，比干扰信号相差几十dB，也就是说，在小期望信号、大干扰信号情况下，也可进行SMI处理。
-SMI pattern and performance (beamforming algorithm, make changes to)
Advantages: fast convergence
Disadvantages: 1 output when the array elements with strong expectations of the letter, or related to the desired signal and interference signals, the performance sharply .2 direction vector containing the weight vector, so the array is very sensitive to the amplitude difference 3, the desired signal The power can not be too large, than the difference between the scores of interference signal dB, that is, in the small desired signal, large signal interference case, SMI can also be handled.
Platform: |
Size: 1024 |
Author:林朝 |
Hits:
Description: 利用角度的不同进行识别，两图像的模式识别范畴-Using the perspective of the different recognition, two images of the field of pattern recognition. Platform: |
Size: 2048 |
Author:王伟楠超 |
Hits:
Description: The principle of back propagation is actually quite easy to understand, even though the maths behind it can look rather daunting. The basic steps are:
Initialise the network with small random weights.
Present an input pattern to the input layer of the network.
Feed the input pattern forward through the network to calculate its activation value.
Take the difference between desired output and the activation value to calculate the network’s activation error.
Adjust the weights feeding the output neuron to reduce its activation error for this input pattern.
Propagate an error value back to each hidden neuron that is proportional to their contribution of the network’s activation error.
Adjust the weights feeding each hidden neuron to reduce their contribution of error for this input pattern.
Repeat steps 2 to 7 for each input pattern in the input collection.
Repeat step 8 until the network is suitably trained.-The principle of back propagation is actually quite easy to understand, even though the maths behind it can look rather daunting. The basic steps are:
Initialise the network with small random weights.
Present an input pattern to the input layer of the network.
Feed the input pattern forward through the network to calculate its activation value.
Take the difference between desired output and the activation value to calculate the network’s activation error.
Adjust the weights feeding the output neuron to reduce its activation error for this input pattern.
Propagate an error value back to each hidden neuron that is proportional to their contribution of the network’s activation error.
Adjust the weights feeding each hidden neuron to reduce their contribution of error for this input pattern.
Repeat steps 2 to 7 for each input pattern in the input collection.
Repeat step 8 until the network is suitably trained. Platform: |
Size: 50176 |
Author:Putra |
Hits:
Description: 通过调用OpenCV库，采用逐差法进行视频跟踪。在源码中输入要跟踪的图样，运行文件，程序可以识别跟踪摄像头摄取场景内与图样相关性最高的目标-By calling the OpenCV library, using a case-by-difference method for video tracking. Input source you want to track the pattern, run the file, the program can identify tracking camera scene and drawings highest goal intake Platform: |
Size: 1024 |
Author:高传清 |
Hits:
Description: 计算duomo阵列的和差方向图，方向系数,副瓣电平。可自行修改参数。-Duomo and computing array difference pattern, the direction coefficient, sidelobe level. You can modify the parameters themselves. Platform: |
Size: 1024 |
Author:Ian |
Hits:
Search - difference pattern