完成基于遗传算法的图像模式识别功能流程描述

HMM用于人脸识别，从OPENCV里提取出来的,非常经典的说-HMM for face recognition, OPENCV extract from the very classical said

本程序实现了一个人工神经网络的算法，相关领域的人可以参考。-This procedure implements a neural network algorithm, in related fields who can refer to.

改进遗传算法－郭涛算法做最优化问题很管用,算法的基本思想是 先任意产生n个随机数，然后从n个数里随机选择m个数，再有这m个 数合成一个新数，将这个新数同n个数中间适应值函数值的最差的比较， 如果好的话就取代最差的那个，如果它比最好的还要好的话，则把最好的 也取代。如果比最差的坏，则重新合成一个新数。依次循环下去。 程序的奇妙之处是GA_crossover()函数，产生的新数确实比较好，看看 那位大侠能改进一下，产生比这跟好的数。-improved genetic algorithm- Guo Tao done optimization algorithm is very effective, and the algorithm is the first basic idea arbitrary n generated random number and then the number li n m randomly selected number, and this number m of a new synthesis of the new middle with n number fitness function of the worst by comparison, if a good case to replace the worst, um, if it even than the best, then ruled the best have replaced. If worse than the worst, a de novo synthesis of new. Followed by the cycle continues. The magic of procedures is GA_crossover () function, the new really good to see heroes who can be improved upon, it generated more than a few with a good.

线性神经网络应用实例-linear neural network application

本书介绍了一些关于机器人的识别技术，值得学习和借鉴，作为基础资料不错-The book introduced a number of robot recognition technology, it is worth to learn from, as a good basis for information

本程序可以对输入.avi 序列中的车辆数进行计算，准确性非常高，是一个值得参考的代码:)-This procedure can be input. Avi sequence to calculate the number of vehicles, accuracy is very high, is a reference to the code:)

一个基于MFC的人工智能的巫魔问题求解，图形化界面显示。-an artificial intelligence-based MFC Witch of the devil problem solving, graphical interface shows.

神经网络（模式识别模块）软件开发包!!!可以识别任何事物，适用于图像识别、语音识别、自动控制等诸多领域，具有简单、易学、开发周期短、识别率高等特点。-neural network (pattern recognition module) software development kit! ! ! Nothing can be identified and applied to image recognition, voice recognition, automatic control and many other areas, which is simple, easy to learn, short development cycle, features high recognition rate.

由于K－均值聚类算法局部最优的特点，而模拟退火算法理论上具有全局最优的特点。因此，用模拟退火算法对聚类进行了改进。20组聚类仿真表明，平均每次对K结果值改进8次左右，效果显著。下一步工作：实际上在高温区随机生成邻域是个组合爆炸问题（见本人上载软件‘k－均值聚类算法’所述），高温跳出局部解的概率几乎为0，因此正考虑采用凸包约束进行模拟聚类，相关工作正在进行。很快将奉献给各位朋友。-as K-means clustering algorithm for optimal local characteristics, and simulated annealing algorithm theory with the characteristics of the global optimum. Thus, simulated annealing algorithm for clustering improvements. Cluster Group of 20 simulations show that the average value of K results improved about eight times, the results are obvious. The next step : In fact, in high temperature generated random neighborhood is a combination of explosives (see my software on the "k-means clustering algorithm" mentioned above), high-temperature solution of partial out almost zero probability, it is considering the use of convex hull bound for simulation cluster, the work under way . Soon dedication to the ladies.

人工神经网络BP算法 1、动态改变学习速率 2、加入动量项 3、运用了Matcom4.5的矩阵运算库（可免费下载,头文件matlib.h)， 方便矩阵运算,当然，也可自己写矩阵类 4、可暂停运算 5、可将网络以文件的形式保存、恢复-Artificial neural network BP algorithm 1, the dynamic change learning rate 2, adding momentum of 3, using the matrix calculation Matcom4.5 library (can be downloaded free of charge, the first document matlib.h), facilitate the matrix calculation, of course, can also write their own matrix Class 4, may be suspended by the operator 5, may be the form of network to the preservation, restoration