搜索资源列表
文字识别程序
- 目前该手写体识别系统主要分为 预处理模块： 主要包括训练数据和识别数据的读取，归一化，二值化 特征提取模块：主要包括笔划方向特征和网格密度特征，还可以根据对识别率的要求继续增加其他特征 识别（分类器）模块：主要包括SVM方法和BP神经网络的方法，其中SVM方法的识别率较高，SVM＋网格密度特征， 在小字符集情况下，达到了识别率97％以上 采用OO思想编写，适合做二次开发-currently the handwriting recognition system consists of pretre
手写识别系统
- 数字识别，识别手写体。 使用神经网络算法编写。 识别率一般，带训练功能-Number recognizing system,used to recogize the hand writing figures.It is writen by nero network and has low recognizing effects but we can exercise it.
神经网络法字符识别
- 这是一个基于神经网络法的字符识别系统，神经网络法的使用使本系统识别率很高。-This is a neural network based on the character recognition system, neural network to enable the use of the system to identify high rate.
LPRecognizer.rar
- 车牌识别，神经网络，识别率高达95％，定位准确。,License plate recognition, neural networks, recognition rate as high as 95 percent, positioning accuracy.
sasdw.rar
- 现有数字信号自动调制识别方法大多只适用于无记忆信号，如PSK、ASK、FSK信号等。将有记忆 信号(MSK信号)和无记忆信号一起考虑，提出了一种改进的数字信号自动识别方法。该方法采用信号的瞬时统 计量作为特征参数，采用多层神经网络作为分类器。计算机仿真表明：当噪声采用高斯白噪声，并且信噪比大于 l5 dB时，识别率高于96％ ；当信噪比不低于l0 dB时，识别率不低于90％。,Existing digital signal automatic modulation recognition
PlateRecognizer
- 车牌识别，神经网络算法，识别率高达95％，识别时间低于80ms。-License plate recognition, neural network algorithm, the recognition rate as high as 95 percent, to identify less than 80ms.
LPR-mobile
- 视频车辆牌照识别，包含汉字处理算法，BP神经网络设计。识别率高达90％以上。-Video vehicle license plate recognition, including the Chinese characters processing algorithms, BP neural network design. Recognition rate as high as 90 .
DigitalRec
- 手写体数字识别的VC实现 使用神经网络算法对手写体数字进行识别，训练后识别率可达90％左右。-Handwritten Digit Recognition of the VC to achieve the use of neural network algorithm for handwritten numeral recognition, training, recognition rate can reach about 90.
neuroocr
- 基于神经网络的字符识别,很高的识别率,80 .-Based on neural network character recognition, a high recognition rate, 80 .
ex3
- 基于BP神经网络识别字符. BP神经网络算法是把一组样本输入输出问题转化为一个非线性优化问题，并通过梯度算法利用迭代运算求解权值的一种学习方法。采用BP网络进行分类，并附加线性感知器来实现单字符的有效识别，算法简便，识别率高，可适用于多种高噪声环境中的印刷体字符识别。-BP neural network based character recognition. BP neural network algorithm is a set of sample input and output is
DCT
- 本文设计基于DCT的人脸识别系统，首先结合当今人脸识别的背景和发展状况讨论了人脸识别的研究内容及在各方面的应用；然后研究了人脸识别进行预处理，讨论了人脸识别预处理的其他方法，分析各种方法的利弊，最后采用DCT（离散余弦变换）实现人脸图像预处理中的降维处理；接下来对人脸图像的特征提取进行了研究，简单叙述了几何特征提取和代数特征提取，同时深入研究了基于DCT和PCA变换的人脸图像特征提取，从而实现是否对人脸识别系统识别率有所提高的研究；对于分类器的选择，本文对两种分类器进行了探讨，即最近邻分类器和B
recognitionprogram
- BP神经网络数字字符识别程序,可以识别各种手写字，识别率还不是很好-BP neural network digital character recognition program that can identify a variety of handwritten word recognition rate is not very good
bpn
- 用matlab实现的神经网络识别字母的程序，识别率达到100 。-Program of neural network regnizing letters using MATLAB, the accuracy is up to 100
GA-WNN
- 基于遗传算法和小波神经网络的语音识别系统，此系统识别率可以到较高-Based on genetic algorithm and neural network speech recognition system, this system can be to a higher recognition rate
face4
- 基于神经网络的人脸识别的一种算法，给出了计算识别率的方法-Neural network-based face recognition algorithm, gives the recognition rate calculation method
基于pca+lda+粗糙集+模糊神经网络的人脸识别
- 各种算法用于人脸识别，各个算法的识别率都不错，有lda,pca等算法(Various algorithms are used for face recognition, and the recognition rate of each algorithm is good. There are algorithms such as LDA, PCA and so on)
testandtrain
- 利用三层卷积神经网络识别信号星座图，准确识别8psk，16psk，32psk，64qpsk四种调制方式，在低信噪比的情况下已然有良好的识别率。(The three layer convolution neural network is used to identify signal constellation, and accurately identify four modulation modes of 8PSK, 16PSK, 32psk and 64qpsk. It has a good
pyDogVsCat
- 识别率85%，kaggle上有名的猫狗大战算法，可以很方便的查看分类结果。每一个epch需要22s左右（GTX1050Ti 4G）(The recognition rate is 85%. The famous dog and dog algorithm on kaggle is very convenient for us to see the classification results. Each epch needs about 22s (GTX1050Ti 4G))
BP神经网络
- 第一个m文件：构造、训练BP神经网络并计算其识别率；第二个文件将进行人脸检测。注意：orl人脸数据库需要在网上下载。(The function of the first m file is to construct and train the BP neural network and calculate its recognition rate. The second is the detection of face. Note: the ORL face database needs to
Tensorflow CNN
- 卷积神经网络识别手写数字，放在jupyter直接跑，99%识别率，已经和Tensorboard联通好了(Convolutional neural network recognizes handwritten numerals and runs directly on jupyter. The recognition rate is 99%. It has been connected with Tensorboard.)