Description: 本文将BP神经网络应用于汽车车牌的自动识别，在车牌图像进行预处理后的基础上，重点讨论了用BP神经网络方法对车牌照字符的识别。并附有部分识别代码。-BP neural network used in the automatic vehicle identification number plates, images of the license plates after pretreatment on the basis of discussion focused on the BP neural network method for car license plate character recognition. Accompanied by some identification code. Platform: |
Size: 857321 |
Author:小博士 |
Hits:
Description: 对具有随机噪声的二阶系统的模型辨识，进行标幺化以后系统的参考模型差分方程为： y(k)=a1*y(k-1)+a2*y(k-2)+b*u(k-1)+s(k) 式中，a1=0.3366,a2=0.6634,b=0.68,s(k)为随机噪声。由于神经网络的输出最大为1，所以，被辨识的系统应先标幺化，这里标幺化系数为5。采用正向建模(并联辨识)结构，神经网络选用3-9-9-1型，即输入层i，隐层j包括2级，输出层k的节点个数分别为3、9、9、1个；由于神经网络的最大输出为1，因此在辨识前应对原系统参考模型标么化处理，辨识结束后再乘以标么化系数才是被辨识系统的辨识结果。-of random noise with the second-order system model, per-unit system after the reference model differential equation : y (k) = y * a1 (k-1) a2 * y (k-2) b * u (k-1) s (k) - style, = 0.3366 a1, a2 = 0.6634, b = 0.68, s (k) as random noise. Because the neural network for a maximum output, therefore, the identification system should be per-unit, per-unit here coefficient of 5. Forward modeling (Parallel identification) structure, neural network-based selection 3-9-9-1, i input layer, hidden layer, including two j, k output layer to the number of nodes 3,9,9,1; The neural network the biggest losers up to one, in the original deal before Identification System Reference Model S Mody treatment, then multiplied by the end of Identification Standard Mody coefficient was recognition system is the ide Platform: |
Size: 874771 |
Author:孙荣超 |
Hits:
Description: 对具有随机噪声的二阶系统的模型辨识，进行标幺化以后系统的参考模型差分方程为： y(k)=a1*y(k-1)+a2*y(k-2)+b*u(k-1)+s(k) 式中，a1=0.3366,a2=0.6634,b=0.68,s(k)为随机噪声。由于神经网络的输出最大为1，所以，被辨识的系统应先标幺化，这里标幺化系数为5。采用正向建模(并联辨识)结构，神经网络选用3-9-9-1型，即输入层i，隐层j包括2级，输出层k的节点个数分别为3、9、9、1个；由于神经网络的最大输出为1，因此在辨识前应对原系统参考模型标么化处理，辨识结束后再乘以标么化系数才是被辨识系统的辨识结果。-of random noise with the second-order system model, per-unit system after the reference model differential equation : y (k) = y* a1 (k-1) a2* y (k-2) b* u (k-1) s (k)- style, = 0.3366 a1, a2 = 0.6634, b = 0.68, s (k) as random noise. Because the neural network for a maximum output, therefore, the identification system should be per-unit, per-unit here coefficient of 5. Forward modeling (Parallel identification) structure, neural network-based selection 3-9-9-1, i input layer, hidden layer, including two j, k output layer to the number of nodes 3,9,9,1; The neural network the biggest losers up to one, in the original deal before Identification System Reference Model S Mody treatment, then multiplied by the end of Identification Standard Mody coefficient was recognition system is the ide Platform: |
Size: 874496 |
Author:孙荣超 |
Hits:
Description: 本文将BP神经网络应用于汽车车牌的自动识别，在车牌图像进行预处理后的基础上，重点讨论了用BP神经网络方法对车牌照字符的识别。并附有部分识别代码。-BP neural network used in the automatic vehicle identification number plates, images of the license plates after pretreatment on the basis of discussion focused on the BP neural network method for car license plate character recognition. Accompanied by some identification code. Platform: |
Size: 3984384 |
Author:小博士 |
Hits:
Description: 利用BP神经网络进行字符识别，非加载图像训练识别，而是利用鼠标绘制字符进行工作。其中在BP神经网络的实现文件中隐层的节点数量计算公式在不同的情况下可能会不准确，需要进行修改。这是理论决定的，不是错误。-BP neural network for character recognition. No need to load images for training, but use the mouse to draw the characters. In the neural network implementation file, formula for the number of hidden layer nodes may be inaccurate under different circumstances, which needs to be modified. It is determined by the theory, not an error. Platform: |
Size: 210944 |
Author:Adrian |
Hits:
Description: ：基于脑电信号的身份识别是通过采集试验者的脑部信号来进行身份认证。对于同一个外部刺激或者主体在思考同一个
事件的时候，不同人的大脑所产生的认知脑电信号不同。选取与运动意识想象有关的电极后，分析不同个体在特定状况下脑
电的个体差异，采用以回归系数、能量谱密度、相同步、线性复杂度多种信号处理结合方法对运动想象脑电信号进行处理来
进行特征提取。组合多元特征向量并运用多层BP 神经网络对不同个体的脑电信号进行分类，并在不同的意识想象及不同数
据长度、不同的波段对试验者进行识别率验证分析。结果表明，不同运动想象的平均识别率均在80 以上，其中以想象舌头
运动的识别率较高，达到90.6 ，不同波段的识别率也表明意识想象的模式及相应波段对身份认别有较大的影响。-EEG-based identification to authenticate through the acquisition of experimental brain signals. For the same external stimuli, or the main thinking of the same
Event, different people s brains produced by cognitive EEG. Select imagine the electrodes and movement awareness, analysis of different individuals in a particular situation brain
Individual differences in electricity, the use of regression coefficients, the energy spectral density, phase synchronization, the linear complexity of a variety of signal processing combined with motor imagery EEG
For feature extraction. The combination of multiple feature vectors and the use of multi-layer BP neural network to classify the EEG signals of different individuals, and in a different sense of imagination and a different number of
Length, the band on the test to verify the analysis of the recognition rate. The results show that the average recognition rates of different motor imagery in more than 80 , which to imagine the tongue
The m Platform: |
Size: 551936 |
Author:王闯杰 |
Hits:
Description: 工神经网络（Artificial Neural Network）又称连接机模型，是在现代神经学、生物学、心理学等学科研究的基础上产生的，它反映了生物神经系统处理外界事物的基本过程，是在模拟人脑神经组织的基础上发展起来的计算系统，是由大量处理单元通过广泛互联而构成的网络体系，它具有生物神经系统的基本特征，在一定程度上反映了人脑功能的若干反映，是对生物系统的某种模拟，具有大规模并行、分布式处理、自组织、自学习等优点，被广泛应用于语音分析、图像识别、数字水印、计算机视觉等很多领域，取得了许多突出的成果。最近由于人工神经网络的快速发展，它已经成为模式识别的强有力的工具。神经网络的运用展开了新的领域，解决其它模式识别不能解决的问题，其分类功能特别适合于模式识别与分类的应用。多层前向BP网络是目前应用最多的一种神经网络形式-Artificial neural network (Artificial Neural Network) connection, also known as machine model, is based on interdisciplinary research in modern neurology, biology, psychology, etc. produced on, it reflects the fundamental processes of biological neural processing of external things, is in the simulation developed on the basis of human brain tissue computing system is constituted by a large number of processing units interconnected through an extensive network system, it has the basic characteristics of biological neural systems, to a certain extent reflects the number of reflecting the human brain function is simulation of certain biological systems, with massively parallel, distributed processing, self-organizing, self-learning, etc., are widely used in many areas of speech analysis, image recognition, digital watermarking, computer vision, and achieved many outstanding achievements . Recently due to the rapid development of artificial neural networks, it has become a powerful tool fo Platform: |
Size: 100352 |
Author:沈阳阳 |
Hits:
Description: 为实现合格和缺陷板栗的分级， 研究了 1 种基于 BP 神经网络与板栗图像特征的板栗分级方法。 试验以罗田板
栗为研究对象， 提取的颜色及纹理等 8 个特征值， 通过主成分分析提取相应的主成分得分向量构成模式识别的输入。 利
用 BP 神经网络方法建立了板栗分级模型。 试验结果表明， 在图像信息主成分因子数为 3， 中间层节点数为 12 时， 建立
的模型最佳， 模型训练时的回判率为 100 ， 预测时识别率达到了 91 .67 。 研究结果表明基于机器视觉技术的针对缺陷
板栗分级检测方法是可行的。- In order to realize grading of eligible and defected chestnut by using machine vision, a classification method
of chestnut was developed based on BP-ANN and image feature of chestnut. In this experiment, Luotian chestnuts were
used as experimental targets. Principal component analysis (PCA) was implemented on these feature variables from
eight eigen values including color parameters and veins characteristics parameters etc., and principal components (PCs)
vectors were extracted as the inputs of pattern recognition. Grading models were built by BP neural network. The test
result showed that when the number of principal component factor was three and the number of nodes of hidden layer
was twelve, the discriminating rate was as high as 100 in training set, and 91.67 in prediction set. The overall results
shows that it is feasible to discriminate chestnut quality with machine vision. Platform: |
Size: 1015808 |
Author:李祥龙 |
Hits:
Description: 包含特征值与特征向量的提取、训练样本以及最后的识别，Wtnjiha参数使用大量的有限元法求解偏微分方程，通过matlab代码，数学方法是部分子空间法，NyYSpzM条件基于人工神经网络的常用数字信号调制，车牌识别定位程序的部分功能。- Contains the eigenvalue and eigenvector extraction, the training sample, and the final recognition, Wtnjiha parameter Using a large number of finite element method to solve partial differential equations, By matlab code, Mathematics is part of the subspace, NyYSpzM condition The commonly used digital signal modulation based on artificial neural network, Part of the license plate recognition locator feature. Platform: |
Size: 6144 |
Author:ybdnfe |
Hits:
Description: 载入数字车牌，智能识别，检测返回识别的数字，此外，也可以单独对打开的图片一步一步进行图像预处理工作，但要注意，每一步工作只能执行一遍，而且要按顺序执行。
“256色位图转为灰度图”－“灰度图二值化”－“去噪”－“倾斜校正”－“分割”－“标准化尺寸”－“紧缩重排”。采用神经网络进行训练识别，识别率达到90 -Loading digital license plate, identification, detection returns the identification number, is also possible to open a separate image for image preprocessing step by step, but be careful, every step can only be d once, and according to the order execution. " 256-color bitmap into grayscale" - " grayscale binary" - " de-noising" - " inclination correction" - " Split" - " standard size" - " austerity rearrangement." Neural network trained to recognize, recognition rate of 90 Platform: |
Size: 3878912 |
Author:李红贤 |
Hits:
Description: 利用大量图像数据对卷积神经网络算法进行训练，通过卷积、池化、下采样以及全连接层训练后的卷积神经网络在图像识别精度越来越高。(By using a large number of image data to train the convolutional neural network algorithm, the accuracy of the image recognition is higher and higher by convolution, pooling, down sampling and full connection layer training.) Platform: |
Size: 3072 |
Author:aperture
|
Hits:
Description: 下载MNIST数据集(手写体数字0-9)后，搭建卷积神经网络，将输入的数据集经过一层一层的卷积，到最后计算交叉熵，用梯度下降算法去优化它，使它变得最小，这就训练出了权重和偏置量，识别的准确率为91%(Download the MNIST data set (handwritten number 0-9), build a convolutional neural network, the input data set by convolutional layers, finally calculate the cross entropy with the gradient descent algorithm to optimize it, so that it becomes the minimum, this training weight and bias, recognition accuracy rate 91%) Platform: |
Size: 11597824 |
Author:未来已来 |
Hits:
Search - number recognition by neural network