搜索资源列表
logistic-regression
- 采用随机梯度算法,来利用逻辑回归对训练集进行分类-Classify the training set by logistic regression
x
- 半导体数据训练集和测试集用来检测故障,使用MATLAB编程对数据采集-Semiconductor data training and test sets used to detect faults, using MATLAB programming for data collection
aam_matlab
- 主动形状模型(active appearance model)的matlab代码 训练集和测试集需要从附带的网址上下载或者根据自己的需求构造-matlab code for active appearance model, the training and test sets could be downloaded the website in readme.txt or constructed by yourself according to your problem
haarcascade_profileface
- 在opencv2.4中找到了侧脸训练集,分享一下,供大家使用-opencv: face detection
IRIS-use-RFclassification
- 用随机森林RF方法分类IRIS数据集,用一百个数据做训练集,五十个作为测试集,并统计出错误率,可直接运行-Classification method with random forests RF IRIS data set, using one hundred data to do training set, and fifty as a test set, and the statistical error rate, can be directly run
PlateIdentify
- 车牌字符识别C++程序 1.先打开一幅图片然后按照顺序灰度化、二值化、灰度拉伸、车牌定位、二值化、倾斜校正、字符分割、训练神经网络、识别字符。 2.测试图像存储在当前目录的img下。 3.测试集、训练集、目标向量均存储在img下的文本文件中。-C++ program license plate character recognition 1. Open a picture first and then follow the order of graying, binarizati
kenstone
- 用于划分训练集及预测集样本数,可用于偏最小二乘,支持向量机等训练集预测集的划分-ks pls svm
paris_sigg_release_v4.5.tar
- What makes Paris Look Like Paris 12年 sigaraph, 通过提取图像特征跟训练集对比得出地理位置-What makes Paris Look Like Paris 12 years sigaraph, by extracting image features contrast with the training set derived location
eigenface
- 用MATLAB实现人脸识别的源代码,包括训练集的实现,识别部分的书写。-Face recognition using MATLAB source code, including writing to achieve the training set, the identification section.
Program
- 模糊逻辑变精度粗糙集算法程序,数据挖掘故障诊断规则,27条训练集,8属性,5类故障。-Rough set of variable precision fuzzy logic algorithm procedures, data mining fault diagnosis rules, 27 train sets, eight properties, five fault.
lpr-system
- 1.先打开一幅图片然后按照顺序灰度化、二值化、灰度拉伸、车牌定位、二值化、倾斜校正、字符分割、训练神经网络、识别字符。 2.测试图像存储在当前目录的img下。 3.测试集、训练集、目标向量均存储在img下的文本文件中。-1. Open a picture first and then according to the order gray, binarization, gray stretching, license plate localization, binarization, t
KNN
- 自己编写的最近邻KNN算法,采用的距离是欧氏距离,附加了详细的中文注释,还有一个测试集和一个训练集- I have written KNN nearest neighbor algorithm, the distance is the Euclidean distance, additional detailed notes in Chinese, as well as a test set and a training set
MIT_persons_jpg
- MIT行人数据库,共924张行人图片,原来是PPM格式,很多人找不到转换工具,这里已转换成JPG。该数据库只含正面和背面两个视角,无负样本,未区分训练集和测试集。-MIT pedestrian dataset of JPG format, for pedestrian detection application.
anglecos
- 利用夹角余弦距离进行样本数据分类。实现步骤主要分为以下两部分:a、待测样品X与训练集里每个样品Xi的距离采用夹角余弦距离公式计算。b、循环计算待测样品和训练集中各已知样品之间的距离,找出距离待测样品最近的已知样品,该已知样品的类别就是待测样品的类别。-Using the sample data classification Angle cosine distance.Implementation steps are divided into the following two parts: a,
TestofFaceRecognition
- 基于BP神经网络的人脸识别系统的训练集源代码。-A face recognition system based on BP neural network the training set of source code.
rec
- Java实现将movielens各种规模数据的划分为测试集和训练集-Split movielens dataset to trainset and test set
AdaBoost
- Adaboost是一种迭代算法,其核心思想是针对同一个训练集训练不同的分类器(弱分类器),然后把这些弱分类器集合起来,构成一个更强的最终分类器(强分类器)。-Adaboost is an iterative algorithm, the core idea is the same training set for training different classifiers (weak classifiers), then these weak classifiers together to f
Transitive-Re-identification
- 行人再识别。人再次鉴定的准确性可以显著提高给定一个训练集,演示了外表的变化与非重叠的两个摄像头。我们测试时是否能保持这种优势直接标注的训练集并非对所有现场camera-pairs可用。给定的训练集捕捉相机A和B之间的对应关系和不同的训练集捕捉相机B和C之间的对应关系,传递鉴定算法(TRID)建议提供了一个分类器(A,C)对外观。该方法是基于统计建模和使用一个边缘化的推理过程。这种方法可以显著减少注释工作固有的学习系统。-Person re-identification accuracy can
svm
- 用matlab编写的SVM模型,用于进行植物分类,包括主程序、训练集和测试集。-Using matlab SVM model, used for plant taxonomy, including the main program, the training and test sets.
pca
- 人脸识别,输入200副人脸照片作为训练集,检测另外200副人脸图像,计算错误率。-Recognition, input 200 people face photo as the training set, testing additional 200 people face image to calculate the error rate.