VOCdevkit_08-Jun-2007.rar，VOC2007官方开发工具包，包含各种python、matlab文件

opencv_3rdparty-ffmpeg-master_20191119.zip

%minFuncfprintf('CompilingminFuncfiles...\n');mexminFunc/lbfgsC.c%KPMfprintf('CompilingKPMfiles...\n');mex-IKPMKPM/repmatC.c%crfChainfprintf('CompilingcrfChainfiles...\n');mexcrfChain/mex/crfChain_makePotentialsC.cmexcrfChain/mex/crfChain_inferC.cmexcrfChain/mex/crfChain_lossC2.c

TUM的一个SLAM数据集，太大了所以拆成了三部分，每部分都是1积分，在国内下载实在是太慢了，我下了两天才下好。
数据集是https://vision.in.tum.de/data/datasets/rgbd-dataset/download的fr1/xyz

数据集++

车牌资源

我的环境：win10+python3.7+TensorFlow1.13.1gpu（conda的虚拟环境）能成功匹配安装版本

Throughaseriesofrecentbreakthroughs,deeplearninghasboostedtheentirefieldofmachinelearning.Now,evenprogrammerswhoknowclosetonothingaboutthistechnologycanusesimple,efficienttoolstoimplementprogramscapableoflearningfromdata.Thispracticalbookshowsyouhow.Byusingconcreteexamples,minimaltheory,andtwoproduction-readyPythonframeworks—Scikit-LearnandTensorFlow—authorAurélienGéronhelpsyougainanintuitiveunderstandingoftheconceptsandtoolsforbuildingintelligentsystems.You’lllearnarangeoftechniques,startingwithsimplelinearregressionandprogressingtodeepneuralnetworks.Withexercisesineachchaptertohelpyouapplywhatyou’velearned,allyouneedisprogrammingexperiencetogetstarted.*Explorethemachinelearninglandscape,particularlyneuralnets*UseScikit-Learntotrackanexamplemachine-learningprojectend-to-end*Exploreseveraltrainingmodels,includingsupportvectormachines,decisiontrees,randomforests,andensemblemethods*UsetheTensorFlowlibrarytobuildandtrainneuralnets*Diveintoneuralnetarchitectures,includingconvolutionalnets,recurrentnets,anddeepreinforcementlearning*Learntechniquesfortrainingandscalingdeepneuralnets

分为省份识别、字母识别、数字识别，对应我blog里面的内容分别使用

本程序根据训练好的网络文件ANN.mat预测新的数据文件，得到均方误差，并画出预测数据和原数据的对比图。