文件名称:ImageClassification-master
介绍说明－－下载内容均来自于网络，请自行研究使用
在这个项目中，我们的目标是建立一个识别和大小231x231图像呈现对象分类系统。我们得到了一组训练图像，每四个标签之一：1飞机；汽车2；3马，否则。我们提供了两个特点：一是方向梯度直方图（HOG），其尺寸为5408；另一个是overfeat ImageNet美国有线电视新闻网的特点，其尺寸37000。关于测试图像，我们只给出了每个图像的功能，没有标签，结果判断由平地机。我们的目标是提供二进制和多个预测。平衡错误率（BER）是我们的性能评估。为了解决这个问题，我们首先减少PCA的问题的维数，处理不平衡数据集，通过向上采样或下采样，去除异常值，通过无监督学习，如k-均值和EM算法。其次，我们使用ML方法，如二进制和多项式logistic回归，二进制和多项式SVM和神经网络。多项式SVM的证明有最好的结果。最后，我们在100分中得了92分。-In this project, our goal was to build a system that recognizes and classifies the object present in an image of size 231x231. We were given a set of training images each with one of four labels: 1 for airplanes 2 for cars 3 for horses 4 otherwise. We were provided with two sets of features: one is Histogram of Oriented Gradients (HOG), which has dimension of 5408 the other one is OverFEAT ImageNet CNN Features, which has dimension of 37,000. Concerning the test images, we were only given the features of each image without label, and the results to be judged by the grader. Our goal was to provide binary and multiple predictions. The Balanced Error Rate (BER) was our performance uator. To solve the problem, we firstly reduced the problem’s dimensionality by PCA, dealt with imbalanced datasets through up-sampling or down-sampling, and removed outliers through unsupervised learning such as K-Means and EM Algorithm. Secondly, we applied ML methods such as Binary and Multinomial Logist
(系统自动生成,下载前可以参看下载内容)
下载文件列表
ImageClassification-master
..........................\ADASYN.m
..........................\DeepLearnToolbox-master
..........................\.......................\NN
..........................\.......................\..\nnapplygrads.m
..........................\.......................\..\nnbp.m
..........................\.......................\..\nnchecknumgrad.m
..........................\.......................\..\nneval.m
..........................\.......................\..\nnff.m
..........................\.......................\..\nnpredict.m
..........................\.......................\..\nnsetup.m
..........................\.......................\..\nntest.m
..........................\.......................\..\nntrain.m
..........................\.......................\..\nnupdatefigures.m
..........................\Model Analysis.xlsx
..........................\README.md
..........................\SMO.m
..........................\ber.m
..........................\binClassPredLogistReg.m
..........................\binClassPredSVM.m
..........................\featureEngineering.m
..........................\gaussianBasis.m
..........................\kmeansUpdate.m
..........................\learnBestModel.m
..........................\learnDistribTestError.m
..........................\learnKmeans.m
..........................\learnLogisticRegression.m
..........................\learnLogisticRegressionOnevsAll.m
..........................\learnMultiClassSVM.m
..........................\learnMultinomLogisticRegression.m
..........................\learnNeuralNetworks.m
..........................\learnPenLogisticRegression.m
..........................\learnPredMulti.m
..........................\learnSVM.m
..........................\linearKernel.m
..........................\lmsvd.m
..........................\logisticRegression.m
..........................\main.m
..........................\multiClassPredLogistReg.m
..........................\multiClassPredSVM.m
..........................\multinomClassPredLogistReg.m
..........................\multinomLogisticRegression.m
..........................\penLogisticRegression.m
..........................\polyKernel.m
..........................\polybasis.m
..........................\pred_binary.mat
..........................\pred_multiclass.mat
..........................\rbfKernel.m
..........................\readTestData.m
..........................\readTrainData.m
..........................\report.pdf
..........................\result.m
..........................\searchGamma.m
..........................\setSeed.m
..........................\solveImbalaceData.m
..........................\splitClass.m
..........................\splitTrainValid.m
..........................\standartscore.m
..........................\test.m