Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

EXAMPLE 2: Artificial Intelligence

Using AlexNet CNN in MATLAB

Introduction

AlexNet is a pretrained Convolutional Neural Network (CNN) that has been trained on approximately 1.2 million images from the ImageNet Dataset (http://image-net.org/index).

AlexNet was created in 2012 and is popularly called AlexNet after the name of the first author in the paper “ImageNet Classification with Deep Convolutional Neural Networks”  by Alex Krizhevsky, Ilya Sutskever, and Geoffrey E. Hinton, which described the CNN.

AlexNet solves the problem of image classification where the input is an image of one of 1000 different classes (e.g. cats, dogs etc.) and the output is a vector of 1000 numbers. The i-th element of the output vector is interpreted as the probability that the input image belongs to the it-h class. Therefore, the sum of all elements of the output vector is 1.

Objectives

· Introduce AlexNet in MATLAB.

· Use analyzeNetwork to inspect AlexNet in Matlab. Study different layers of AlexNet.

· Visualise Features of a CNN using deepDreamImage. Observe the features and make conclusions.

· Use AlexNet with transfer learning to adapt it to a new dataset. Use Deep Network Designer for this purpose.

· Relate practical work to the theory learnt in lectures including the types of AlexNetlayers, training parameters and how to set them, training, validation and testing subsets, etc.

· Analyse the training process graphs and the results and make appropriate conclusions.

· Use another pre-trained CNN, ResNet, to extract features from new images and train a different classifier using these features.

Instructions

· Follow the commands in the AlexNet lecture presentation. Investigate different command options for those commands that you use. Relate what you do to the theory from the lectures to achieve the objectives listed for this lab.

· Apply transfer learning to your own image data set. Create the dataset by taking photographs of several different objects with your own phone camera or by downloading suitable photos from Internet.

Extract Image Features Using Pretrained Network

This example uses:

· Deep Learning Toolbox

· Deep Learning Toolbox Model for ResNet-18 Network

· Statistics and Machine Learning Toolbox

This example shows how to extract learned image features from a pretrained convolutional neural network and use those features to train an image classifier. Feature extraction is the easiest and fastest way to use the representational power of pretrained deep networks. For example, you can train a support vector machine (SVM) using fitcecoc (Statistics and Machine Learning Toolbox™) on the extracted features. Because feature extraction only requires a single pass through the data, it is a good starting point if you do not have a GPU to accelerate network training with.

Load Data

Unzip and load the sample images as an image datastore. imageDatastore automatically labels the images based on folder names and stores the data as an ImageDatastore object. An image datastore lets you store large image data, including data that does not fit in memory. Split the data into 70% training and 30% test data.

unzip('MerchData.zip');

imds = imageDatastore('MerchData','IncludeSubfolders',true,'LabelSource','foldernames');

[imdsTrain,imdsTest] = splitEachLabel(imds,0.7,'randomized');

There are now 55 training images and 20 validation images in this very small data set. Display some sample images.

numTrainImages = numel(imdsTrain.Labels);

idx = randperm(numTrainImages,16);

figure

for i = 1:16

    subplot(4,4,i)

    I = readimage(imdsTrain,idx(i));

    imshow(I)

end

Load Pretrained Network

Load a pretrained ResNet-18 network. If the Deep Learning Toolbox Model for ResNet-18 Network support package is not installed, then the software provides a download link. ResNet-18 is trained on more than a million images and can classify images into 1000 object categories, such as keyboard, mouse, pencil, and many animals. As a result, the model has learned rich feature representations for a wide range of images.

net = resnet18

net =

  DAGNetwork with properties:

 

         Layers: [71x1 nnet.cnn.layer.Layer]

    Connections: [78x2 table]

     InputNames: {'data'}

    OutputNames: {'ClassificationLayer_predictions'}

Analyze the network architecture. The first layer, the image input layer, requires input images of size 224-by-224-by-3, where 3 is the number of color channels.

inputSize = net.Layers(1).InputSize;

analyzeNetwork(net)

Extract Image Features

The network requires input images of size 224-by-224-by-3, but the images in the image datastores have different sizes. To automatically resize the training and test images before they are input to the network, create augmented image datastores, specify the desired image size, and use these datastores as input arguments to activations.

augimdsTrain = augmentedImageDatastore(inputSize(1:2),imdsTrain);

augimdsTest = augmentedImageDatastore(inputSize(1:2),imdsTest);

The network constructs a hierarchical representation of input images. Deeper layers contain higher-level features, constructed using the lower-level features of earlier layers. To get the feature representations of the training and test images, use activations on the global pooling layer, 'pool5', at the end of the network. The global pooling layer pools the input features over all spatial locations, giving 512 features in total.

layer = 'pool5';

featuresTrain = activations(net,augimdsTrain,layer,'OutputAs','rows');

featuresTest = activations(net,augimdsTest,layer,'OutputAs','rows');

whos featuresTrain

  Name                Size              Bytes  Class     Attributes

  featuresTrain      55x512            112640  single

Extract the class labels from the training and test data.

YTrain = imdsTrain.Labels;

YTest = imdsTest.Labels;

Fit Image Classifier

Use the features extracted from the training images as predictor variables and fit a multiclass support vector machine (SVM) using fitcecoc (Statistics and Machine Learning Toolbox).

classifier = fitcecoc(featuresTrain,YTrain);

Classify Test Images

Classify the test images using the trained SVM model using the features extracted from the test images.

YPred = predict(classifier,featuresTest);

Display four sample test images with their predicted labels.

idx = [1 5 10 15];

figure

for i = 1:numel(idx)

    subplot(2,2,i)

    I = readimage(imdsTest,idx(i));

    label = YPred(idx(i));

    imshow(I)

    title(char(label))

end

Calculate the classification accuracy on the test set. Accuracy is the fraction of labels that the network predicts correctly.

accuracy = mean(YPred == YTest)

accuracy = 1

Train Classifier on Shallower Features

You can also extract features from an earlier layer in the network and train a classifier on those features. Earlier layers typically extract fewer, shallower features, have higher spatial resolution, and a larger total number of activations. Extract the features from the 'res3b_relu' layer. This is the final layer that outputs 128 features and the activations have a spatial size of 28-by-28.

layer = 'res3b_relu';

featuresTrain = activations(net,augimdsTrain,layer);

featuresTest = activations(net,augimdsTest,layer);

whos featuresTrain

  Name                Size                      Bytes  Class     Attributes

  featuresTrain      28x28x128x55            22077440  single

The extracted features used in the first part of this example were pooled over all spatial locations by the global pooling layer. To achieve the same result when extracting features in earlier layers, manually average the activations over all spatial locations. To get the features on the form N-by-C, where N is the number of observations and C is the number of features, remove the singleton dimensions and transpose.

featuresTrain = squeeze(mean(featuresTrain,[1 2]))';

featuresTest = squeeze(mean(featuresTest,[1 2]))';

whos featuresTrain

  Name                Size             Bytes  Class     Attributes

  featuresTrain      55x128            28160  single

Train an SVM classifier on the shallower features. Calculate the test accuracy.

classifier = fitcecoc(featuresTrain,YTrain);

YPred = predict(classifier,featuresTest);

accuracy = mean(YPred == YTest)

accuracy = 0.9500

Both trained SVMs have high accuracies. If the accuracy is not high enough using feature extraction, then try transfer learning instead. For an example, see Train Deep Learning Network to Classify New Images. For a list and comparison of the pretrained networks, see Pretrained Deep Neural Networks.