A Step by Step Tutorial for Image Classification using CNN as a Feature Extractor in MATLAB

#matlab #cnn #featureextractor #deeplearning
This example shows how to extract learned image features from a pre-trained 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.