Autoencoders for Content-based Image Retrieval with Keras and TensorFlow - PyImageSearch

In this tutorial, you will learn how to use convolutional autoencoders to create a Content-based Image Retrieval system (i.e., image search engine) using Keras and TensorFlow. The tutorials were a big hit; however, one topic I did not touch on was Content-based Image Retrieval (CBIR), which is really just a fancy academic word for image search engines. Image search engines are similar to text search engines, only instead of presenting the search engine with a text query, you instead provide an image query -- the image search engine then finds all visually similar/relevant images in its database and returns them to you (just as a text search engine would return links to articles, blog posts, etc.). I'll show you how to implement each of these phases in this tutorial, leaving you with a fully functioning autoencoder and image retrieval system. To learn how to use autoencoders for image retrieval with Keras and TensorFlow, just keep reading!