Training Convolutional Networks with Web Images
This work addresses the challenge of building large-scale image datasets for deep learning, but it is incremental as it replicates an existing approach with minor variations.
The study replicated the ImageNet database using web images from four download strategies and applied AlexNet for classification, achieving results comparable to the original dataset.
In this thesis we investigate the effect of using web images to build a large scale database to be used along a deep learning method for a classification task. We replicate the ImageNet large scale database (ILSVRC-2012) from images collected from the web using 4 different download strategies varying: the search engine, the query and the image resolution. As a deep learning method, we will choose the Convolutional Neural Network that was very successful with recognition tasks; the AlexNet.