This project was created at University as part of my final year project in order to learn more about the use of Convolutional Neural Networks and their efficacy within the branch of computer vision known as Content Based Image Retrieval.
You can access the PDF write-up here.
The datasets used within this project were the Omniglot dataset and a modifed version of the Stanford Dogs dataset.
The network architecture is based on that of Gregory Koch et al. with the original paper is listed here.