Patrick Sambrook Undergraduate Dissertation 2017/18

Deep Learning for Image Classification using Tensorflow

Supervised by R.Gaizauskas

Abstract

The field of computer vision has seen rapid change in recent years, as previously limited Deep Learning techniques have overtaken the performance of once dominant machine-learning methods.

This project aims to implement Supervised Deep Learning in the form of a Convolutional Neural Network, using Google's recently released, yet hugely popular machine learning library, Tensorflow.

The project will test the Convolutional Neural Network implementation with benchmark computer vision competitions, and explore the effects of changes to its architecture.