Using MATLAB, several image processing algorithms were used to augmented the Stanford Cars dataset - so that improved image classification accuracy could be attained. Methods used include image transformations, colour filters, edge detection, and seam carving. Achieved improved classification accuracy of 99%. Deep learning model used was an AlexNet Convolutional Neural Network (CNN) - which was fine tuned on the aforementioned training data

Using MATLAB, a series of raw Electroenphalograms (EEGs), and matrices of coherence between sets of EEGs, were converted into images, before being used to generated a training / testing set for use in a foundational (pre-trained) AlexNet CNN - fined tuned with the aforementioned data. The work investigated how classification accuracy is sensitive to different splits in training and testing data - in other wordws, intra and inter-patient.