Rita Pucci
PhD Researcher, University of Udine
Bio:
Rita Pucci is a research assistant in machine learning in computer vision. She is enrolled as a PostDoc at the University of Udine in the Machine Learning and Perception Lab. Her research is focused on the interdisciplinary field of Computer science combined with the study of animal behaviour in the wild. She has been exploring deep networks for the unsupervised tasks such as image reconstruction and image colourisation and the supervised tasks such as species classification and behaviour classification.
Workshop:
A Gentle Introduction to Capsule Networks
Abstract:
We will dive into the architecture of the CapsNet model proposed by Hinton et al. in 2017. It presents a new strategy for image classification and features clusterisation. In this class, we will cover the theoretical aspects of capsules and their application in CapsNet. We will also analyse the coding strategies applied to build the capsules in a classification model.
Workshop:
Automatic Old Image Colourisation
Abstract:
Starting from the capsule concept, we will analyse a work presented by the University of Udine for grayscale image colourisation. In this class, in the first part, we will introduce the theoretical concepts behind the colourisation task, and in the second part, we will go through the code to understand how to construct a new architecture for automatic colourisation and how to integrate capsules into it.
