Towards routine discovery from egocentric photostreams

Abstract

[en] When I was thinking what topic to choose and what work would it be. On the one hand I wanted it to be from the area that is particularly interesting and I can use my knowledge. On the other hand I wanted to study something new and exciting. This work exceeded my expectations. In this work I apply all my skills developed during my education in University of Barcelona to work on real scientific case. I do my research reading many papers on topic, develop methods that suit my case, brainstorm new ideas, spend hours coding and debugging them, spend days waiting calculations to finish and finally - obtain results I’m satisfied with. This work is about turning 106 days and 70GB of egocentric photostreams into a simple predictions if a particular day is a routine or no.