Projects

Bird song learner

A simple website to practice your bird identification skills! The website is coded in Python using Flask and uses xeno-canto's API to retrieve audios of singing birds to test you on your sound ID skills. I made it because there were no online resources where I could select a few bird species that I wanted to practice by themselves.

Tumorous area detection in breast cancer biopsy images

This project was part of my bachelor's thesis and it involved building a multi-resolution Neural Network called HookNet in order to perform semantic segmentation of microscopy images of breast tissue samples. This project was conducted within the Image Processing Group at the Polytechnic University of Catalonia and in collaboration with the Catalan Institute of Health.

Indoor navigation app

During this project, my team developed an indoor navigation app for patients in the Catalan Institute of Oncology. I trained Random Forests to predict the user's location within a 3000 square-meters area using the signal from 20 BLE beacons. We deployed and tested these models as an on-device predictor in an app using Flutter. The data was also uploaded to a cloud database in Firebaseused to feed a webapp to study the flows of people within the hospital.

Learn more about the project in the following video: