Deep Learning for person re-identification with an RGB-D Camera in a top- view configuration

Degree Course: 
Electronic/Computer science engineer (Information-engineering track)
Location: 
Department of Information Engineering - UNIVPM
Description: 

Person re-identification is an important topic in retail, scene monitoring, human computer interaction, people counting, ambient assisted living and many other types of computer vision research. A dataset for person re-identification TVPR(Top View Person Re-Identification) has been previously built. This dataset uses an RGB-D camera in a top-view configuration that allows to extract anthropometric features for the recognition of people in view of the camera. The goal of this thesis is to find a method for person re-identification based on a number of significant features derived from both depth and colour.
In this context, the candidate will be required to study, investigate and apply machine-learning and deep learning approaches for person re-identification.

Aim: Applying machine learning and deep learning for recognising person passing under a camera installed in top-view configuration.

Supervisors: Marina Paolanti, Andrea Felicetti, Massimo Martini

Start: Sept, 2018

Expected graduation: Feb/July, 2019

Skills that will be acquired: Programming (Python, MATLAB)

Contacts: {e.frontoni,m.paolanti}@univpm.it

Academic Tutor: 
Emanuele Frontoni