PhD position on «Machine Learning on non-linear Manifolds: Application to Human Body Analysis»

We are opening a PhD position on «Machine Learning on non-linear Manifolds: Application to Human Body Analysis» in University of Florence and IMT Lille Douai/CRIStAL (UMR CNRS) and are seeking candidates for a selection which will take place shortly.
Title - Machine Learning on non-linear Manifolds: Application to Human Body Analysis
Description - Analysis of the movement of the human body is central in problems like human action recognition, human behaviour understanding, human-object and human-human interaction, avatar animation, emotion detection, gait recognition, etc. A recent trend in this area is that of investigating such aspects using RGB-D sensors that jointly capture photometric and depth data (the body skeleton is typically also available from these data). Extracting representations of such dynamic sequences of RGB-D frames often results in descriptors with an underlying structure, which lay in a non-Euclidean space.
In this PhD theme proposal, we aim to investigate methods for representing human movements in RGB-D sequences. In particular, we are interested in matrix manifold solutions that shown the potential to effectively manage the non-linearity of such data. In addition, such geometric data are large and complex, and are natural targets for machine learning techniques. In many applications, Deep neural networks have been recently proven to be powerful tools, but these tools have been most successful on data with an underlying Euclidean or grid-like structure, and in cases where the invariances of these structures are built into networks used to model them. We also aim to investigate emerging techniques attempting to generalize (structured) deep neural models to non-Euclidean domains such as graphs and manifolds.
Requested expertise - Strong preference will be given to candidates with experience in Computer Vision and Pattern Recognition, and a good knowledge of written and spoken English. The following expertise is especially considered:
Excellent record of academic and/or professional achievement
Very good English skills, written and spoken (B2 level appreciated). Good written and spoken communication skills in Italian will be appreciated
Strong mathematical skills
Solid programming skills
Strong interests in one or more of the involved research areas (machine learning, computer vision, high performance computing).
The position is for a duration of three years. Eligibility criteria:
A Master degree in Computer Science, Mathematics, Physics or closely related disciplines
Appropriate experience to undertake PhD research in the specified area.
Salary - The grant for the PhD student will correspond to a standard salary of PhD student. The PhD program is expected to start in November 2017.
This PhD program will take place at the Media Integration and Communication Center (MICC) at University of Florence (UNIFI), and 3D SAM CRSItAL (UMR CNRS) at IMT Lille Douai. MICC is a pioneer laboratory in Computer Vision and Multimedia, and 3D SAM has a long experience on the application of Riemannian geometry in biometrics. The PhD student will work under the supervision of
Prof. Stefano Berretti (MICC/UNIFI), Prof. Pietro Pala (MICC/UNIFI), and Prof. Mohamed Daoudi (IMT Lille Douai, CRIStAL).
Candidates should send a CV, a cover letter and the grades obtained during the last two years to stefano.berretti-AT-unifi.it
Opening of the official application is expected in end June / beginning of July, 2017.