On January 26th, 2016 the final dissertations of the candidates of the Ph.D. in Information Technology will be held at DEIB Seminar Room and will start at 1.30 pm:

Dr. Luca BAROFFIO – XXVIII Cycle
Visual Analysis Tools for Energy-Aware Heterogeneous Networks
Supervisors: Prof. Matteo Cesana, Prof. Marco Tagliasacchi

Abstract: The potential of the Internet of Things is leading to a paradigm shift with an ambitious long-term vision, in which battery-operated sensing nodes are empowered with sight capabilities and are able to accomplish complex visual analysis tasks. Unfortunately, this is out of reach with the current technology. This thesis addresses a comprehensive set of new methodologies to empower sensor nodes with vision capabilities comparable to those achievable by power-eager visual analysis systems. The key tenet is that most visual analysis tasks can be carried out based on a succinct representation of the image, which entails both global and local features, while it disregards the underlying pixel-level representation. Still, under severe energy and bandwidth constraints it is imperative to optimize the computation, the coding and the transmission of the visual features. On the coding side, this thesis tackles the problem by reversing the conventional Compress-Then-Analyze (CTA) paradigm in which compressed images are transmitted and processed by a central server. Instead, according to the proposed approach, image features are computed by sensing nodes, compressed with suitable coding algorithms, and delivered to final destinations in order to enable higher-level visual analysis tasks by means of either centralized or distributed algorithms, somewhat mimicking the processing of visual stimuli in the early visual system. The extraction of visual features is subject to tight application-dependent requirements regarding computational and memory resources and bandwidth/delay guarantees. Therefore, on the feature extraction side, this thesis addresses the design of energy-efficient tools for optimizing the operation of sensor nodes.

Dr. Lucio BIANCHI – XXVIII Cycle
A Unified Framework for Acoustic Scene Analysis, Synthesis, and Processing
Supervisor: Prof. Augusto Sarti

Abstract: In the past few years, the expectation of quality in acoustic signal processing has significantly grown. Research has focused on techniques to exploit multiple microphones and loudspeakers to analyze and to reproduce high-quality acoustic fields. This thesis proposes a new framework that acts as a fundamental structure for a wide range of acoustic signal processing tasks: it encompasses the most recent advances in signal processing theory (frame representation of signals, estimation theory, etc.) for a twofold purpose. 1) Design and the evaluation of new signal processing techniques based on conventional acoustic representations (plane waves, spherical waves, etc.); 2) introduction of a novel representation for acoustic fields, based on the idea of decomposing them into a set of spatially confined wave objects. This thesis pursues both purposes and gives rise to a promising global framework, in which acoustic fields are represented in a rigorous but extremely flexible and intuitive fashion.