NEoN project ("Neuron on Network") aimed at realizing a system for the analysis of electrical data coming form biological neural networks (BNN), such as the time stamps of neuron spikes. Data are retrievied from the BNNs thought a Multi Electrode Array (MEA) devices. In particular, NEoN wants to analyze the communication patterns among the neurons in order to classify their behavior (inhibitory or excitatory) and to build a map of their connections (epresented as an unweighted directed graph). This project started as a joint project between Politecnico di Milano and Università Statale Milano Bicocca.
This project aims at creating an automatic flow to generate modular hardware architectures that implement complex iterative algorithms, especially those in the multimedia domain. Starting from a high-level description of an algorithm, the proposed flow analyzes the dependences between subsequent iterations, and it finds the circuit that best exploits the data locality and minimizes the resource usage. At present, the proposed methodology has been successfully applied to many multimedia problems, and in particular to the parallelization of Chambolle, an important algorithm for the optical flow estimation that is characterized by a complex dependency schema. The work has produced two conference publications (DATE 2011 – DAC 2013) and a journal article (IEEE Design and Test). This is a joint project between Politecnico di Milano and École Polytechnique Fédérale de Lausanne (EPFL).
FASTER is an european project that aims to facilitate the use of reconfigurable technology by providing a complete methodology that enables designers to easily implement and verify applications on platforms with general-purpose processors and acceleration modules implemented in the latest reconfigurable technology. Reconfigurable logic allows the definition of new functions to be implemented in dynamically instantiated hardware units, combining adaptivity with hardware speed and efficiency. This project is financed by European Commision.
Aim of this project is to create a multi-camera system to gather data from different (and possibly moving) points-of-view, in order to build a 3D model of the target scene. The system is able to perform run-time preprocessing, filtering and analysis on the raw images flow, thanks to a set of synchronized FPGA boards (one for each camera) that guarantee performance and flexibility. The camera provides an enabling technology for such applications as high spatial and temporal resolution movies, as well as highly-interactive devices for video communication. At present, the first working prototype has been successfully employed in important events such as Balelec and Montreux Jazz Festival. This project is owned by the Ecole Polytechnique Fédérale de Lausanne (EPFL).