FRANCESCO BELLOCCHIO S CURRICULUM VITAE ET STUDIORUM April 2011 Index Personal details and education 1 Research activities 2 Teaching and tutorial activities 3 Conference organization and review activities 4 Publications 5 Francesco Bellocchio www.dti.unimi.it/bellocchio francesco.bellocchio@unimi.it
1. PERSONAL DETAILS AND EDUCATION Date and place of birth: March 4, 1980, in Bollate, Italy. Nationality: Italian. April 2007: Master in Computer Science at Università degli Studi di Milano, Italy, 104/110. Master s Thesis: Costruzione real-time di modelli 3D acquisiti mediante scanner ( Real-time computation of models acquired by 3D scanners ). Supervisor: Prof. N. Alberto Borghese. April 2007- December 2007: research activities at Laboratory of Applied Intelligent Systems of Department of Computer Science, Università degli Studi di Milano, Italy. The research has been devoted to real-time 3D surface reconstruction. July 2008: International Computer Vision Summer School (ICVSS 2008), organized by Department of Computer Science of Università degli Studi di Catania, Italy. December 2007- March 2011: Ph.D. in Computer Science at Università degli Studi di Milano, Italy. Ph.D. dissertion (defended on March 25, 2011): Online Hierarchical Models for Surface Reconstruction. Supervisor: Prof. Vincenzo Piuri He is currently post-doc researcher at Department of Information Technology, Università degli Studi di Milano, Italy. The research activities are focused to 3D surface reconstruction and computational intelligence techniques for regression problems. 1
2. RESEARCH ACTIVITIES The research activities have been focused on the following topics: systems for the computation of digital models from real objects, techniques for reconstruction of tridimensional surfaces and Machine Learning paradigms for regression problems. The work, mainly, has been developed following two directions: the first one, more practical, was about the use of specialized techniques for real-time computation of tridimensional models from a set of measurements realized on the physical objects. The second one, more theoretical, was about the analysis of computational intelligence techniques for the function approximation problem. In particular, the following arguments have been considered: 3D reconstruction systems: the study has been devoted to the analysis of different available technologies for the acquisition of geometrical features of physical objects. The analysis has considered the techniques implemented on these systems considering also the hardware used to perform the measurement. Particular attention has been given to those systems whose output is a set of 3D coordinates of points sampled on the physical object surface [7]. Moreover the study has been focused on the parallel computing devices that can be used for processing the acquired data. In fact, the operations needed for the computation of a 3D model are generally characterized by computations that can be performed in parallel and then parallel hardware can be an important resource for real-time models calculus. On-line learning for surface computation: the objective of this study has been the analysis of techniques for on-line approximation of surfaces from data gathered by 3D digital system. The research has been devoted to the use of robust computational models for domains in which the a priori knowledge is limited. The work has been focused on the development of paradigms in which the configuration phase of the parameters is performed without the entire knowledge of the input dataset, taking into consideration the robustness of the solution [2, 6, 8, 9]. Computational intelligence techniques for regression problem: the objective of this research activity has been the study of Machine Learning paradigms (Artificial Neural Network and Support Vector Machines) for regression problem. The work has been focused on the development of new techniques for function approximation from a limited number of samples that allows a higher robustness degree and a good efficiency in the strategy of the parameters configuration. In particular hierarchical models for multi-scale approximation have been considered [1, 3, 4, 5, 8]. The preliminary results of this study have been presented in a paper awarded as the IEEE IJCNN runner-up best paper award for 2010 [3]. 2
3. TEACHING AND TUTORIAL ACTIVITIES 3.1 Teaching Tutor for Fundamentals of Computer Science course for undergraduated students, Università degli Studi di Milano, Italy (Academic Years 2007/08, 2008/09 and 2009/10). Tutor for Operating Systems course for undergraduated students, Università degli Studi di Milano, Italy (Academic Year 2010/11). 3.2 Tutorials February, 2008: 3D Scanners: state of the art presented at Department of Computer Science, Università degli Studi di Milano, Italy. December, 2008: Scanner tridimensionali: copie virtuali da oggetti reali (Tridimensional scanners: virtual copies from real objects) presented at Department of Information Technology, Università degli Studi di Milano, Italy. He co-authored the tutorial 3D Scanners: state of the art (V. Piuri, F. Bellocchio, and S. Ferrari), presented at I2MTC 2009 (International Instrumentation and Measurement Technology Conference), 5-7 May 2009, Singapore. June, 2009: Hierarchical Support Vector Machine presented at Department of Computer Science, Università degli Studi di Milano, Italy. 3
4. CONFERENCE ORGANIZATION AND REVIEW ACTIVITIES 4.1 Activities in international conferences International Program Committee Member for INISTA 2010 (2010 IEEE International Symposium on INnovations in Intelligent SysTems and Applications), 21-24 June 2010, Kayseri & Cappadocia, Turkey. Publicity Chairs for EESMS 2010 (2010 IEEE Workshop on Environmental, Energy, and Structural Monitoring Systems), 9 September 2010, Taranto, Italy. Publicity Chairs for CIMSA 2010 (2010 IEEE International Conference on Computational Intelligence for Measurement Systems and Applications), 9 September 2010, Taranto, Italy. International Program Committee Member for INISTA 2011 (2011 IEEE International Symposium on INnovations in Intelligent SysTems and Applications), 15-18 June 2011, Instanbul, Turkey. Publicity Chairs for EESMS 2011 (2011 IEEE Workshop on Environmental, Energy, and Structural Monitoring Systems), 28 September 2011, Milan, Italy. 4.2 Journal reviews Mathematics and Computers in Simulation, Elsevier. 4.3 Conference reviews INISTA 2010 (2010 IEEE International Symposium on INnovations in Intelligent SysTems and Applications), 21-24 June 2010, Kayseri & Cappadocia, Turkey. WCCI 2010 (2010 IEEE World Congress on Computational Intelligence), 18-23 July 2010, Barcelona, Spain. INISTA 2011 (2011 IEEE International Symposium on INnovations in Intelligent SysTems and Applications), 15-18 June 2011, Instanbul, Turkey. IJCNN 2011 (2011 IEEE International Joint Conference on Neural Networks), July 31- August 5 2011, San Jose, California, USA. 4
5. PUBLICATIONS 5.1 Journal papers [1] F. Bellocchio, S. Ferrari, V. Piuri, and N. A. Borghese, A Hierarchical Scheme for Multi-Kernel Support Vector Regression, submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011. [2] S. Ferrari, F. Bellocchio, V. Piuri, and N. A. Borghese, A Hierarchical RBF On-line learning Algorithm for Real-time 3D Scanner, IEEE Transactions on Neural Network, vol. 21, no. 2, pp 275-285, 2010 5.2 Conference papers [3] S. Ferrari, F. Bellocchio, V. Piuri, and N. A. Borghese, A Multi-scale Support Vector Regression, in Proceeding of IJCNN 2010 (IEEE International Joint Conference on Neural Networks), July 2010. (IJCNN runner-up best paper award for 2010) [4] F. Bellocchio, N. A. Borghese, S. Ferrari, and V. Piuri, Kernel Regression in HRBF Networks for Surface Reconstruction in Proceedings of HAVE 2008 (IEEE International Workshop on Haptic Audio and Visual Environments and Games), October 2008. [5] S. Ferrari, F. Bellocchio, N. A. Borghese, and V. Piuri, Refining Hierarchical Radial Basis Function Networks in Proceedings of HAVE 2007 (IEEE International Workshop on Haptic Audio and Visual Environments and Games), pp. 166-170, October 2007. [6] F. Bellocchio, S. Ferrari, V. Piuri, and N. A. Borghese, On-line Training of Hierarchical RBF in Proceedings of IJCNN 2007 (IEEE International Joint Conference on Neural Networks), pp. 2159-2164, August 2007. 5.3 Book chapters [7] F. Bellocchio, and S. Ferrari, 3D Scanner, state of the art, in Depth Map and 3D Imaging Applications: Algorithms and Technologies, IGI Global (accepted). 5.4 Ph.D. Thesis [8] Francesco Bellocchio, Online Hierarchical Models for Surface Reconstruction, Ph.D. Thesis, Università degli Studi di Milano, Italy, Archivio Istituzionale della Ricerca, 2011. 5.5 M.S. Thesis [9] Francesco Bellocchio, Costruzione real-time di modelli 3D acquisiti mediante scanner, ( Real-time computation of models acquired by 3D scanners ), M.S. Thesis, Università degli Studi di Milano, Italy, Biblioteca del Dipartimento di Scienze dell Informazione, Milan, Italy, 2007. 5