Current page: Members->List of Members->Paolo Bottoni
 

Paolo Bottoni
University of Rome
Membership Number: 35
Address: Dipartimento di Scienze dell'Informazione, Via Salaria 113, 00198 Roma, Italy
Email: bottoni@dsi.uniroma1.it
Phone: +39 06 4991 8426
Fax: +39 06 8541842
URL: http://www.dsi.uniroma1.it/personal/bottoni/eng/index.html

Biographical Sketch
Paolo Bottoni, Research Associate

- Born in Milan (Italy), 1960

- Degree in Physics, University of Milan, 1988

- Ph.D in Computer Science, University of Turin, 1995

From 1988 through 1994 he collaborated with the Image Processing and Interpretation Group and with the Centre for Theoretical Medicine Studies of the University of Milan on the design of image interpretation and simulation environments for the study of biological systems. In 1992 he started collaborating with the Pictorial computing Laboratory of the University of Rome "La Sapienza" on the design of visual interactive tools. In 1994 he was visiting Ph.D. student at the Rank Xerox Research Centre in Grenoble. At the end of 1994 he joined the Department of Computer Science of the University of Rome, as a research associate. In 1995 he obtained the Ph.D. in Computer Science. In 2000 he was appointed Associate Professor His research interests evolved from Structural Pattern Recognition to the definition of models of visual interactive computing. The models are formally defined by means of Conditional Attributed Rewriting Systems, and implemented as agent-based systems. He is member of the ACM and IEEE CS

University of Rome - PCL: Pictorial Computing Laboratory
A central interest of the PCL is the definition of a formal model for image or diagram based activities, from person-to-person sketch-based communication to image understanding, to image retrieval and visual interactive systems. This formal model has at its core the notion of visual sentence, which links a pictorial sentence (a concrete image, graph, diagram, etc.) to its description (be it used to generate the pictorial component or be its interpretation). This allows the definition of visual languages in terms of invariant properties that the visual sentences have to satisfy, and the derivation of programs from these programs. The same theory is at the basis of a perspective on usability as agreement between the user and system descriptions of a same pictorial sentence. Such an agreement can be improved by the interactive exploration of the sentence. Studies on web sites and human-computer interaction based on this approach are performed within the PCL. In particular, activities related to vision and image processing concern interactive and automatic annotation of images for purposes as diverse as image query and retrieval, communication between domain experts, or reference to textual content. This becomes specially important in image retrieval, where we face a shift of focus from the "classical" problems of image understanding and pattern recognition, to the challenges related to human-user interaction in image retrieval systems, more specifically in query specification, image annotation, and result visualization. The goal is to provide tools that enable users to interact in a more semantically rich way, by having a correct interpretation of the tools and metaphors used in the program and of the at times intricate distance functions used in region-based image similarity search. A prototype system is currently being developed which, thanks to its plug-in architecture, acts as a testbed for new image segmentation algorithms, new features for describing the content of segmented regions, and new distance functions for image retrieval. Its user interface enables the user to explore and interact with the inner structures of the system and to examine and modify the image database in a visual and interactive way, even by adding new types of descriptions. Studies are being performed on the shape descriptions used in the MPEG-7 experiments, in order to determine their feasibility in image retrieval systems. Annotation becomes a fundamental tool for e-learning has taken place, where the PCL has designed and implemented the MultiCom II platform for managing and fostering communication between students, between students and tutors and making interaction during the learning sessions as easy as possible. Moreover, a new tool is under development to allow students to annotate following their personal styles (using different symbols, colours and signed versions) any document belonging to authorware within an e-distant course.


Site generated on Friday, 06 January 2006