was born in Tesero, Trento (Italy) in 1959. She received her degree in Electronic Engineering from the University of Padova, ITALY, in 1987. Ensuing her graduation she worked on radar electronics for fighters jets before joining the VLSI research group at IRST in 1989. Since 1990 she is an active member of the Computer Vision Group.

hera.itc.it:3003/~mich/

Abstract 1:

COMPASS
COMputer Aided Search System

R. Brunelli and O. Mich

COMPASS is a distributed application for content-based image retrieval using remote databases. The system is configured as a client-server architecture in which a client application can submit a user query to multiple image servers.
Statistical methods are used to quantify image descriptor effectiveness, to simplify user queries and to select an image comparison strategy which optimizes search speed while preserving answer quality.
The COMPASS system can be used for two main activities: to browse still image databases, and to search image databases for pictures similar to a set of photos or drawings provided by the user.
The COMPASS system has been developed in the context of the Esprit Project No. 25389: OPAL.

Abstract 2:

VIDEO: Video and Image Data Exploration and Organization

R. Brunelli and O. Mich

The future of digital telecommunications is strictly linked to systems which are able to manage multimedia databases providing type specific interfaces. The purpose of the VIDEO (Video and Image Data Exploration and Organization) system, developed at ITC-IRST is to provide a leading edge tool to cope with image and video data in type and application specific ways.
A set of algorithms to partition a video stream into shots has been developed.
The algorithms are able to detect abrupt changes of scenes (cuts) and gradual transitions (edits) as well. Camera operations are also estimated and provide additional information useful in characterizing video structure. The resulting shots are then described at different level of detail by appropriately chosen key-frames.
Building on the functionalities of the developed algorithms an interactive graphical interface has also been developed whose features are: compact layout to optimize video stream presentation by representative key-frames (linear and lighttable visualization); multiple detail levels for the key-frame representation; simultaneous multiple video streams presentation; possibility of simultaneous multiple shots playing with VCR-like functionalities; shot annotation capabilities; detection of similar shots (e.g. to speed-up the annotation process).
Significant applications of the resulting system are: support for video annotation; video abstracting for Video on Demand systems; video browsing.