I'm currently working on the different projects of sugarcube: they are described on its website

In this project, an on-line meeting assistant provides real-time and natural interaction between the participants and printed papers. The main capabilities of the assistant are:

1. Control the slideshows by using printed paper to select the currently projected slide;
2. Add indexes and notes during the meeting;
3. Permit the participant to access additional information related to the content of a slide.

The assistant analyzes the video stream captured with a camera over the head of a participant and recognizes her fingers, as well as manipulated documents. Furthermore, the assistant beams augmented information over the paper, e.g. bookmarkers, a keyboard, a search text box, etc., in order to improve user experience.

The goal of this project is to manage large repositories of software in an enterprise. Usually, administrators create builds (e.g. packages of software) to install automatically on a PC when a new user joins the company or its technologies are updated. Currently, the definition of builds depends on administrators’ personal experience and instinct.
In this project, Enrico Bertini and I propose a solution based on the data collected by NEXThink, which provides technologies that monitor the network of an enterprise and gather the effective usage of applications. Thanks to those data, it has been possible to conceive BuildShape, a tool for helping administrators to manage efficiently the repositories of software, and to define optimal builds relatively to the needs of the enterprise and to the real usage of applications.

In this project, Enrico Bertini and I extend an existing technique for dynamic labeling, called Excentric Labeling and invented by Catherine Plaisant and Daniel Fekete. This technique allows users to visualize the labels associated with graphical objects that are enclosed in a zone around the mouse position. Visualized labels change when the user moves the mouse. In our Extended Excentric Labeling, we try to improve the original technique, by considering graphical objects included in high density areas; by automatically resizing the zone around the mouse position; and by providing statistics that summarize the content of the zone focused by users.

The ChartX API is an advanced graphical toolkit designed in cooperation with NEXThink. The charting toolkit is developed in Flex 2.0/ActionScript 3.0 for implementing completely customizable dashboards.
The chartx API allows developers to visualize various categories of charts, e.g. histograms, bars, lines, etc., and furthermore it offers a view inspired and emulating that of Google Finance.

FaericWorld is a multimedia browser, which intends to visualize a large collection of heterogeneous multimedia documents. The major issues of FaericWorld are: 1) the cross-media navigation; 2) the validation of annotations on media and of links between documents; 3) the abstraction and creation of new documents, by hyper-linking existing ones.
The archives imported, analyzed, and indexed by FaericWorld browser are composed of media (static documents, images, videos, or audio recordings), multimedia documents (slideshows or websites), people (person name, its publications, etc.) and events (conferences, meetings, etc.)

Inquisitor and JFriDoc are static documentic-centric browsers developed for the IM2 project. Inquisitor is a mono-modal browser for visualizing static documents and for correcting the results of document analyses. It has been developed as master thesis by Florian Evéquoz.
JFriDoc is a multimedia browser that allows to navigate in a meeting.

xed (extracting electronic documents) is a tool for extracting text, images, and graphics from PDF documents in a structured and handily format (an XML language called XCDF). The extracted data is analyzed, in order to extract the physical and the logical structures of the original document. The final expected output of xed is a logical document representation containing structured text, images, and graphics elements.

DocMining is a general framework that allows to design scenarios for document's image processing. The framework is the result of the collaboration between four academic partners (Laboratoire PSI, Laboratoire L3i, LORIA, and DIVA Group) and an industrial partner, leading the project (France Telecom R&D). The main issues of DocMining are the description and the execution of document analysis scenarios.
The task of DIVA group in the DocMining Consortium consists of providing the user interface for visualizing and validating the results of document analysis methods. This interface is called xmillum (XML illuminator) and is the result of Oliver Hitz's thesis. In this project, my task was to develop new tools and interactive visualizations.