Section: Other Grants and Activities
National Projects
QUAERO
Participants : Matthijs Douze, Hervé Jégou, Frédéric Jurie, Arnau Ramisa, Harsimrat Sandhawalia, Cordelia Schmid, Jakob Verbeek.
Quaero is a French-German search engine project, supported by OSEO. It runs from 2008 to 2013 and includes many academic and industrial partners, including INRIA, CNRS, the universities of Karlsruhe and Aachen as well as LTU, Exalead and INA. LEAR/INRIA is involved in the tasks of automatic image annotation, clustering and search. See http://www.quaero.org for details.
ANR Project GAIA
Participants : Hervé Jégou, Cordelia Schmid, Oksana Yakhnenko.
GAIA is an ANR (Agence Nationale de la Recherche) “blanc” project that is running for 4 years starting October 2007. It aims at fostering the interaction between three major domains of computer science—computational geometry, machine learning and computer vision—, for example by studying information distortion measures. The partners are the INRIA project-teams Geometrica and LEAR as well as the University of Antilles-Guyane and Ecole Polytechnique.
ANR Project RAFFUT
Participants : Matthijs Douze, Hervé Jégou, Benoit Mordelet, Cordelia Schmid.
RAFFUT is an ANR (Agence Nationale de la Recherche) “audiovisuel et multimédia” project that started in December 2007 for two years. This project aims at detecting pirated videos. The main issues addressed by this project are 1) how to handle the scalability issues that arise when dealing with extremely large datasets ; 2) how to improve the accuracy of the search if the videos have suffered very strong attacks, as for example low-quality camcorded copies of movies. The partners are the company Advestigo (http://www.advestigo.com ) and LEAR. Advestigo is one of the leaders in the growing “digital asset management market”. Its technology is oriented towards video piracy, in particular for detecting fraudulent content on user-generated websites such as YouTube or DailyMotion.
ANR Project R2I
Participants : Moray Allan, Frédéric Jurie, Josip Krapac, Cordelia Schmid, Jakob Verbeek.
R2I (Recherche d'Image Interactive) is an ANR “masse de données et connaissances” project that is running for 3 years starting in January 2008. R2I aims at designing methods for interactive image search, i.e., to extract semantics from images, to cluster similar images and to enable user interaction via semantic concepts related to images. The final goal of this project is a system for interactive search, which can index about one billion of images and provide users with advanced interaction capabilities. The partners are the company Exalead, a leader in the area of corporate network indexing and a specialist for user-centered approaches, the INRIA project-team Imedia, a research group with a strong background in interactive search of multi-media documents, as well as LEAR and the University of Caen, both specialists in object recognition.
ANR Project SCARFACE
Participants : Frédéric Jurie, Cordelia Schmid, Gaurav Sharma.
Video surveillance systems are currently installed in many public areas. As their number increases, the manual analysis becomes impossible. The ANR project SCARFACE will develop tools to automatically access large volumes of video content in order to help investigators solve a crime. These tools will search videos based on human attributes, which describe the suspect. SCARFACE is a three-year project (2009-2011) funded by the French Research Agency (ANR). The participant are: the university of Lille (investigating legal issues of such tools), the INRIA-Imedia research group (developing image clustering and interactive retrieval algorithms), SpikeNet (developing technology for face/person detection), EADS (developing the experimental framework) as well as the University of Caen and the INRIA-LEAR group (developing vision algorithms for semantically characterizing persons in images and videos).
