Search

 

PressCNRS international magazine

Table of contents

Ipal

A High-Tech Partnership

In Singapore, Southeast Asia's small but very high-tech country, French and Singaporean researchers are inventing tomorrow's technologies at the IPAL joint laboratory.

 

portable

© IPAL Singapour

Currently being tested in Singapore, the newest version of SnapToTell works on most cell phones (not just smartphones).


 

This is your first visit to Singapore. You explore the city, cross its numerous bridges, and remain puzzled by the grandeur of the Theaters on the Bay. You decide to pull out your cell phone, point it towards the buildings, and take a snapshot. A second later, you receive a text message informing you that the aluminum sunshades adorning this complex–opened in 2002–change their angle in accordance with the sun's movement. You also learn that this “skin” of automated shutters has earned the building its nickname: the durian, a prickly-shelled fruit highly appreciated in Asia.

This is one application of the many ongoing research programs at Singapore's IPAL (Image Perception Access and Language) joint laboratory, a collaborative lab between French1 and Singaporean researchers.2

An agreement renewing the existing collaboration through the IPAL joint lab was signed on January 22nd, 2007 at the French Embassy, a few steps from Singapore's internationally renowned Botanic Gardens. This signature gives an official frame to the lab which was created in 1998, based on a friendly partnership, to work on image treatment. “It is a new milestone in French-Singaporean collaboration, and CNRS will now boost its support,” explains the French co-director of IPAL, CNRS scientist Jean-Pierre Chevallet. CNRS researchers will be encouraged to spend long-term periods at IPAL, where post-doc and PhD fellowships will be provided. “I am personally very happy,” adds I2R scientist Joo Hwee Lim, the Singaporean co-director of IPAL. “French researchers are usually very motivated and experienced. Bringing together our two different cultures will be a great asset.” France has a long history of research whereas Singapore is younger with a research culture in full expansion. And already, the results stemming from IPAL's three main research axes are remarkable. The cell phone application mentioned earlier is just one of the numerous projects carried out under the “Contextual Multimodal Interaction for Mobile Information Access” research axis. This research area focuses on information access via mobile devices. Due to their physical constraints, modalities other than traditional text keywords are explored: active visual input for instance. In Singapore, over one hundred landmarks have already been photographed and tested with success, and the prototype is continuously being improved.

The second research axis focuses on medical applications. If you have a headache, for example, but the head scan doesn't reveal anything, your surgeon could send the image for instant comparison in a worldwide medical database for further diagnosis. Just one of the applications that will stem from the “Medical Image Indexing and Retrieval for Assisted Diagnosis, Medical Research, and Teaching” axis. In the medical field, computer-aided diagnoses for radiological practices are on the rise and create a need for powerful data-retrieval methods.

The third axis links images and text. When the need arises for comparing similar pathological cases, it could be done using text queries (e.g.,  “show me head scans of people with such and such characteristics”). This research is tightly coupled with the previous field but has a wider frame. “The trickiest thing to understand is that we don't use words but concepts,” explains Jean-Pierre Chevallet. “For instance, when you type 'cancer,' our device will ask you if you are referring to an illness or to astrology. If you proceed with illness, it will not just search for all pages with the words 'skin cancer' in it, but also for all pages that include the term melanoma, because documents are indexed at a conceptual level. This system is already more efficient than textual indexing.” This third research axis is called “Precise Multilingual mixed-mode access to multimedia contents.” The team is working on a querying system that merges text with other media, so as to obtain a real inter-media index. Today, even with multimedia documents, textual queries are still the current standard for information access. There is still much to achieve. Researchers are focusing on how to index multimedia documents with concepts using fused multi-source knowledge bases.

What place could be better than high-tech Singapore to work on devices that will soon prove essential to everyday life? “Technology develops very quickly here,” acknowledges Jean-Pierre Chevallet. “Singaporeans are pragmatic. They think about what could be very useful in the future and build the best teams locally and from abroad to help them realize their visions.” Next year, IPAL will be moving to a 24-story skyscraper called Fusionopolis, where every area of Computer Science will be represented, from wifi technology to infocomm research. “Singapore gives itself the means to build and change the landscape, both literally and figuratively,” reckons Jean-Pierre Chevallet.

 

Samantha Maguire

Notes :

1. CNRS and the Université Joseph Fourier in Grenoble.
2. A*STAR's (Agency for Science, Technology and Research), Institute for Infocomm Research (I2R) in Singapore and the National University of Singapore (NUS).

Contacts :

> Jean-Pierre Chevallet Director of IPAL, Singapore.
Jean-Pierre.Chevallet@imag.fr
> Minh-Hà Pham-Delègue
Head of the Asia-Pacific Division at CNRS, Paris.
minh-ha.pham-delegue@cnrs-dir.fr


Top

Back to homepageContactcredits