CNRS Summer School
Paris, July 1-5, 2012
How do we interpret images’ content? How do we tell images from other visual media? What can images represent? What ontology better describes their content? How do humans and machines recognize and classify images?
Roberto Casati, Institut Nicod, CNRS-ENS-EHESS
Anouk Barberousse, Université de Lille 1
Alberto Voltolini, Università degli Studi di Torino
Images are universal instruments of representation and communication. In many intellectually complex activities (the execution of plans and projects, the identification of people and places, navigation, data collection, medical diagnoses) the use of images is essential. Their interpretation requires little teaching (as opposed to, say, that of written language). But at the same time images are inherently ambiguous, and their interpretation may pose difficult problems.
This is particularly evident now that countless images are available in online archives. Their content is often made explicit by annotations (captions, tags, place and time stamps). Software for automatic image interpretation has developed at an impressive rate in recent years, but some problems remain hard to tackle, especially when moving from the identification of instances of objects (tokens) or the recognition of simple categories (plants, vehicles) to attempts to work with more complex categories.
Ontological/philosophical issues interface here with widening knowledge about cognitive processes and technological development.
- provide multi disciplinary state-of-the-art information
- establish a network of people working on images from many different point of views.
Building on a tradition of highly interactive summer schools, we would like to create a three-day labor-intensive mini-lab. Each day-long session is devoted to a different theme, around the work of one-two instructors. Breakout sessions allow for in-depth analysis of the issues. Writing of research documents is encouraged.
The language of the school is English. One session will be offered in French.
Sun, July 1, 2011
Arrival and registration
Mon, July 2
John Kulvicki (Dartmouth): Levels of content in images
Tue, July 3
Simon Thorpe (CNRS, Toulouse, and Spikenet Corp.): Fast image recognition in man, monkey and machine
Keynote evening lecture: Paul Taylor (Warburg Institute, London): The complex management of an art historical photographic collection
Wed, July 4
Matthieu Cord (UPMC-Sorbonne Universités, LIP6): Visual categorization, indexing and interactive retrieval: challenges for the new generations of image processing systems
Tutorial: Stéphane Pouyllau (CNRS -TGE Adonis): Informatisation documentaire d'images scientifiques : problématique de métadonnées ; stockage, diffusion et enrichissement de données
Thu, July 5
To appy for the school
This is a simple
application method: please register with Easychair.org as
author, fill in your personal infomation, give your name
as "title", and paste or attach a short letter of
motivation describing your background and interests (.pdf
only if attached).
Applications should be sent by Feb 10, 2011. Successful applicants will be contacted by Feb 20, 2012.
A 150€ fee covers accommodation and tuition. CNRS members will be exempted and their full costs, including travel costs will be covered. Some bursaries will be available for non-CNRS members.
Venue: CISP, 6 avenue Maurice Ravel, 75012 Paris. Accommodation will be provided at the same address. Accommodation includes the nights of Jul 1st throuhg July 4th, as well as breakfasts and lunches of Jul 2nd through 4th.
An active involvement of participants is the key feature of our school. Some material ought to be read and commented prior to the school’s start. Each instructor has a full day of interaction with attendees, in both plenary and breakout sessions. At the end of each day a final meeting summarizes the main teachings.
Institut Jean Nicod
casati AT ehess DOT fr
Download a pdf of the flyer ISP2012 Images School Paris 2012