DISCOVER is an ERC funded project (starting grant 101076028).
Goal
Our goal is to develop approaches to assist experts in identifying and analyzing patterns. Indeed, while the success of deep learning on visual data is undeniable, applications are often limited to the supervised learning scenario where the algorithm tries to infer a label for a new image based on the annotations made by experts in a reference
dataset. In contrast, we want to take as input images without any annotation, automatically
identify consistent patterns and model their variation and evolution, so that an expert can
more easily analyze them.
The concept we will develop is the one of visual structures. Their key features will be their
interpretability, in terms of correspondences, deformations, or properties of the observed
images, and their ability to incorporate prior knowledge about the data and expert feedback.
It will explore two complementary approaches to formally define and identify visual
structures: one based on analyzing correspondences, the other on learning interpretable
image models.
We will develop visual structures in two domains: historical documents and Earth imagery.
For example, from temporal series of Earth images, we want to identify types of
moving objects, areas with different types of vegetation or constructions, and model the
evolution of their characteristics, which may correspond to changes in their activity or life
cycle. Ultimately, experts will still be needed to select relevant visual structures and perform
analysis, requiring to work closely with them, to design relevant features in our algorithms
and adapted interfaces for interaction.
People
Current collaborators:
Past collaborators
Elliot Vincent (PhD student with Jean Ponce at INRIA/ENS, now at IGN, the french National Geographic instute)
Job offers
We have positions for M2/PhD, research engineer and postdoc. Please fill this form if you are interested (I will not answer e-mail applications).
Selected presentations
- DH Nord Colloquium, Lille, France, November 6th, 2024, panel and hands on workshops (slides, website)
- VISART Vision for Art workshop keynote at ECCV 2024 in Milano, Italy, September 30th, 2024
(slides, website)
- Workshop on Archaeology & Machine Learning (WAML24), Paris, France, September 16th, 2024 (slides, website)
- Online DH Lecture Series 2024 @ MPIWG, June 19th 2024, Berlin, Germany, invited lecture on Computer Vision for History (slides, website)
- Symposium on Advances in applied mathematics and learning approaches for cultural heritage and the arts, April 22-24th, 2024 , Cagliari, Italy, invited presentation on Analysis-by-synthesis for History (slides, website)
- Workshop on Computing the Early Modern Printed Page, March 15th 2024, Oxford, UK, invited presentation on Interpretable computer vision analysis of historical prints (slides, program)
- Big Historical data conference, Nov 22-25th, 2023, Jena, Germany. Remote presentation on Discovering and analyzing visual patterns in historical image databases (slides, website)
- Conference on artificIAl Intelligence and applied MAthematics for History and Archaeology, Nov 27-28th, 2023, Nice, France. Keynote presentation on Computer Vision for Historians (slides, website)
- Machine Learning for Geometry workshop, Oct. 26th, 2023, Paris, France. Invited presentation on 3D analysis-by-synthesis (slides, website)
|