Image processing
Scientific referents
Coordinator: Gabriele Facciolo
Scientific leaders:
- Miguel Colom
- Gabriele Facciolo
- Rafael Grompone
- Enric Meinhardt Llopis
Presentation
The team led by Gabriele Facciolo works on:
- Mathematical theory and algorithms for image and video processing,
- The design of cameras and satellites,
- The theory of statistical detection applied to images and videos,
- Variational restoration methods and applications of deep learning to computer vision.
Key words
Image processing; Video processing; Image analysis; Video analysis; Detection a contrario; Image authentification; Deep learning ; Satellite imagery.
Key facts
- 2024 : Pierre Lamoure prize. Thibaud Ehret, post-doctoral student at the Centre Borelli received the 2024 Pierre Lamoure prize 2024 awarded by FMJH (Fondation mathématiques Jacques Hadamard) which rewards mathematical research work in direct relation to a problem arising from the socio-economic world.
- 2023 : thesis prize awarded by GDR sécurité informatique 2023. Tina Nikoukhah post-doctoral student at the Centre Borelli is awarded for the quality and scientific impact of her thesis defended in November 2022.
- 2023 : Thesis prize Gilles Kahn 2023. Tina Nikoukhah post-doctoral student at the Centre Borelli is rewarded by the Société informatique de France for her work on anomalies making it possible to identify whether an image has undergone suspicious processing.
- 2023 : Best student paper at the Workshop EarthVision for Nguyen Ngoc Long (article co-authored with Jérémy Anger, Axel Davy, Pablo Arias and Gabriele Facciolo)
- 2022 : Prix Jeunes Talents France 2022. The Fondation L’Oréal supports 35 young female researchers each year in order to overcome the obstacles that limit their progress and accelerate their careers. Tina Nikoukhah, doctoral student at the Borelli Center, is one of the 2022 recipients.
- 2022 : Elyes Ouerghi winner of the "Use Case Competition" challenge organized ny EODisrupt. The research of Elyes Ouerghi, doctoral student at the Centre Borelli, on the automatic detection of methane plumes using satellite images was rewarded by EODisrupt.
- 2022 : Jérémy Anger winner of the 2022 thesis awarded by the Agence Innovation Défense. The awarded work concerns the deblurring of images in the presence of noise.
- 2021 : Tristan Dagobert and Thibaud Ehret are winners of the Earthvision CVPR challenge on unsupervised change detection using satellite images.s
- 2021 : Best student paper for Nguyen Ngoc Long at the Workshop EarthVision (article co-authored with Jérémy Anger, Axel Davy, Pablo Arias and Gabriele Facciolo).
- 2021 : Best paper prize at the International Workshop for Biometrics and Forensics (IEEE-IWBF 2021) for Marina Gardella (article co-authored with Miguel Colom, Jean-Michel Morel and Pablo Musé).
Structuring project of the theme
IPOL (Image Processing on Line) is a unique scientific journal that promotes reproducible research in image processing and signal processing (sound, video, physiological). Each published algorithm must include an online demo that can be run on a user-uploaded image or signal. This journal founded in 2011 has published nearly 260 algorithms with a mathematical description, a code, an online demo and a public experimental archive offering the numerous experiments carried out online.
Applications
The Image Processing team not only develops theories, but also supports their development and long-term applications in industry and society in general. She collaborates in particular with:
- The Agence spatiale française (CNES) on the design of Earth observation satellites;
- The laboratory DxOMark (word leader on image processing) on image denoising;
- The Agence France Presse and the french scientific police on the detection of falsified images and videos;
- Kayrros on the automatic monitoring by earth observation satellites (optical and SAR) to detect oil and gas activities and sources of methane and CO2 emissions;
- Thalès on the registration of aerial or satellite images for the guidance of flying objects and on the denoising and deblurring of videos
