Automatic relighting of images of day and night outdoor scenes
This work was conducted in collaboration with the University of Évry-Val-d’Essonne. It consists of building generative AI solutions to automatically extract intrinsic lighting information in order to be able to relight the same scene under different lighting conditions, including night times.
- Hala Djeghim, Céline Loscos et Désiré Sidibé (2026) « SAIL: Self-supervised Learning of Lighting-Invariant Representations from Real Images with Latent Diffusion », The IEEE/CVF Winter Conference on Applications of Computer Vision (WACV 2026), Tucson (AZ), United States, mars 2026. pdf
