Informations générales
Intitulé de l'offre : (M/F) Thesis offer : Task oriented multi-sensor fusion based on a differentiable approach: Application to coastal areas (H/F)
Référence : UMR5216-VIRFAU-056
Nombre de Postes : 1
Lieu de travail : ST MARTIN D HERES
Date de publication : mardi 14 octobre 2025
Type de contrat : CDD Doctorant
Durée du contrat : 36 mois
Date de début de la thèse : 1 décembre 2025
Quotité de travail : Complet
Rémunération : 2200 € gross monthly
Section(s) CN : 07 - Sciences de l'information : traitements, systèmes intégrés matériel-logiciel, robots, commandes, images, contenus, interactions, signaux et langues
Description du sujet de thèse
Remote sensing has demonstrated great potential for characterizing the chemical and physical properties of the Earth's surface, including coastal area. Several systems are equipped with multiple imaging sensors with complementary specifications in terms
of resolution (e.g. Pléiades, PRISMA and Sentinel-2). Image fusion is a necessary step to obtain an image with both optimal spatial and spectral resolution. Most of the image fusion algorithms focus on generating a high (spatially and spectrally) resolution image that will then be used in other downstream tasks (e.g., object detection, land cover classification, change detection…). Most of the fusion algorithms proposed in the literature rely on task-agnostic assumptions to find the optimal fusion
product, using explicit energy functional, or implicitly learning them from the data. The validation protocols rely typically on strategies to cope with the absence of a reference image at the target resolution using standard image quality metrics at low
resolution. There is an extensive literature on these metrics, most of which are based on the assumption of scale invariance of the fusion process, an assumption that is questionable. Additionally, the correlation between these image metrics and the
downstream task has not been extensively studied.
The objective of this PhD thesis is to propose an alternative approach to the image fusion process by considering the various stages of the processing chain— including image acquisition, the fusion itself, downstream tasks, and validation metrics—as
tunable processes including learnable parameters, whose precise forms can be optimized according to a specific objective. This approach aims to not only provide a principled comparison between fusion algorithms but also improve the quality of
downstream task performance, by optimizing previous elements of the processing chain to this end. These parameters can be optimized using conventional image quality metrics and/or task-based indicators. Such a unified view of the whole processing chain as a digital twin of operational acquisition and processing is made possible by the recent democratization of automatic differentiation frameworks (Pytorch, Tensorflow, JuliaDiff, JAX), and have been used in many areas of science recently (see e.g. the CLIMA initiative for global climate modeling). An end goal of this thesis is to unify the characterization of sensor specification, design of fusion algorithms, and their application to downstream tasks, in order to characterize the impact of the sensor characteristics on the quality of the fused images.
Contexte de travail
The Gipsa-lab is a joint research laboratory of the CNRS, Grenoble-INP -UGA and the University of Grenoble Alpes. It is under agreement with Inria and the Observatory of Sciences of the Universe of Grenoble. He conducts theoretical and applied research on AUTOMATICS, SIGNAL, IMAGES, SPEECH, COGNITION, ROBOTICS and LEARNING.
Multidisciplinary and at the interface between the human, the physical and digital worlds, our research is confronted with measurements, data, observations from physical, physiological and cognitive systems. They focus on the design of methodologies and algorithms for processing and extracting information, decisions, actions and communications that are viable, efficient and compatible with physical and human reality. Our work is based on mathematical and computer theories for the development of models and algorithms, validated by hardware and software implementations.
By relying on its platforms and partnerships, Gipsa-lab maintains a constant link with applications in a wide variety of fields: health, environment, energy, geophysics, embedded systems, mechatronics, processes and industrial systems, telecommunications, networks, transport and vehicles, operational safety and security, human-computer interaction, linguistic engineering, physiology and biomechanics, etc.
Due to the nature of its research, Gipsa-lab is in direct and constant contact with the economic environment and society.
Its potential as teacher-researchers and researchers is invested in training at the level of universities and engineering schools on the Grenoble site (Grenoble Alpes University).
Gipsa-lab develops its research through 16 teams or themes organized into 4 divisions:
• Automatic and Diagnosis (PAD)
• Data Science (PSD)
• Speech and Cognition (PPC)
• Geometries, Learning, Information and Algorithms (GAIA).
The staff supporting research (38 engineers and technicians) is distributed in the common services distributed within 2 divisions:
• The Administrative and Financial Pole
• The Technical Pole
Gipsa-lab has around 150 permanent staff, including 70 teacher-researchers and 41 researchers. It also welcomes guest researchers and post-docs.
Gipsa-lab supervises nearly 150 theses, including around 50 new ones each year. All the theses carried out in the laboratory are financed and supervised by teacher-researchers and researchers, including 50 holders of an HDR.
Finally, around sixty Master's trainees come each spring to swell the ranks of the laboratory.
Le poste se situe dans un secteur relevant de la protection du potentiel scientifique et technique (PPST), et nécessite donc, conformément à la réglementation, que votre arrivée soit autorisée par l'autorité compétente du MESR.
Contraintes et risques
NONE
Informations complémentaires
NONE