PHD MSCA ITN EU UNVEIL M/F

Job Description

Thesis Subject

Title: Processing and fusion of data coming from a variety of techniques (XRF, HSI, HMI, FUV) for improved material characterisation and pigments identification.

Sub-title: Uncertainty-aware multimodal learning for pigment mapping: joint X-Ray Fluorescence & Hyperspectral data fusion

Project overview:
In the context of museum artwork conservation (2D and 3D such as paintings, photographs, sculptures…) where material identification plays a crucial role, this PhD will focus on material and pigment identification. The research aims to develop a data-efficient and uncertainty-aware framework for pigment mapping in Cultural Heritage. Heterogeneous spectral and imaging data from different sensors will be jointly exploited, including X-ray fluorescence (XRF) and optical reflectance and fluorescence imaging (hyperspectral/multispectral).

The primary objective is multimodal unmixing, i.e., estimating pigment abundances at each location under scarce supervision and strong domain shifts across sensors and artworks. Pigment identification can then be derived from the unmixing outputs (endmember presence and abundance distributions), while optional auxiliary classification objectives may be introduced to stabilize training.

The research will explicitly account for modality-specific mixture formation laws: XRF mixtures can often be approximated by (quasi-) linear mixing in appropriate representations, while optical spectral mixtures are commonly approximately described by Kubelka–Munk theory or related radiative transfer approximations. These physical laws will guide model design, constrain the latent space, and improve interpretability.

The main research axes will be:

Multimodal data acquisition, preprocessing, and cross-modal alignment: Development of painted mock-ups in controlled conditions, acquisition of multimodal datasets (XRF and optical spectral imaging), and design of robust preprocessing and cross-modal registration pipelines. This axis provides the experimental foundation and enables quantitative evaluation.
Representation learning with physics-guided data augmentation and adaptation: Design of self-supervised learning strategies to learn shared and modality-specific embeddings, using physics-inspired data augmentations and mixture models (linear/quasi-linear behavior for XRF; non-linear Kubelka–Munk-inspired transformations for optical imaging) to regularize training and improve robustness under domain shifts (across artworks and sensors). This axis aims to obtain stable, transferable representations despite scarce labels, heterogeneous sensors, and varying imaging conditions.
Multimodal unmixing, uncertainty quantification, and adaptive fusion: Development of multimodal unmixing models to estimate pigment abundances, combined with uncertainty-aware mechanisms for dynamic modality weighting, open-set rejection, and calibrated confidence maps for expert interpretation. This axis delivers physically meaningful pigment abundance maps and trustworthy fused outputs for conservation scientists

Expected Results:
1) A multimodal dataset combining controlled mock-ups and selected artworks, with standardized acquisition and documentation protocols;
2) A set of algorithms for multimodal representation learning and data fusion tailored to Cultural Heritage imaging;
3) A robust unmixing and pigment-mapping workflow producing interpretable abundance maps, allowing conservators and imaging scientists to apply the developed methods to new artworks.

Your Work Environment

UNVEIL is a Marie Sklodowska-Curie Doctoral Network linking 21 academic, cultural, and industrial partners to develop advanced nondestructive evaluation and data-driven digital tools for paintings and 3D artworks (https://unveil-dn.eu/). The goal is to understand how objects are built and age and to support conservation, restoration, and authentication—while training 12 doctoral researchers across engineering and the humanities.
We combine terahertz, ultrasound, infrared thermography, hyperspectral/multispectral, and X-ray imaging with multimodal data fusion, interactive visualisation, and digital twins. Results are validated on mock-ups and artworks, showcased via AR exhibits at partner museums, and shared openly (data, software, DTs) through European platforms.
The Centre for Research and Restoration of the Museums of France in Paris (C2RMF), in collaboration with the SATIE (Systèmes et Applications des Technologies de l'Information et de l'Energie) and the Centro per la Conservazione ed il Restauro dei Beni Culturali “La Venaria Reale” (CCR) in the province of Turin, invites applications for 3-year full time early stage Researcher (DC8) position in pigment identification in the context of the EU-funded Marie Sklodowska-Curie Initial training network (ITN) project entitled UNVEIL.
The position will be based in France, at the C2RMF in Paris, and SATIE in Gif-sur-Yvette (near Paris). The PhD Enrollment is Université ED STIC (Sciences et Technologies de l'Information et de la Communication, ED#580). Collaborations with the firm Profilocolore in Roma and the CCR near Turin are planned.
The C2RMF is the national research centre in France responsible for the research, conservation and restoration of the items held in the collections of the French museums (https://c2rmf.fr/). C2RMF also carries out extensive scientific studies and data recording for these collections, and is active both nationally and internationally in the field of cultural heritage conservation and analysis. The C2RMF is involved in the development of technologies and scientific procedures employed in the preservation of artworks and artefacts, both on its own and in partnership with other museums and research institutions across the globe. The recruited researchers will work in a professional environment in the domain of conservation, restoration and research on cultural heritage, the C2RMF. During their 36 month contract the researchers will be based in the imagery team and in close collaboration with the other departments of the center as well as other partners from the UNVEIL project all around Europe.
SATIE is a joint research unit (UMR CNRS 8029) bringing together researchers, engineers and doctoral students across a broad spectrum of electrical engineering topics, including electronics, electrical energy systems, signal processing, and emerging interdisciplinary fields. The laboratory maintains strong academic and industrial partnerships. Within SATIE, the SICOIA research group, located in Gif-sur-Yvette, develops both theoretical and applied methodologies in Data and Image Analysis, Computational Imaging, Statistical Learning, Uncertainty Quantification, Robust Estimation, and Deep Neural Networks. The group combines expertise in stochastic modeling, Bayesian inference, data fusion and modern machine learning. Its research activities span various application domains such as security, non-destructive testing, infrared imaging and cultural heritage analysis. SICOIA offers a stimulating scientific environment with access to high-performance computing resources within the Paris-Saclay ecosystem.
The Centro per la Conservazione ed il Restauro dei Beni Culturali "La Venaria Reale" (CCR) is a conservation centre dedicated to cultural heritage and managed by a non-profit Foundation established by the Italian Ministry for Cultural Activities, the Piemonte Region, the University of Turin, the Fondazione per l'Arte of the Compagnia di San Paolo, and the Fondazione Cassa di Risparmio di Torino. Over time, additional institutions have joined the original founders, including the Metropolitan City of Turin, the City of Turin, and the City of Venaria Reale (https://www.centrorestaurovenaria.it/eng/). The CCR has nine conservation laboratories organized into specialized workshops: wooden furniture; contemporary art; paper and photography; mural paintings, stone materials, and construction sites; panel paintings; paintings on canvas; metals, ceramics, and glass; wooden sculptures; fabrics, tapestries, leather, and carpets. Conservation activities are carried out on behalf of the founding institutions as well as other public and private bodies. The CCR hosts scientific laboratories dedicated to photographic and 3D documentation, multiband imaging, scientific analysis, and environmental monitoring for cultural heritage. These facilities support both the study of artworks and the development of conservation and restoration strategies.
In collaboration with the University of Turin, the CCR hosts the Master's Degree Course in Conservation and Restoration of Cultural Heritage, which trains professional conservators in accordance with Italian legislation (DLgs 42/2004). It has an Advanced Training School (SAF), which organizes and manages training and professional development courses and contributes to the activities of the Master's Degree Course.
Through its integrated approach to conservation, research, and training, the CCR promotes the development of professional skills and fosters employment opportunities in the field of cultural and environmental heritage at both national and international levels.
Main supervisor(s):
Dr Clotilde Boust (C2RMF & CNRS-SATIE)
Prof. Sylvie Le Hégarat-Mascle (Université Paris-Saclay / CNRS-SATIE-SAC)
Assoc. Prof. Aurélia Fraysse (Université Paris-Saclay / CNRS-SATIE-SAC)
Co-supervisor: Dr. Tiziana Cavaleri (CCR)
Referee: Gui Yun Tian (Newcastle University)
Planned secondments
Year 2 : 3 months at the CCR (Venaria Reale, province of Turin) for training on optical and chemical analysis.
Year 3 : 3 months in Roma with Profilocolore Firm for training on HMI and methods for classification and comparison of different techniques.

Compensation and benefits

Compensation

2300 € gross monthly

Annual leave and RTT

44 jours

Remote Working practice and compensation

Pratique et indemnisation du TT

Transport

Prise en charge à 75% du coût et forfait mobilité durable jusqu’à 300€

About the offer

About the CNRS

The CNRS is a major player in fundamental research on a global scale. The CNRS is the only French organization active in all scientific fields. Its unique position as a multi-specialist allows it to bring together different disciplines to address the most important challenges of the contemporary world, in connection with the actors of change.

CNRS

The research professions