Internship Scalable Local Inversion For Real-Time Ultrasonic Imaging-Saclay H/F - CEA

Le CEA est un acteur majeur de la recherche, au service des citoyens, de l'économie et de l'Etat.


Il apporte des solutions concrètes à leurs besoins dans quatre domaines principaux : transition énergétique, transition numérique, technologies pour la médecine du futur, défense et sécurité sur un socle de recherche fondamentale. Le CEA s'engage depuis plus de 75 ans au service de la souveraineté scientifique, technologique et industrielle de la France et de l'Europe pour un présent et un avenir mieux maîtrisés et plus sûrs.


Implanté au coeur des territoires équipés de très grandes infrastructures de recherche, le CEA dispose d'un large éventail de partenaires académiques et industriels en France, en Europe et à l'international.


Les 20 000 collaboratrices et collaborateurs du CEA partagent trois valeurs fondamentales :


- La conscience des responsabilités

- La coopération

- La curiosité

As an intern at the CEA, you will have the opportunity to work within a world-renowned research environment. Our teams consist of passionate and dedicated experts, providing an ideal setting for learning and collaboration. You will have access to state-of-the-art equipment and top-tier research resources to carry out your assignments. The work conducted may potentially lead to a scientific publication.

Context :

Ultrasonic imaging from Full Matrix Capture data can exceed the resolution of the Total Focusing Method (TFM) when the individual transmitter-receiver contributions are inverted locally.

Local linear direct model

The current physics-informed unrolled method [1] can resolve point reflectors below the Rayleigh limit. The learned inversion parameters can also transfer across nearby depths.

Two point-like reflectors separated by 0.8 mm (left) and 0.4 mm (right). Top: TFM images. Middle: reconstructions obtained at
inference with configuration B. Bottom: lateral profiles at z = 16 mm. Blue crosses and dotted lines indicate the reflector positions.

However, the method currently processes isolated windows and requires the computation of large local operators.

Extending this approach to whole-component inspection requires a scalable strategy that preserves reconstruction quality while meeting real-time constraints.

The objective of the internship is to develop a scalable implementation of local ultrasonic inversion for extended inspection areas. The work will focus on identifying reusable structures across neighbouring operators, combining overlapping local reconstructions consistently, and determining how far locally learned parameters can transfer before retraining becomes necessary.

The internship work will include:

• analysing the spatial regularities and the low-rank or spectral structure of local operators;
• designing overlap and stitching strategies for neighbouring reconstruction windows;
• investigating parameter-transfer rules across windows;
• optimising the implementation.

[1] K. Diallo, F. M. Ngolè Mboula and F. Le Bourdais, Physics-Informed Super-Resolved Ultrasonic Imaging, 11th Junior Conference on Data Science and Engineering, 2026 [].

#Cea-List

The internship is intended for a final-year Master's student or engineering student specialising in applied mathematics, signal processing or scientific computing.

The candidate should have strong Python programming and numerical optimisation skills. Experience with PyTorch, inverse problems, GPU programming or ultrasonic imaging would be an advantage.