Post-Doc In Deep Learning And Computer Vision For Tracking Dislocations During In-Situ Electron...

Context : Mechanical properties of metallic alloys are largely governed by the motion of nano-scale linear defects called dislocations and their interactions with the microstructure. Hence, understanding dislocation dynamics is of fundamental interest to predict material strength in many modern applications for aeronautics and automobile. At CEMES, moving dislocations, appearing as fine black lines, are directly observed under a mechanical stress during in-situ transmission electron microscopy experiments (see movies: https:// mompiou.free.fr/videos/)LEG. Lots of quantitative information such as dislocation speed, curvature, anchoring points distributions... can be retrieved from a still image analysis MOM. However, this work is up to date performed manually, which limits statistical treatments, although a large database of observations is available. Moreover, this approach misses a large amount of information by sampling observations and averaging quantities. In the proposed project, the overall objective of is to take benefit of computer vision coupled to deep learning methods SHE to exploit databases in order to construct numerical twins of in-situ observations. From this, we expect to retrieve quantitative data statistically significant which will allow direct comparison with meso-scale simulations, such as Dislocation Dynamics (https:// www. numodis.fr/)DRO, in modern metallic alloysLIL.

Mission : The PPM group at CEMES is looking for post-doc candidates for a one year position within the NanoX labex project DISTRACK. The applicant is expected to adapt/develop deep learning models for dislocation segmentation and/or computer vision strategy for tracking. She/he will participate to the collect and constitution of an appropriate database in link with researchers. Physically informed models could eventually be investigated based on Dislocation Dynamics simulations in collaboration with CEA Saclay.

Ref :

SHE: S Mingren, G Li, D Wu, Y Yaguchi, J. C. Haley, K. G. Field, and D. Morgan. Computational Materials Science 197 110560 (2021)

DRO: J. Drouet, L. Dupuy, F. Onimus, F. Mompiou, Scripta Materialia. 119 (2016) 71–75.

MOM: F Mompiou, D Tingaud,Y Chang, B Gault, G Dirras Acta Mater. 161 420-30 (2018)

LIL: L Lilenstein et al., Acta Mater. 142, 131-141 (2018)

LEG: M Legros, Comptes Rendus Physique 15, no. 2–3 224–40 (2014)

phD in materials science, computer science. Open to other profiles.

Provide CV, cover letter first and any documents that prove motivation and adequation to the job. Selection based on document examination and interview.

ENGLISH: Excellent

Strong interest in the field of deep learning, especially convolution neural networks, and computer vision. Demonstrated skills in coding. Working experience with electron microscopy images will be appreciated. Ability to communicate clearly on the technical areas concerned by adapting to the interlocutors and to discuss constructively the progress of the project.