Post-Doctoral Fellow (PerForms): Computer Vision and 3D Modeling at IMT Atlantique

30 Novembre 2023

---

Catégorie : Post-doctorant

---

Post-Doctoral Fellow in the field of Computer Vision and 3D Modeling (W/M) ANR PerForms 2024-2026 project

1- WORK ENVIRONMENT

This post-doctoral position (24-month fixed-term contract) is offered by IMT Atlantique, an elite technological university in France, renowned for its academic excellence and cutting-edge research in the field of information science, energy technologies, environment and health, which is ranked in the top 10 in France.
The work will be carried out within the LaTIM Laboratory (https://latim.univ-brest.fr), which is
developing multidisciplinary research in medical information processing led by members of IMT Atlantique, INSERM, the University Bretagne Occidentale and the Hospital of Brest. LaTIM's translational research is supported by a wide range of industrial partnerships and a large number of partnership projects (regional, national, European). The laboratory is involved in 2 LabEx, 1 Infrastructure Nationale (France Life Imaging), 1 IRT (shared industrial research institute), has a LabCom and is the holder of an inter-regional Hospital-University Federation. The ANR PerForms project will be carried out within one of LaTIM's 4 teams, IMAGINE, which is developing its activities in the field of interventional therapies.

2- GLOBAL PROJECT SUMMARY

Fracture of the distal radius is the most frequent fracture. Especially, it is a public health issue related to women over 50- year-old with osteoporosis. In such fractures, the decision- making process remains challenging and controversial. Nevertheless, the final functional result is highly correlated to the restitution of certain morphometric parameters. Till then, most of the authors focused on the design of a single dedicated implant for fracture. However, as it has been proved in other joints and mainly in arthritis management, there is a major interest in preoperative planning first to confirm the surgical indication, then to determine the surgeon's objectives. Thus, we propose an original approach by providing an automatic planning software. This planning must be based on a reliable model while most of the clinical studies often skip this essential step. The main objective of the PERForms project is to develop a forearm 3D model that will be automatically generated by a planning software and that will provide a basis to calculate morphometric information. This model will be static and dynamical and generated in a reliable, precise, reproducible, and automated way. Thus, the surgeons will be helped in their diagnosis and their decision-making process. To achieve this objective, a specific Statistical Shape Model (SSM) method that has already proven its usefulness for other joints in terms of automatic segmentation, anatomy and functional prediction, and dynamic repositioning, will be used. The project will follow the stages listed below: 1) Data collection and labelisation 2) SSM building and validation 3) Generation of the premorbid forearm bones and realignment in an anatomical position 4) Automatic segmentation, landmarking process and validation 5) Model optimization and preclinical validation.

Keywords: Statistical shape model, forearm, preoperating planning.

3- POSTDOCTORAL MISSIONS

As part of the global project, under the supervision of Prof. Valérie Burdin (IMT Atlantique/LaTIM), in collaboration with Guillaume Dardenne (INSERM Researcher), the person recruited will have to:

1. Establish a state-of-the-art and identify the relevant solution.
2. Develop a robust approach for building a combined SSM including the forearm bones from healthy patients
3. Design a kinematic model of the forearm to adjust and correct the forearm in the good anatomic supination position
4. Publish and disseminate research results in recognized journals and conferences, and draft ANR project deliverables.
5. Participate in and lead working meetings with other partners.

4- POSTDOCTORAL ACTIVITIES

1. Establish a state-of-the-art overview and identify a relevant solution.
   • Carry out a state-of-the-art review of multi-bones statistical shape models
   • Make a state-of-the-art review on the forearm kinematics
2. Develop a robust approach for building a multi-bone digital twin of the patient's forearm and the kinematics: 3D statistical and dynamic model
   • Design and develop the code to build the statistical shape and motion model for the forearm
   • Infer the combined SSM to realign structures in the surgical reference position
3. Publish and disseminate research results in recognized journals and conferences, and draft ANR project deliverables.
   • Write publications resulting from research carried out in with supervisors
   • Write project deliverables
4. Participate in and lead working meetings with other partners.

5- TRAINING and SKILLS

Minimum education and/or experience required:

• PhD obtained less than 3 years before the hiring date (01/02/2024) with a strong expertise in computer vision (especially Machine Learning) and 3D modeling (Computational Mathematics)

Essential skills, knowledge and experience:

• Computer Vision, Machine Learning, Computational mathematics
• C++/Python development
• Software libraries in the field (VTK, OpenCV, etc.)
• Knowledge of the Scalismo library would be appreciated

Abilities and skills:

• Teamwork skills
• Interpersonal skills (multiple interactions with project partners)
• Writing skills
• Creativity/sense of innovation
• Autonomy

6- ADDITIONAL INFORMATION

• Participation in cadaver validation of the prototype may be requested

7- APPLICATION DOCUMENTS

• CV
• Covering letter
• Copy of PhD's degree diploma
• Thesis report for both reviewers and defense report
• List of publications
• Letter of recommendation from thesis supervisor

Send a zip file to guillaume.dardenne@univ-brest.fr and valerie.burdin@imt-atlantique.fr
Before 2024 January 4th

More information: https://institutminestelecom.recruitee.com/l/en/o/post-doctorat-en-vision-par-ordinateur-modelisation-3d-iti