Call for a research grant
The Institute of Systems and Robotics (ISR) opens a call for granting 2 research fellowships financed by project "VisArthro: Computer Vision For Computer Aided Arthroscopy ", with the reference PTDC/EEI-AUT/3024/2014, founded by Fundação para a Ciência e a Tecnologia I.P. (PIDDAC) and Fundo Europeu de Desenvolvimento Regional (FEDER) and Orçamento de Estado - OE, through COMPETE – Programa Operacional Fatores de Competitividade (POFC).
The VisArthro project aims to develop a Computer-Assisted Surgery (CAS) system for knee and hip arthroscopy procedures. Arthroscopy is a modality of minimally invasive orthopedic surgery in which the articulation is accessed through two small incisions that allow the insertion of instruments and a small video camera (the arthroscope). The navigation system to be developed will process the arthroscopic video in real-time in order to guide the surgical procedure by overlaying information on images using augmented reality techniques (for more information go to http://home.deec.uc.pt/~jpbar/visarthro/).
The Project has the objective of not only developing and implementing the system, but also validating it in “post-mortem” experiments. The candidates will integrate a multi-disciplinary team that includes the ISR-UC, the Coimbra Hospital and University Centre (CHUC), University College London (UCL) and Perceive3D (P3D), which is a spin-off of the University of Coimbra strongly committed to developing and commercializing this technology.
Scientific area / Key words: Computer Vision, Computer-Assisted Surgery, Multi-View Geometry, Machine Learning.
1. Duration and regime of activity: The selected applicants will be hired for 3 months, in an exclusive dedication regime, starting in February 2019, according to the regulation of the Portuguese Science Foundation available through the following link:
http://www.fct.pt/apoios/bolsas/normasbolsasemprojectosunidades. The contract can be eventually renewed if both sides agree.
2. Job description: The candidates will integrate a multi-disciplinary team with the aim of developing a system for Computer-Aided Surgery (CAS), being the first of the kind not requiring additional intra-operative sensing modalities such as opto/magnetic tracking. The research work will be carried at ISR-UC in close collaboration with P3D. The selected candidates will perform the following activities:
Research grant 1 (BI-1): Normal estimation from affine correspondences for contactless 3D registration
3D registration is a crucial step in CAS and enables to overlay a pre-operative 3D image or plan with the actual patient anatomy in the Operating Room (OR). Our current system for navigated arthroscopy performs 3D registration by asking the surgeon to touch the intercondylar region with an instrumented touch-probe. The candidate will develop research towards avoiding this explicit action by the surgeon. The idea is to investigate the estimation of 3D normal from affine correspondences in order to accomplish 3D registration without requiring explicit contact with a touch probe.
Research grant 2 (BI-2): 3D Registration / 6D Pose estimation in CAS using deep learning
The goal is to investigate methods of deep learning to perform 3D registration and/or 6D pose estimation of femur bone from video acquired intra-operatively. In a first stage we will employ a depth sensor together with the video camera. The idea is to use deep learning to detect the femur in image, segment the point cloud and feed the 3D result into our state-of-the-art registration algorithms. In a later stage we will investigate the application of recent methods of 6D pose estimation to our surgical navigation problems.
3. Financial conditions: According to the table of the values for national fellowships granted by FCT (available at http://www.fct.mctes.pt/pt/apoios/bolsas/valores/ ). The monthly value for the fellowship is 980 euros.
4. Start of work: The fellowship will start on the beginning of February, 2019.
5. Skills and Requirements: Master degree in Electrical Engineering, Informatics Eng., Biomedical Eng., Physics, Mathematics, or similar, with provided knowledge in programming. Previous experience in image and signals processing, computer vision, or machine learning are not mandatory but might strengthen the application.
6. Selection methods: Criteria to be used for curriculum evaluation and analysis:
(A) Merit and adjustment of the CV to requirements;
(B) Previous experience relevant for the proposed tasks;
(C) Motivation and autonomy (to be evaluated during the interview).
Each item will be rated on a scale from 0 to 5 and the final score (FS) will be given according to the following formula FS=0.6xA+0.25xB+0.15xC. The applicants should have at least 2.6, if none of the applicants meets the requirements, the competition closes automatically.
7. Jury: The jury responsible for selection of the applicants is composed by Professor João Pedro Barreto, Professor Hélder Araújo and Professor Paulo Peixoto.
8. Scientific Orientation: The scientific responsible at ISR-UC will be Professor João Pedro Barreto.
9. Applicants: The applicants should send a motivation letter, eventual recommendation letters and a detailed CV including a courses list and their ratings to firstname.lastname@example.org. The candidate must indicate for which grant is applying (BI-1 or BI-2) and, in case he/she applies to both, should indicate which one is the favorite.
10. Results publications: The evaluation results will be announced after the meeting for the selection by email. All candidates will be notified.
11. Period to submit applications: 15/01/2019 to 31/01/2019.