Université d'Auvergne Clermont1 | CNRS


AR2OR – Augmented Reality in Gynecological Surgery


Type: Engineering school or master degree

  • Deadline: when it's filled
  • Salary: 508,20€ net/month
  • Team: ALCoV
  • ISIT supervisors:
  • Description

    ALCoV is a CNRS research group in the Université d’Auvergne that pioneers research in Augmented Reality (AR) in gynecological Minimal Invasive Surgery (MIS). One of the main projects is to assist surgery of the uterus during myomectomy. This surgery involves removing myomas (which are fibrous benign tumors also known as fibroids) that grow in the uterus wall by surgical resection using a laparoscope and incision tools.  AR can help this surgery by showing where the exact position of the fibroids are, even when they are deep inside the uterus [1]. We achieve this by scanning the patient before surgery in an MR machine, and semi-automatically segmenting the uterus and the myomas from background structures in the MR image. During surgery, the MR image is registered to the laparoscope’s image, and the myomas are then visualized on the laparoscope’s image. This visualization gives the impression the uterus is transparent, and the surgeon can see inside it [1]. This process involves automatic laparoscopic image segmentation and registration in realtime.

    In addition to myomas, there is other useful information that can be augmented onto the laparoscopic images. We have identified three types of information. The first are important vessels in the uterus, the second is the uterine cavity and the third are the muscle fiber directions of the uterus. Major vessels are important to visualize because the surgeon has to stop the blood supply to the myoma, and this can be difficult and time consuming. The uterine cavity is important to visualize because the surgeon should not cut into it during the operation. The uterus muscle fibers are important to visualize because we believe healing time can be reduced by having the surgeon cut along muscle fibers to access the myoma, rather than cutting across the fibers.

    The goal of this internship is to (i) investigate methods to acquire the above information with pre-operative MR scans (T2 weighted MRI and Diffusion Tensor Imaging), and (ii) investigate the best way to visualize all the information during surgery. The candidate will work with computer vision experts, radiologists and surgeons in the ALCoV team. The candidate will gain a good understanding of MR acquisition and segmentation methods, inter-modal medical image registration and visualization methods in AR. If successful, the results will be publishable in both the technical and medical literature. This internship is an opportunity for a well-motivated bright student to participate to a larger project and to contribute to the new trend of using AR for medical applications, with a direct impact of the patient’s life by improving the quality of surgery.

    [1] Computer-Aided Laparoscopic Myomectomy by Augmenting the Uterus with Pre-operative MRI Data
    T. Collins, D. Pizarro, A. Bartoli, N. Bourdel and M. Canis
    ISMAR'14 - IEEE International Symposium on Mixed and Augmented Reality, Munich, Germany, Sep. 2014.

    Software/Hardware needs and skills

    Knowledge of segmentation in computer vision, camera models and 3D rendering are mandatory, as well as good coding skills (C++) and preferably some experience in OpenCV. GPU coding, OpenGL and knowledge of medical imaging is advantageous.