From Medical Imaging to Physics-Based 3D Liver Animation
Type: Engineering school or master degree
Today most patients with liver cancer are treated through traditional open liver surgery, which involves a high risk of serious post-operative complications. Evolving open liver surgery to a minimally invasive liver surgery (laparoscopy) is the key for fast patient recovery. In laparoscopy, resection is performed by a camera and surgical tools that are inserted into the patient's body through a few small holes. However, this minimally invasive surgery challenges surgeons to see the vascular structures, to find the hidden tumors and to navigate on and inside the liver. We can relieve the task of surgeons by building a patient specific physics-based 3D liver model with the hidden vascular system, tumors, etc., using the pre-operative MRI (or CT) scan images. In MRI scan images these structures are clearly visible and segmented weeks before the surgery by a radiologist. Once this physics-based 3D liver model is available, then the surgeons can animate (translate, rotate, deform) it during surgery so that the model fits the real liver of the patient seen from the laparoscope (camera) in a fixed pose. In this way surgeons will be able to see the hidden structures inside the liver using augmented reality (AR).
The goal of this internship is thus to build a physics-based 3D liver model from pre-operative MRI scan images and implement a user interface for surgeons to animate this model. This internshiprequires skills to understand, improve and implement current techniques, and solve some theoretical problems.
Successful outputs of this internship will first of all improve the quality of minimally invasive liver surgerywhich has a direct impact on the patient’s future life. Secondly, these outputs may lead to a scientific publication in one of the medical conferences which is a plus for the intern for his/her future academic career.
Software/Hardware needs and skills
Basics in computer vision: sensor models, 2D/3D images (MRI), features, segmentation; good coding skills (Matlab, C++) and some experience in OpenCV and OpenGL is a plus.