Automatic abdominal segmentation using deep learning
At the Netherlands Cancer Institute (NKI), a dedicated abdominal surgical navigation system was developed. This system combines patient-specific 3D segmentation of abdominal anatomy and target tumors (e.g., surgical “roadmap”), segmented from preoperative contrast-enhanced CT scan, with intraoperative tracking of surgical instruments, to provide a real-time guidance within the model. These segmentations are commonly used for intraoperative guidance during complex oncological resections. Precision and the level of details of these 3D segmentation models significantly define the accuracy of intraoperative image guidance.
This 3D segmentation of patient-specific anatomy should include general anatomical landmark (i.e., bones, main veins and arteries), malignancies and all organs at risk (e.g., ureters, nerves, seminal vesicles). However, only a few anatomical structures (i.e., bones and arteries) are automatically extracted from contrast-enhanced CT images based on HU-based segmentation methods. All remaining structures require manually delineation, which is time-consuming and strongly dependent on the experiences of the user. Use of deep learning for automatic segmentation of patient anatomy already showed promising results in various applications of medical imaging.
The Diagnostic Image Analysis Group is part of the Departments of Radiology and Nuclear Medicine, Pathology, and Ophthalmology of Radboud University Medical Center. They develop computer algorithms to aid clinicians in the interpretation of medical images and thereby improve the diagnostic process. The group has its roots in computer-aided detection of breast cancer in mammograms, and has expanded to automated detection and diagnosis in multiple fields. The technology primarily used is deep learning.
In this project, you will be working on development of deep learning network for segmentation of abdominal anatomy used by surgical navigation. This is a new collaboration project of Diagnostic Image Analysis Group of the Radboud UMC and Surgical Oncology department of the NKI-AvL.
– Perform a literature research on existing methods and other related work
– Development of deep learning method for abdominal anatomy segmentation in pre-operative CT
– Write paper and/or thesis report
- Enthusiastic Master student in electrical engineering, biomedical engineering, computer science, technical medicine or a related field
- A good team player with excellent communication skills
- A creative solution-finder
Duration: 40 weeks (M3)
Start date: a.s.a.p.