GE HealthCare is a leading global medical technology and digital solutions innovator. Our purpose is to create a world where healthcare has no limits. In Norway, entire systems including the underlying software for cardiac ultrasound have been designed with a long history in research and development. At Forskningsparken, there is a R&D department working on software solutions for the next generation of ultrasound scanners and building the forefront when it comes to medical research with strong collaborations within Norway and with the rest of the world. In Olso in particular, one can find not only strong technical but also strong clinical expertise with tight connections to local and global hospitals.
Motivation:
Cardiac care is increasingly shifting from open-heart surgery to minimally invasive therapeutic approaches, driven by the need for faster recovery and reduced patient risk. In this transition, TEE (transesophagal echocardiography) imaging is essential, providing high-resolution, real-time ultrasound (US) guidance during procedures. Since the probe is inserted through the esophagus, the heart can be visualized without occlusion from surrounding anatomical structures.
US is a cornerstone of these interventions, but the presence of intracardiac devices can interfere with imaging quality—blocking acoustic windows or introducing artifacts. To maintain the effectiveness of TEE, solutions are needed to address these imaging challenges, ensuring compatibility between devices and ultrasound-based visualization.
Future solutions need to be designed to be robust towards occlusion and potentially detect and track interventional devices and instruments.
Project:
In this project, different supervised and unsupervised solutions could be explored to detect external objects in 2D and 3D ultrasound images. Here, the use of methods is not limited to AI but also looking at ultrasound signal processing and image processing.
Ideal Candidates We’re looking for curious and motivated students with backgrounds in:
- Machine learning / deep learning
- Biomedical engineering or medical imaging
- Signal and image processing
- Programming (Python, PyTorch, TensorFlow)
Contact UiO:
Sarina Thomas – Sarinat@ifi.uio.no
Eigil Samset – eigilsa@ifi.uio.no