Ultrasound imaging has been widely used in clinical applications. Ultrasound images have also been intensively studied on various image processing approaches, including segmentation and registration, etc. However, the raw data of ultrasound (i.e. radio frequency data) are typically neglected and receive less attention. In this MS thesis project, ultrasound raw data will be revisited and explored using deep learning methods to unveil some unique features that can be used for medical applications. The potential application will focus on orthopedics and bone detection. This work is a collaborative work with RadboudUMC ORL lab (https://www.radboudumc.nl/en/research/research-groups/orthopaedic-research-lab) and the biomechanical engineering group (BE) at UT.
The required skills:
- Python Programming
- Familiar with DL networks, e.g. TensorFlow, PyTorch or Keras
- Willing to work with clinical datasets and conduct clinical studies.