Ultrasound signal processing acceleration with CUDA
Ultrasound imaging is a non-invasive imaging technique that provides visible information on the structure of musculoskeletal tissues and organs. Thanks to the increased availability of digital computation capabilities, fully-digital ultrasound systems are progressively replacing the more expensive and bulky analog solutions. In a fully-digital solution, the ultrasound signals are digitized directly on the probe, and sent to a commodity device (e.g. PC) for further processing. To support continuous imaging with high frame rates, the remote PC should perform efficient image reconstruction with minimal execution time.
Goal & Tasks
The goal of the mini-project is to optimize ultrasound signal processing routines and the Delay-And-Sum image reconstruction algorithm with PyCuda. The optimized code will be integrated into the PyBF ultrasound beamformer library developed at IIS.
- Python (sklearn, tensorflow)
- 20% Literature Study
- 80% Coding