Compressed sensing – with applications to medical imaging

Compressed sensing is a new approach for acquiring signals. It captures and represents signals and images at a rate significantly below Nyquist rate. In certain areas like magnetic resonance imaging (MRI), it is urgent to reduce the time of the patients’ exposure in the electromagnetic radiation. Compressed sensing breaks the canonical rules and effectively reduces the sampling rate without losing the essential information, so it has a wide application in medical imaging. In this project, three different recovery strategies Orthogonal matching pursuit (OMP), Compressive Sampling Matching Pursuit (CoSaMP) and Model-based recovery will be explored to investigate the performance of algorithms on different MRI images. Peak Signal-toNoise Ratio is used to measure the quality of reconstruction.