Medical Image Reconstruction and Processing

EEE 575 treats medical imaging as an inverse problem: you're handed incomplete, noisy measurements from MRI, CT, or MPI scanners and have to recover a faithful image, which means thinking carefully about sampling, Fourier-domain encoding, and the priors that make reconstruction well-posed. Expect five homeworks and a project where you actually implement the core machinery — gridding/NUFFT for non-Cartesian data, SENSE/GRAPPA for parallel imaging, compressed sensing recovery, and registration/segmentation on the post-processing side. It's a graduate elective that assumes you're comfortable with signals, linear algebra, and a bit of optimization, and it's the natural bridge from classical DSP into the reconstruction pipelines used in clinical scanners and modern imaging research.