The Block Diagonally-Relaxed Orthogonal Projection Algorithm for Compressed Sensing Based Tomography
The theory of compressed sensing has recently shown that signals and images that have sparse representations in some orthonormal basis can be reconstructed from much less data, at high quality, than what the Nyquist sampling theory requires. In this talk, we will introduce a block diagonally-relaxed orthogonal projection algorithm for computed tomography image reconstruction in the compressed sensing framework and derive its convergence.
International Symposium on Photonics and Optoelectronics (SOPO)
"The Block Diagonally-Relaxed Orthogonal Projection Algorithm for Compressed Sensing Based Tomography."
Mathematical Sciences Faculty Presentations.