Information
- Publication Type: Student Project
- First Supervisor: Erald Vuçini
Abstract
Compressed sensing (or Compressive sampling) tries to exploit the sparsity of signals to either improve the quality of reconstructions, or reduce the required number of samples. The samples are usually taken by multiplication with some matrix with independent — normally distributed — random values, or by selecting values from the Fourier transform of a signal. Important is that the signal should not be sparse in the transformed domain. How many samples need to be taken depend on the characteristics of the matrix and the signal. Several introductory texts, as well as a lot of reference material, can be found on http://www.dsp.ece.rice.edu/cs/.Additional Files and Images
Additional files: report |
