Shepp-Logan phantom

by Yubo "Paul" Yang

Compressive Sensing

Compressive sensing takes advantage of sparsity to reconstruct full signal from sparse samples in a way that is not limited by Nyquist-Shannon. It effectively performs compression at the time of sensing so that few detector/sensors are needed. It has many practical applications such as single-pixel camera, digital-to-analog conversion, and lattice dynamics in atomic simulations.

Presentation Summary

In these slides, I present:

  • Basic idea behind the compressive sensing (cs).
  • A few simple examples.

Examples

  • cs_simple.py: simplest CS implementation.
  • beat_nyqsha.py: beat Nyquist-Shannon frequency for perfect reconstruction.
  • cs_phantom.py: reconstruct Shepp-Logan phantom. You also need phantom.py to generate the phantom image and config_h5 to store the constructed A matrix (or code is pretty slow).

References

All Signal Processing

ALGORITHM
signal processing

Archive