A priori estimation of a real-time AER transformation

Determine whether a single transformation matrix W for adaptive eigenspectrum remapping can be estimated a priori for an entire ultrasound dataset and applied in real time.

Background

AER currently derives its weighting transformation from the measured covariance eigenspectrum, requiring estimation of the relevant subspaces and eigenvalues. The paper notes that preliminary analysis with randomized SVD still requires accurate estimation of the full eigenvalue spectrum, making full subspace estimation preferable for the present implementation.

For practical real-time processing, an unresolved question is whether one transformation matrix can be determined in advance for an entire dataset rather than repeatedly estimated during processing. Such a result could reduce computational demands and facilitate real-time ultrasound imaging applications.