November 23, 2004

Speaker: Ignace Loris
PACM, Princeton University

Title: A review of the "Empirical Mode Decomposition" method

 Abstract:

Empirical Mode Decomposition (or EMD) was first introduced by Huang et al. (1998). The algorithm decomposes one-dimensional signals into "intrinsic mode functions", and allows for the calculation of a meaningful multicomponent instantaneous frequency. It distinguishes itself from other decomposition techniques by not depending on fixed basis functions; it derives a basis from the signal data only.
The algorithm, its justifications, some examples and also its shortcomings will be discussed.

Time-Frequency Brown Bag Seminar's homepage.