This repository contains all the files necessary to perform the algorithm BoundEffRed (Boundary Effects Reduction). This is an efficient forecasting approach for the real-time reduction of boundary effects in time-frequency representations. The theoretical background relative to the algorithm can be found in the following paper:
A. Meynard and H.-T. Wu, "An Efficient Forecasting Approach to Reduce Boundary Effects in Real-Time Time-Frequency Analysis", IEEE Transactions on Signal Processing, vol. 69, pp. 1653–1663, 2021, available online.
![]() |
![]() |
---|---|
Ordinary syncrosqueezing transform of a cardiac signal | Boundary-free syncrosqueezing transform of the same signal, obtained via BoundEffRed |
WARNING: This algorithm has been designed to work optimally with MATLAB R2020a. We do not guarantee its compatibility with previous releases.
The folder Signals
contains some biomedical and synthetic signals used to implement the above-mentioned algorithm. The following folders are specifically related to the implementation of BoundEffRed:
Algorithm
contains the MATLAB functions enabling the implementation of BoundEffRed, using differents extensions schemes such as SigExt, EDMD, or GPR (see the paper for details).TimeFrequencyScaleRep
contains the functions generating Time-Frequency and Time-Scale representations such as STFT, Synchrosqueezing Transform, Reassignment, or ConceFT.Scripts
contains the scripts to perform the experiments detailed in the paper. Some corresponding results are directly provided in subfolderResults
.
The folder Paper
contains the article and the associated Supplementary Materials, while the folder Animations
contains some videos featuring the real-time boundary-free synchrosqueezing transforms of a few biomedical signals, provided by BoundEffRed.
Authors: Adrien Meynard and Hau-Tieng Wu
Contact email: [email protected]