![]() |
Synchronization
Toolbox |
![]() |
This Matlab program (name: warpGUI) synchronizes univariate and multivariate trajectories. There are two different algorithms that can be used. The first algorithm is named symmetric dynamic time warping (SDTW). The second algorithm is called parametric time warping (PTW). Some example data sets are automatically loaded in the main workspace of Matlab when warpGUI is started. The following buttons can be clicked in warpGUI:
|
||
Load
Data |
A
window is opened that allows you to use a Matlab variable (from the main
Matlab memory space) as a reference and sample vector |
Warp |
The actual
warping action is performed |
Print |
Prints
the warping window, including the figures and the settings that are used
for synchronization |
SDTW |
When this
radio button is selected, the SDTW algorithm is used for synchronization |
PTW |
When this
radio button is selected, the PTW algorithm is used for synchronization |
When the 'SDTW' option is selected, the following options are available:
|
|
Global
Constraints |
Constraints
exclude certain regions from the search area for the best warping path.
These constraints are calculated by the routine and drawn in the plot
(upper right) showing the warping path |
Normalize |
The SDTW
algorithm is normalized |
Direction
Penalties |
The direction
penalties implemented make the incorperation of horizontal and vertical
steps in the warping path less favorable. Increasing the size of the penalty
value makes the penalty more severe. The software returns a sensible quess
for this value when the data is loaded for the first time |
Processing your own data is possible by first loading your own data in the Matlab main window, then pressing the "load data" button and selecting the variables that contains your data. This program is tested for Matlab v6.1 and v6.5. Note that version 6.5 of Matlab requires windows-XP. This program was made by Hans Boelens (Process and Analysis group, University of Amsterdam), Henk-Jan Ramaker (TIPb) and Paul eilers (Department of Medial Statistics, University of Leiden). For questions about SDTW, comments or source code, please contact Hans Boelens (Hans.Boelens@science.uva.nl) )or Henk-Jan Ramaker (HJ.Ramaker@tipb.nl). Questions regarding PTW are best answered by Paul Eilers (P.Eilers@lumc.nl).
References Henk-Jan Ramaker, Eric van Sprang, Johan Westerhuis, Hans Boelens and Age Smilde, DYNAMIC TIME WARPING OF SPECTROSCOPIC DATA, accepted (2003), Analytica Chimica Acta.
|
This graphical user interface and the code in it belongs to the holder of the copyrights and is made public under the following constraints: it must not be changed or modified and code cannot be added. The file must be regarded as read-only. Furthermore, the code cannot be made part of anything but the 'Synchronization Toolbox'. In case of doubt, contact the holder of the copyrights. © 2002 by the Process and Analysis group of the University of Amsterdam. |