/Waveform-Extractor

Reconstructing waveforms using Fourier series from any figures of your wish.

Primary LanguagePythonMIT LicenseMIT

Waveform Extractor

About

Waveform Extractor is a lightweight toolbox allowing you to extract any waveform $Q$ (preferably, periodic) from a raster figure and reconstruct into 21 Fourier terms (surely the code is scalable for any 2D regression method upon request).

The code is inspired by Dr Zhuo Cheng's in-house MATLAB code initially developed in 2007. The new version

  • incorporates a simple tkinter-based user interface for i/o path selection and data input,
  • incorporates a data point removal feature (right mouse button) that allows users to revoke any wrongly labelled points on the current canvas, and
  • fixed an inherited hidden bug in the original MATLAB code, which would result in unreasonable $y$-shifting under the presence of negative-valued data.

To Use

The directory /asset provides three figures for rapid testing purposes.

Waveform Extractor v1.0.3 is also compiled with PyInstaller and ready for the quick Windows deployment. The package is downloadable at https://imperialcollegelondon.box.com/s/ncbg6oxqfu43raw4aez47hc9921adwde.

Note: the pre-compiled package disabled the debug mode for users' access to more detailed backend info. To enable the debug mode, simply execute the code

python WaveformExtractor.py --debug

or

python WaveformExtractor.py -d

The debug script debug_fourier.py allows you to visualise the reconstructed waveform against the original points taken.

last update: 27/04/2024