How to read these slides?

Navigate via arrow keys between slides Overview

Click on the menu bar items to navigate to chapters

Click here for PDF version
SCIENCE PASSION TECHNOLOGY

Visual Fingerprints of Vibration Signals
Using Time Delay Embeddings

Julian Rakuschek, Adrian Boesze,
Johanna Schmidt, Tobias Schreck

05.06.2025 - EuroVis Short Paper

The Problem with Vibrations

Can you tell the difference?

A hidden signal

You cannot see the hidden signal with a line chart.

The Idea: Vibration $\rightarrow$ Point Cloud

1. Sliding Window View
2. Apply PCA

Press "Arrow Down" to see other dimensionality reduction techniques

Projecting Time Series has been explored already

M. Ward et al. "Visual Exploration of Time‐Series Data with Shape Space Projections" Computer Graphics Forum 30 (2011)
J. Bernard et al. "TimeSeriesPaths : Projection-Based Explorative Analysis of Multivariate Time Series Data" 20 (2012): 97-106

But not with Vibrations!

Noise is not exciting ...

... but oscillations result in circles!

Grounded in Theory

Sliding Windows and Persistence: An Application of Topological Methods to Signal Analysis

Jose Perea and John Harer, Foundations of Computational Mathematics (2015).

They explained: Why the circles?

Yes, this is also detectable with spectrograms, but ...

... the TDE is suitable to gain an overview!

Why the coloring?

Through the turbo coloring, the fingerprint guides users towards vibrations within the signal.

Press "Arrow Down" to see other colormaps

Viridis

Plasma

Cividis

Jet

RdYlBu

bwr

The only parameter: Window Size

Large window sizes in general advisable

The only parameter: Window Size

Large window sizes in general advisable

TDE under Noise Influence

We can use the fingerprints to observe change ...

... and analyze the radii on a global scale

We can compare fingerprints through the radius distribution

Clustering

Observing Wear

Conclusion

  • Detecting periodicity through fingerprints
  • Robust to window size parameter and noise to some degree
  • Applicable to detect wear

Future Work

  • Scalability: Visually and computationally
  • Application to stream monitoring
  • Counterfactuals: Effect of data changes

Questions?

Slides:

Corresponding Author: Julian Rakuschek

julian.rakuschek@tugraz.at