Natural Imaging Data and Artifacts - Konnsy/REAML2022-hackathon GitHub Wiki

When working with sensor image data, imaging artifacts should always be taken into account. This applies most of all to on-site use, but can also not be avoided completely in a laboratory environment. Influences like external impacts and contaminations of the analyzed samples, e.g. with dust or air bubbles, but also imperfect calibration of the optical instruments and temperature dependencies have the potential to make a detection of nanoparticles, especially smaller particles, harder.

Spatial examples

Examples for imaging artifacts can be seen in the preprocessed images of datasets recorded earlier.

Not only external influence can cause artifacts. Below you can see gold foil with defects that are visible as plain gray regions.

Temporal examples

Also on the temporal level, artifacts become visible.

Here you can see examples for time series (mean pixel intensities per frame) for regions with attaching particles of interest:

And here you can negative examples, i.e., examples showing time series where no particle of interest is present: