Automated Pollen Classification (von Allmen et al., 2024)

This basic template helps you to test whether different supplemented images will correctly be classified by the model trained in von Allmen et al. (2024) for a few taxa important for early post glacial period (i.e., Apiaceae, Artemisia, Betula, Chenopodiaceae, Cyperaceae, Helianthemum, Pinus, Poaceae, Salix), diverse non-pollen palynomorphs (i.e.,NPP classes 1 to 4), as well as other proxies (i.e., Lycopodium)

For the model to be able to evaluate your image, make sure that the following applies to your input:

  • The image is cut to the specific object of interest.
  • The format of the image should be squared (1:1).
  • The resolution of the image should be above 120x120 pixels. (strong deviations may result in artifacts through preprocessing steps)
  • The color scheme of the image must be RGB (greyscale and CMYK are not accepted).
  • OBS: Remove any text from the images if possible.

    • For guidance, here are some reference images, downloaded online, which were successfully classified:



    Images Uploaded: 0

    Automated Analysis