Automatically measured birch pollen grains from the KM23B and Li-BPa records

The data set includes pollen grain images produced for the manuscript "First automatic size measurements for the separation of dwarf birch and tree birch pollen in MIS 6 to MIS 1 records from Northern Germany". 

The images are cut-outs of birch pollen produced to automatically determine the size (area and mean diameter) of birch pollen grains. 

The workflow includes the following three steps, each of which is conducted using trained convoulutional neural networks: 

automatic detection and classification of pollen with the Tofsi algorithm (Theuerkauf et al. 2023, https://doi.org/10.1177/09596836231211876)

automatic distinction between pollen grain suitable (i.e. in polar view) or not suitable (some other orientation) for size measurement 

automatic outline detection and size measurement (the green area indicates the area that was identified as the area of each pollen grain)