search

NHMDenmark / Pinned-Insects-workstation

Work on the Pinned Insects workstation and workflow for mass digitisation in Denmark (DaSSCo)
0 stars 0 forks source link

How to address problems encountered digitizing Dung Beetles #68

Open chelseagraham opened 7 months ago

chelseagraham commented 7 months ago

On 1 February 2024, Pip and Chelsea met with Alexey and Lucas to discuss problems encountered during Dung Beetle digitization. They were later summarized by Alexey in an email from the same date as

  1. Manual handling of specimens demanded by the current protocol is time consuming
  2. Condition of some specimens (thin, blunt pins etc., certain type of labels etc) make their handing even more time consuming.
  3. Examples of indirectly labelled specimens (i.e. they do not have their own label but in some way associated with a certain properly labelled specimen) are common.
  4. Examples of completely unlabeled specimens (no any label information) are not rare.
  5. A combination of cases 2, 3 and 4 is very common because indirectly labelled or completely unlabeled specimens are mostly from the old collections and thus with bad pins. <<

The following day, Lucas sent illustrative examples

Image Image Image

chelseagraham commented 5 months ago

@PipBrewer I think that the images Lucas shared would be illustrative enough for digitizers to understand which specimens to image, how to count the specimens connected and where to resume imaging, if we go that route.

PipBrewer commented 1 month ago

@PipBrewer I think that the images Lucas shared would be illustrative enough for digitizers to understand which specimens to image, how to count the specimens connected and where to resume imaging, if we go that route.

I would have no idea (as a digitiser) looking at those images and explanations as to how to spot these myself and feel confident. I don't think we could add this into a workflow without clearer instructions or examples.

chelseagraham commented 1 day ago

As communicated in the June Catch-Up #76 https://github.com/NHMDenmark/Pinned-Insects-workstation/issues/76#issuecomment-2206029287

We would like to avoid skipping specimens for now in order to fully test the current model. This does not mean that the suggestion has been taken off the table completely, but that we want to fully explore the current methodology right now, before adjusting our workflow

We therefore need to decide if we require a datamatrix in these instances (like the MOS herbarium sheets without labels), or if the shared labels will remain on the angle until the final specimen how to include this in our workflow how to train digitizers document the protocol in our guides train digitizers in the protocol