We are a group of researchers motivated to improve the Inat’s CV capacity to ID earthworms species. (Curated collection of Earthworms ID project) .
We would like to have insights on the evolution of the iNat’s CV capacity over time to identify species of this group of taxa.
This would be helpful for us to show the progress that were made and to show what is still necessary to do, in order to recruit other colleagues to help us, and also to see if our efforts are useful.
So my question is: is there a way to use different former versions of the CV model to estimate its capacity to identify different species? Like different version of the CV demo ? (or API, but this looks touchy)
We were thinking to test it with the same group of annotated images on the different versions of the CV.
Thanks