Lightroom: AI Data management automation to identify and tag plants using taxonomic keys & morphology

  • 2
  • Idea
  • Updated 2 months ago
  • (Edited)
Please excuse errors and not sure where to post this. i request the ability to identfy tree saplings grown in nurseries and eventually full grown trees and their leaf or flower samples serving as the keys and we can identify and begin to put together a database of every species of plants. I would like to contrribute my time as a human manually identifying the plants for the adobe ai facial recognition team, volunteer to identify the plants if that will help with the process. So then i can photograph and then face recognition for ash trees, which are being lost of a disease carried by the emerald ash bore as its vector and if we could crowd source the inventory then we can measure its progress as the disease vector moves northward according to the rising temperatures. When perfomance of urban forest populations is measured then the performance will improve. And when performace is measuered and recorded and catalogued then the rate of improvement will increase.
Photo of Aaron Edwards

Aaron Edwards

  • 2 Posts
  • 0 Reply Likes
  • concerned we will lose so many Ash trees to EAB

Posted 2 months ago

  • 2
Photo of John R. Ellis

John R. Ellis, Champion

  • 3690 Posts
  • 963 Reply Likes
I think it's highly unlikely Adobe itself would take such a specific project on.

However, there are a number of companies and open-source packages that make it possible for non-experts to build custom object-recognition models using "training sets" of manually labeled photos.  E.g. see this article:
https://medium.com/coinmonks/watson-visual-recognition-custom-models-9cbcc20a6f98

With such a custom model, it would then be straightforward for a programmer to create a Lightroom plugin that applied the custom object recognition to selected photos and tagged them with the result (e.g. with keywords).  
(Edited)
Photo of Aaron Edwards

Aaron Edwards

  • 2 Posts
  • 0 Reply Likes
thank you i appreciate the link