Mobile App Identify Result for Metalmarks

Bug Reports
1

Usually, I’ve found the eButterflyAI tool quite good at butterfly image identification. Not too surprised when it misses, but was that it doesn’t come close with my local Metalmarks (Apodemia mejicanus, A. mormo, A. nais, A. palmerii), while SEEK and even Google Lens does. Any thoughts? Thanks!

1 Like
2

Your comparison is quite interesting, and we would love it if you could share those images with us. We can use them explicitly when we retrain our model.

Also, please keep in mind that our algorithm uses location and date to fine-tune its results, unlike Google. So, when trying out your images, please make sure to set the actual date and location when you take them.

Thanks

-Rodrigo

3

Thanks for getting back to me so quickly. I was wondering if that might be the explanation; will get some of those images (with date and location) to you soon.

1 Like
4

Sounds good. Feel free to share them here or email them directly to me at ebutterfly@vtecostudies.org

Thanks

-Rodrigo

5

Thanks, Rodrigo. Ended up sending them directly to your email in four emails. Please let me know if there’s anything else I can do to help.
-Joe

1 Like
6

Excellent! I am on a meeting right now, will check those images and will get back to you soon

-Rodrigo

7

I am having the same problem. All of the 2024 iNaturalist fatal metalmarks were mis-IDed as Rawson’s by the eButterfly App:

Observations · iNaturalist

Andrew Hogan